language en

Neuroelectrophysiology Analysis Ontology - Analysis Steps

Release: 2024-12-06

This version:
http://purl.org/neao/0.1.0/steps#
Latest version:
http://purl.org/neao/steps#
Revision:
0.1.0
Issued on:
2024-12-06
Authors:
Cristiano Köhler, Forschungszentrum Jülich
Michael Denker, Forschungszentrum Jülich
License:
https://creativecommons.org/licenses/by/4.0/
Visualization:
Visualize with WebVowl
Provenance of this page
Ontology Specification Draft

Introduction back to ToC

This module in the Neuroelectrophysiology Analysis Ontology (NEAO) contains classes that represent (atomic) steps in the analysis of neuroelectrophysiology data.

Namespace declarations

Table 1: Namespaces used in the document
biro<http://purl.org/spar/biro/>
dcterms<http://purl.org/dc/terms/>
neao_base<http://purl.org/neao/base#>
neao_bib<http://purl.org/neao/bibliography#>
neao_data<http://purl.org/neao/data#>
neao_steps<http://purl.org/neao/steps#>
owl<http://www.w3.org/2002/07/owl#>
rdf<http://www.w3.org/1999/02/22-rdf-syntax-ns#>
rdfs<http://www.w3.org/2000/01/rdf-schema#>
skos<http://www.w3.org/2004/02/skos/core#>
vann<http://purl.org/vocab/vann/>
xml<http://www.w3.org/XML/1998/namespace>
xsd<http://www.w3.org/2001/XMLSchema#>

NEAO Analysis Steps: Overview back to ToC

This ontology has the following classes and properties.

Classes

Object Properties

Data Properties

NEAO Analysis Steps: Description back to ToC

The main classes are subclasses of the AnalysisStep class (defined in the base module), and represent the different methods and procedures used in the analysis to generate new data or to perform specific operations aimed to transform or extract additional information from the input(s). The classes are organized in a taxonomy, and the lowest level of the hierarchy represent specific, fine-grained descriptions of a step used in the analysis.

Grouping classes are provided to identify analysis steps according to their semantic similarities. This is used to keep the fine-grained descriptions associated with specific analysis methods (e.g., a step that used either the Welch or multitaper method to compute the power spectral density) while providing the ability to identify the steps in a more general nature (e.g., a step that computed a power spectral density). The grouping classes are defined in the taxonomy hierarchy by subclasses, and also as classes inferred by reasoning with the hasPurpose object property (that points to individuals of the AnalysisPurpose class, such as FunctionalConnectivityPurpose) or data properties defining boolean values (e.g., isBivariate). Therefore, groupings across multiple semantic dimensions are available (e.g., bivariate analyses or functional connectivity analyses).

The information regarding the inputs and outputs can be associated with each step by the hasInput and hasOutput object properties that point to individuals that represent data entities (using the classes defined in the data module).

The parameters used to control the behavior of the analysis step are defined with the usesParameter object property that points to individuals that represent specific parameters (using the classes defined in the parameters module).

The relevant bibliographic references are provided with the hasBibliographicReference annotation property that points to individuals of the biro:BibliographicReference class (defined in the bibliography module). This structures the specific information associated with the analysis step represented by the class, and helps to disambiguate the description of the diversity of methods that are available to analyze neuroelectrophysiology data. This would be the case, for example, of different algorithms and assumptions (e.g., computing the power spectral density using either the Welch or multitaper approach), and the evolution/modifications of a method (e.g., different computations of phase lag index estimates).

The details of the software code associated with each analysis step can be provided by the isImplementedIn object property, that points to individuals of the SoftwareImplementation class (defined in the base module).

Finally, the hasSubstep object property is used to describe compound analyses. These analyses involve the execution of multiple smaller steps (substeps) that are associated with specific parameters and intermediary data inputs and outputs. The main compound process can point to other individuals of the AnalysisStep class using the hasSubstep property. Therefore, the appropriate semantic information and description associated with either the compound analysis or each individual substep can be provided in the ontology.

Cross-reference for NEAO Analysis Steps classes, object properties and data properties back to ToC

This section provides details for each class and property defined by NEAO Analysis Steps.

Classes

analysis purposec back to ToC or Class ToC

IRI: http://purl.org/neao/steps#AnalysisPurpose

Analysis purpose refers to the specific objective or goal that an analysis is intended to achieve. It outlines what the analysis step aims to discover, understand, or demonstrate. This class is intended for grouping steps that perform analyses with similar goals/outputs but with distinct methodological or algorithmic approaches. It is used as a normalization class via the hasPurpose object property.
is in range of
has purpose op
has members
correlation purpose ni, data smoothing purpose ni, distance purpose ni, field-field coupling purpose ni, functional connectivity purpose ni, instantaneous firing rate purpose ni, latent dynamics purpose ni, neural synchronization purpose ni, neuronal firing regularity purpose ni, spike-field coupling purpose ni, spike-spike coupling purpose ni

apply adaptive kernel smoothingc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ApplyAdaptiveKernelSmoothing

A kernel smoothing that uses a variable-width kernel. The kernel width varies adaptively depending on the local density of the data. This method takes into consideration local variations in data density, resulting in a more accurate representation of the underlying patterns and structures.
has super-classes
kernel smoothing c

apply analytic signal conversionc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ApplyAnalyticSignalConversion

A data transformation that uses the Hilbert transform of a real-valued input time series to construct the analytic signal, a complex-valued time series where the real part is the original real-valued signal and the imaginary part is the Hilbert transform. The analytic signal does not have negative frequency components.
has super-classes
data transformation c

apply Butterworth filterc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ApplyButterworthFilter

An infinite impulse response filtering that uses a Butterworth type filter, i.e., a filter designed to have a maximally flat frequency response in the passband (no ripples). The frequency response gradually decreases to zero in the stopband, and the steepness of the decrease (roll-off) is controlled by the order of the filter.
has super-classes
infinite impulse response filtering c

apply canonical polyadic tensor decompositionc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ApplyCanonicalPolyadicTensorDecomposition

A tensor component analysis (TCA) that expresses the input high-dimensional tensor as a sum of rank-one tensors (tensor components). Each dimension in the tensor component corresponds to a dimension in the input high-dimensional tensor. For neural data, a tensor could be used to represent trial-by-trial spiking activity, with neurons, time, and trials as dimensions. Therefore, each tensor component produced from that input will have a rank-one tensor for the neurons, time, and trial, which describes both within- and between-trial changes.
has super-classes
tensor component analysis c
is disjoint with
apply coupled canonical polyadic tensor decomposition c, apply non-negative tensor component analysis c

apply coupled canonical polyadic tensor decompositionc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ApplyCoupledCanonicalPolyadicTensorDecomposition

A tensor component analysis (TCA) that expresses multiple high-dimensional input tensors as a sum of rank one tensors (tensor components). The output tensor components have shared vectors that summarize an input dimension across all input tensors. For example, if analyzing two tensors, each representing trial-by-trial spiking activity obtained from a distinct experimental subject (neurons X time X trials), the coupled canonical polyadic tensor decomposition (CCPD) could produce tensor components for each dataset where the vector for the trial and time dimensions are the same, but the neuron dimension is unique for each dataset (hence, each subject). Therefore, CCPD is useful for scenarios with multiple and related datasets, allowing for the exploitation of shared information to enhance the decomposition results.
has super-classes
tensor component analysis c
is disjoint with
apply canonical polyadic tensor decomposition c, apply non-negative tensor component analysis c

apply data concatenationc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ApplyDataConcatenation

A data transformation that joins two or more data inputs into a single data element. For example, in a multitrial experimental session, where data was acquired separately as one epoch per trial, a continuous data segment for the session can be constructed by concatenating all trial epochs together.
has super-classes
data transformation c

apply demixed principal component analysisc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ApplyDemixedPrincipalComponentAnalysis

A principal component analysis (PCA) that uses a modified version of the standard PCA for neural activity data analysis. Demixed PCA (dPCA) not only obtains a low-dimensional representation of the input data, but it also demixes the dependencies of the population activity on the task parameters. Therefore, dPCA can show the dependence of the neural representation on parameters such as stimuli, subject decisions, or rewards.
has super-classes
principal component analysis c
is disjoint with
apply probabilistic principal component analysis c, apply standard principal component analysis c

apply discrete Fourier transformc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ApplyDiscreteFourierTransform

A frequency-domain transformation that applies the discrete Fourier transform (DFT) to an input time series acquired in equally-spaced samples. The DFT is used to obtain the frequency representation of the time-domain input. The DFT output is a sequence of coefficients of complex sinusoids, each representing a frequency component in the input signal. Each frequency component consists of an interval (or bin), and the width of the bin determines the frequency resolution of the DFT. The number of frequency components and the frequency resolution is determined by the length of the input signal (number of samples) and the sampling frequency. For large datasets, the computation is computationally expensive (O(N^2) complexity).
has super-classes
frequency domain transformation c
has sub-classes
apply fast Fourier transform c

apply discrete Fourier transform noise removalc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ApplyDFTNoiseRemoval

A line noise removal that uses a discrete Fourier transform (DFT) filter and estimates the power line component amplitude in the input data by fitting a sine and cosine at a user-specified line noise frequency (e.g., 50 Hz), followed by the subtraction of those components from the input.
has super-classes
line noise removal c

apply distance covariance analysisc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ApplyDistanceCovarianceAnalysis

A dimensionality reduction that identifies linear and nonlinear relationships between multiple input datasets. The method identifies linear projections (DCA dimensions) that maximize the distance covariance statistic (an Euclidean-based correlational statistic). For example, for recordings from different brain regions (two neuronal populations), the distance covariance analysis (DCA) can identify the dimensions in the population activity in the different brain areas that are related to each other. The dimensionality reduction can also take other dependent variables into account (e.g., stimulus or behavioral variables).
has super-classes
dimensionality reduction c

apply downsamplingc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ApplyDownsampling

A resampling that reduces the number of samples in the input data (i.e., reduces the sampling frequency). This is often accomplished after applying an anti-aliasing filter (i.e., to remove frequencies above half the value of the new sampling frequency).
has super-classes
resampling c
is disjoint with
apply upsampling c

apply fast Fourier transformc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ApplyFastFourierTransform

A frequency-domain transformation that uses the fast Fourier transform (FFT) algorithm to compute the discrete Fourier transform (DFT). The computation of the DFT is computationally expensive for large datasets. The FFT reduces the number of computations significantly by using a divide-and-conquer approach, leveraging the symmetry and periodicity properties of the DFT. The complexity is O(N log N). The output of the method is similar to the DFT, i.e., a sequence of coefficients of complex sinusoids, each representing a frequency component in the input signal. The FFT takes a size parameter, that refers to the number of points in the input data sequence that is used for the computation. The FFT size determines the frequency resolution, and, for maximal efficiency, should be a multiple of 2.
has super-classes
apply discrete Fourier transform c

apply finite impulse response filterc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ApplyFiniteImpulseResponseFilter

A finite impulse response (FIR) filtering where a custom-designed FIR filter is applied to the input data.
has super-classes
finite impulse response filtering c

apply finite impulse response filter with Kaiser windowc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ApplyFiniteImpulseResponseFilterKaiserWindow

A finite impulse response (FIR) filtering that uses a FIR filter whose impulse response is controlled by applying a Kaiser window function.
has super-classes
finite impulse response filtering c

apply fixed kernel smoothingc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ApplyFixedKernelSmoothing

A kernel smoothing that uses a fixed-width kernel. The kernel type and kernel width can be specified as parameters.
has super-classes
kernel smoothing c

apply general linear model polynomial detrendingc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ApplyGeneralLinearModelPolynomialDetrending

A detrending that uses a general linear model to fit a polynomial from the input data and remove the mean and linear trend.
has super-classes
detrending c

apply Hilbert transformc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ApplyHilbertTransform

A data transformation that applies the Hilbert transform to a real-valued time series. The Hilbert transform shifts the phase of each frequency component of the signal by 90 degrees: positive frequencies are shifted by -90 degrees, and negative frequencies are shifted by +90 degrees. This can be used to construct the analytic signal.
has super-classes
data transformation c

apply independent component analysisc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ApplyIndependentComponentAnalysis

A data transformation that separates a multivariate input signal into additive subcomponents (independent components). The independent components are statistically independent from each other.
has super-classes
data transformation c

apply infinite impulse response filterc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ApplyInfiniteImpulseResponseFilter

An infinite impulse response (IIR) filtering where a custom-designed IIR filter is applied to the input data.
has super-classes
infinite impulse response filtering c

apply interpolationc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ApplyInterpolation

A data transformation that estimates new (intermediate) values between the known values in the input data. This can be accomplished using several methods, such as linear interpolation (i.e., estimating the values along a straight line connecting adjacent points), polynomial interpolation (i.e., using polynomials to estimate the values between points), or spline Interpolation (i.e., using piecewise polynomials that pass through the known data points and provide a smooth curve).
has super-classes
data transformation c
has sub-classes
apply spike waveform interpolation c

apply linear discriminant analysisc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ApplyLinearDiscriminantAnalysis

A dimensionality reduction that finds a linear combination of features that separates two or more classes defined in the input data. The input data must have a variable that defines the class for each observation, and continuous variables that are used for the linear discriminant analysis (LDA). LDA finds an optimal projection vector that maximizes the distance between the means of the different classes, and minimizes the variance within each class. LDA projects the input data into the lower-dimensional space, therefore reducing the number of features while retaining the information needed for classification. The optimal projection vectors found by LDA are the Fisher linear discriminants.
has super-classes
dimensionality reduction c

apply local linear regression detrendingc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ApplyLocalLinearRegressionDetrending

A detrending that removes a running line fit using local linear regression. Local linear regression estimates a function by fitting a low-order polynomial to data within a sliding window (local neighborhood) across the input data.
has super-classes
detrending c

apply local regression and likelihood smoothingc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ApplyLocalRegressionAndLikelihoodSmoothing

A data smoothing that estimates a low-order polynomial in a local neighborhood (window) of any value in the input data. Polynomial coefficients are estimated using the least mean squares method. Contrary to kernel smoothing methods, this is a non-parametric approach and has reduced bias at the boundaries of the input data.
has super-classes
data smoothing c

apply median rescalingc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ApplyMedianRescaling

A data normalization that uses the median and interquartile range (IQR) to rescale the values of the input data. This method is less sensitive to outliers (compared to the z-score transform), and therefore is known as robust scaling.
has super-classes
data normalization c

apply min-max normalizationc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ApplyMinMaxNormalization

A data normalization that adjusts the range and distribution of values in the data input such that they fall within a fixed range, based on the minimum and maximum values in the input data. In the typical case, the values are normalized to the [0, 1] interval.
has super-classes
data normalization c

apply movement artifact removalc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ApplyMovementArtifactRemoval

An artifact removal that identifies and removes artifacts originating from movements of the experimental subject (e.g., eye movement, head movement).
has super-classes
artifact removal c

apply neural trajectory Gaussian process factor analysisc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ApplyNeuralTrajectoryGaussianProcessFactorAnalysis

A dimensionality reduction that uses the Gaussian process factor analysis (GPFA) method described by Yu et al. (2009). GPFA extracts smooth, low-dimensional neural trajectories that summarize the activity recorded simultaneously from many neurons on individual experimental trials over time. The input is a set of spike trains representing multitrial activity of multiple neurons recorded in parallel. The input spike trains are binned, and factor analysis is applied to reduce the dimensionality while smoothing the resulting low-dimensional trajectories by fitting a Gaussian process (GP) model to them. The identified trajectories are called neural trajectories, and show the evolution of the activity of the population of neurons over time.
has super-classes
dimensionality reduction c

apply non-negative tensor component analysisc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ApplyNonNegativeTensorComponentAnalysis

A tensor component analysis (TCA) where a non-negative constraint is applied to the decomposition. This is desirable when the underlying components have physical interpretation and negative values are not possible.
has super-classes
tensor component analysis c
is disjoint with
apply canonical polyadic tensor decomposition c, apply coupled canonical polyadic tensor decomposition c

apply notch filterc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ApplyNotchFilter

An infinite impulse response filtering that uses a filter designed to attenuate or eliminate a narrow band of frequencies in the input data (stopband) while allowing other frequencies to pass through relatively unaffected.
has super-classes
infinite impulse response filtering c

apply notch filter noise removalc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ApplyNotchFilterNoiseRemoval

A line noise removal that employs a notch filter to remove line noise. A notch filter is designed to attenuate or eliminate a narrow band of frequencies in the input data (stopband) while allowing other frequencies to pass through relatively unaffected. The notch filter stop band is usually centered at the power line frequency (i.e., 50 Hz or 60 Hz depending on the location).
has super-classes
line noise removal c

apply outlier removalc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ApplyOutlierRemoval

An artifact removal that identifies and removes values in the input data that differs significantly from other values (outliers). Outliers may arise from the variability in the measurement or be the result of experimental error.
has super-classes
artifact removal c

apply paddingc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ApplyPadding

A data transformation that adds extra data (often zeros or other predefined values) to the beginning, end, or both sides of the input data.
has super-classes
data transformation c

apply probabilistic principal component analysisc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ApplyProbabilisticPrincipalComponentAnalysis

A principal component analysis (PCA) that assumes a probabilistic model for the generation of the observed data, according to Tipping & Bishop (1999). The model assumes that the values in the data input are generated from the lower-dimensional subspace of latent variables (principal components) plus an additive Gaussian noise. This generalizes the standard PCA for the case where the noise covariance approaches zero. The probabilistic PCA (pPCA) allows for uncertainty estimation and modeling of the data generation process, and can be employed when there are missing values in the input data.
has super-classes
principal component analysis c
is disjoint with
apply demixed principal component analysis c, apply standard principal component analysis c

apply rectificationc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ApplyRectification

A data transformation that computes the absolute value of the input data (rectification).
has super-classes
data transformation c

apply rereferencec back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ApplyRereference

A data transformation that changes the reference point of the data recorded from an electrode. This can be performed by calculating the average across the data obtained from all electrodes and subtracting it from each individual electrode’s data (reducing common noise) or referencing each electrode to its nearest neighbor or a defined pair, subtracting one signal from another (bipolar referencing).
has super-classes
data transformation c

apply spectrum interpolation noise removalc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ApplySpectrumInterpolationNoiseRemoval

A line noise removal that uses spectral interpolation to remove power line noise. After obtaining the discrete Fourier transform (DFT) of the input signal with noise, the original frequency component at the power line oscillation frequency can be estimated by interpolating the amplitude spectrum (obtained from the DFT) at the power line frequency (e.g., 50 Hz), followed by the inverse DFT to reconstruct the signal.
has super-classes
line noise removal c

apply spike extraction from time seriesc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ApplySpikeExtractionFromTimeSeries

A data transformation that obtains a series of spike times (i.e., a spike train) from an input time series (e.g., voltages recorded from an electrode). The spike times can be estimated, for example, by taking all the time points where the values in the input data are greater or lower than a threshold value.
has super-classes
data transformation c

apply spike train binarizationc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ApplySpikeTrainBinarization

A data transformation that takes an input spike train and returns an array of boolean values indicating if at least one spike occurred at individual time points.
has super-classes
data transformation c

apply spike train binningc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ApplySpikeTrainBinning

A data transformation that performs a binning operation on the input spike train data. The transformation discretizes the duration of the input spike train(s) into smaller time intervals (bins), and obtains the number of spikes occurring into each bin (binned spike train). Additionally, the occurrence of spikes into each bin can be converted into a binary form (i.e., bins with or without spikes). This is known as clipping. The width of the bin interval is specified by a parameter (bin size).
has super-classes
data transformation c

apply spike waveform interpolationc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ApplySpikeWaveformInterpolation

A data transformation that estimates additional (unknown) values between sample points of a spike waveform input.
has super-classes
apply interpolation c

apply spike waveform outlier rejectionc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ApplySpikeWaveformOutlierRejection

An artifact removal that identifies and removes spike waveforms that differ significantly from the other spike waveforms in the input. This usually involves identifying waveforms with too late peaks, or in which the rising phase of the potential does not align with the peaks of all other waveforms in the input.
has super-classes
artifact removal c

apply spike waveform peak alignmentc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ApplySpikeWaveformPeakAlignment

A data transformation that modifies input spike waveforms in order to align their peak values in time.
has super-classes
data transformation c

apply standard principal component analysisc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ApplyStandardPrincipalComponentAnalysis

A principal component analysis (PCA) that operates by computing the covariance matrix of the input data, which is then decomposed into its eigenvectors and eigenvalues. The eigenvectors, corresponding to the principal components (PCs), are sorted by the magnitude of their associated eigenvalues. The eigenvectors with the largest eigenvalues explain the most variance in the data and thus form the primary PCs.
has super-classes
principal component analysis c
is disjoint with
apply demixed principal component analysis c, apply probabilistic principal component analysis c

apply stimulation artifact removalc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ApplyStimulationArtifactRemoval

An artifact removal that identifies and removes artifacts originating from presenting a stimulus during the recording (e.g., electrical stimulation).
has super-classes
artifact removal c

apply sumc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ApplySum

A data transformation that performs the addition of two or more data inputs to obtain a sum.
has super-classes
data transformation c

apply synchronous spike removalc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ApplySynchronousSpikeRemoval

An artifact removal that identifies and removes spikes across two or more spike train inputs that occurred simultaneously within a temporal precision specified by a parameter. The temporal precision is usually the sampling rate used by the recording equipment: if different neurons fired within an interval equal to or smaller than the sampling period, this suggests that this synchronous activity does not come from temporal synchronization of the neurons but rather due to an interference in the recording (e.g., electrical noise picked simultaneously by multiple channels).
has super-classes
artifact removal c

apply Thomson regression noise removalc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ApplyThomsonRegressionNoiseRemoval

A line noise removal that uses Thomson's regression method (1982) for detecting sinusoids, that identifies and removes significant sine waves from the input data. The desired frequencies can be specified by parameter, or determined using an F-statistic.
has super-classes
line noise removal c

apply trial extractionc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ApplyTrialExtraction

A data transformation that extracts trial segments from an input data entity storing a longer and continuous data stream. A trial is a single instance of a repeated experimental procedure. For example, during an electrophysiology experiment, a visual stimulus might be presented several times, with neural activity recorded each time. Each presentation of the stimulus defines a trial. Since the data is recorded continuously, this data transformation identifies and isolates the segments of data corresponding to each individual stimulus presentation, returning them as separate data entities.
has super-classes
data transformation c

apply upsamplingc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ApplyUpsampling

A resampling that increases the number of samples in the data (i.e., increases the sampling frequency). This is often accomplished by adding new (zero-valued) samples between existing ones followed by applying a lowpass filter to replace the zeros and smooth out the discontinuities.
has super-classes
resampling c
is disjoint with
apply downsampling c

apply windowed-sinc filterc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ApplyWindowedSincFilter

A finite impulse response filtering that employs the convolution with a sinc function kernel multiplied by a window function (e.g., Blackman or Hamming). The kernel is obtained by evaluating the sinc function for the cutoff frequencies specified as parameters, followed by truncation of the filter skirt, and applying the window to reduce the artifacts introduced from the truncation. The windowed-sinc filter is stable and very efficient to separate one band of frequencies from another.
has super-classes
finite impulse response filtering c

apply z-score transformc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ApplyZscoreTransform

A data normalization that transforms the values of the input data to have zero mean and unit variance.
has super-classes
data normalization c

artifact removalc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ArtifactRemoval

A data transformation that aims to identify and remove artifacts from the input data. Artifacts are unwanted disturbances that distorts the data. In an electrophysiology experiment, they can arise from various sources: environmental interference (e.g., electromagnetic interference from nearby equipment), physiological processes (e.g., eye movement, heart beat), and technical instrumentation issues (e.g., baseline drift of the recorded potentials).
has super-classes
data transformation c
has sub-classes
apply movement artifact removal c, apply outlier removal c, apply spike waveform outlier rejection c, apply stimulation artifact removal c, apply synchronous spike removal c

artificial data generationc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ArtificialDataGeneration

A data generation that produces artificial data. Artificial data refers to data that is generated programmatically rather than obtained from experimental recordings (e.g., neural simulations or by using specific statistical procedures). The artificial data generation procedure does not take experimentally-recorded data as input, and the generation of the output data depends only on parameters to the method employed. The isArtificial data property is defined as True for outputs of artificial data generation steps.
is equivalent to
has output op some artificial data c
has super-classes
data generation c

ASSET analysis probability matrix substepc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ASSETAnalysisProbabilityMatrixSubstep

ASSET analysis substep that computes the probability matrix (PMAT). The probability matrix contains, for each entry in the intersection matrix, the cumulative probability representing the probability of having the overlap in the IMAT under the assumption that the spike trains are independent. If an entry in the PMAT is large, the null hypothesis of independence is rejected, and the alternative hypothesis that the observed overlap reflects active synchronization between the involved neurons at the time bins of the intersection is accepted. The PMAT computation can be done with either an analytical or Monte Carlo approach.
has super-classes
ASSET analysis substep c
has sub-classes
compute ASSET probability matrix (Monte Carlo method) c, compute ASSET probability matrix (analytical method) c

ASSET analysis substepc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ASSETAnalysisSubstep

An analysis step that is an individual part of the ASSET analysis method to identify neuronal activity patterns in spike train data.
has super-classes
analysis step c
has sub-classes
ASSET analysis probability matrix substep c, compute ASSET cluster matrix c, compute ASSET intersection matrix c, compute ASSET joint probability matrix c, compute ASSET mask matrix c, compute ASSET sequence of synchronous events extraction c

autocorrelation analysisc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#AutocorrelationAnalysis

An analysis step used to compute a measure of autocorrelation, i.e., the correlation of the input with displaced (lagged or advanced) versions of itself. The computation produces the autocorrelation value for every lag considered.
has super-classes
analysis step c
has sub-classes
compute autocorrelation function c, compute spike train autocorrelation histogram c

bivariate analysisc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#BivariateAnalysis

An analysis step that has only two distinct inputs considered for the computation of the output (e.g., the two time series with the local field potential recorded from two electrodes used to compute a cross-correlation).
is equivalent to
is bivariate dp value true
has super-classes
analysis step c
is disjoint with
multivariate analysis c

central tendency statistical analysisc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#CentralTendencyStatisticalAnalysis

A statistical analysis to compute a measure of central tendency, i.e., that represents the center or typical value of the input data.
has super-classes
statistical analysis c
has sub-classes
compute mean c, compute median c

coherence analysisc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#CoherenceAnalysis

A spectral analysis that computes a measure of coherence between two or more inputs. Coherence is a real measure that describes the linear association of the distinct inputs (e.g., two time series) in different frequency bands. It corresponds to the absolute value of the coherency.
has super-classes
spectral analysis c
has sub-classes
compute canonical coherence c, compute coherence c, compute partial coherence c

coherency analysisc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#CoherencyAnalysis

A spectral analysis that computes a measure of coherency between two inputs. Coherency is a complex-valued measure that describes the linear association of the distinct inputs (e.g., two time series) in different frequency bands.
has super-classes
spectral analysis c
has sub-classes
compute coherency c, compute imaginary coherency c, compute maximized imaginary coherency c

compound analysisc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#CompoundAnalysis

An analysis step that is composed by two or more substeps, each performing a part of the analysis with its own data inputs/outputs and analysis parameters.
is equivalent to
has substep op some analysis step c
has super-classes
analysis step c

compute angular mean of spike phasesc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ComputeAngularMeanSpikePhases

A phase analysis that computes the angle obtained from averaging the phases of an input signal at the time points where spikes occurred. For the computation, the phases are represented as vectors in the unit circle, the mean phase vector is computed, and the angle is extracted.
has super-classes
phase analysis c

compute ASSET cluster matrixc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ComputeASSETClusterMatrix

ASSET analysis substep that computes the cluster matrix (CMAT) in the ASSET analysis, using DBSCAN with a modified distance metric. It takes the mask matrix (MMAT) as input. The cluster matrix groups the significant entries in the MMAT according to each diagonal structure that they belong to. For each significant entry in the MMAT, the CMAT will have an integer value: -1 for significant entries that do not belong to any diagonal structure, or any value greater than zero with the identification of the cluster that the entry belongs to.
has super-classes
ASSET analysis substep c

compute ASSET intersection matrixc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ComputeASSETIntersectionMatrix

ASSET analysis substep that computes the intersection matrix (IMAT). For a set of input spike trains, binned with a bin width, each entry in the IMAT corresponds to a pair of distinct bins (i.e., distinct time points in the data). The value in the entry corresponds to the number of neurons that fired in both bins corresponding to that entry. When groups of neurons fire in a sequence that repeats in time, the IMAT will show patterns that follow a diagonal direction (diagonal structure). The ASSET method aims to identify the diagonal structures by automated statistical testing and clustering procedures.
has super-classes
ASSET analysis substep c

compute ASSET joint probability matrixc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ComputeASSETJointProbabilityMatrix

ASSET analysis substep that computes the joint probability matrix (JMAT). For every entry in the probability matrix (PMAT), the computation produces the combined probability of a fixed number of neighbors in a rectangular kernel (with fixed length and width as parameters) covering a diagonal structure. A value in the JMAT reflects how likely entries with high probability in the PMAT are to be located in the same diagonal structure.
has super-classes
ASSET analysis substep c

compute ASSET mask matrixc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ComputeASSETMaskMatrix

ASSET analysis substep that computes the mask matrix (MMAT). The parameters are the threshold values that are used to determine if the entries in the probability matrix (PMAT) and joint probability matrix (JMAT) are significant. Entry significance in either matrix is defined as a probability value greater than the provided threshold value. Significant entries in both probability and joint probability matrices are defined as 1 in the mask matrix.
has super-classes
ASSET analysis substep c

compute ASSET probability matrix (analytical method)c back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ComputeASSETProbabilityMatrixAnalytical

ASSET analysis probability matrix substep that computes the probability matrix (PMAT) using the assumption that the input spike trains are independent and Poisson. The computation can take as input the firing rate profiles of the spike trains used for the intersection matrix (IMAT), or those will be automatically computed using convolution with a boxcar kernel of specified width. The probability distribution of the value in the intersection matrix (IMAT) is approximated by a Poisson distribution computed using LeCam's approximation. The output is a matrix with the cumulative probabilities representing the probability of having each overlap in the IMAT strictly lower than the observed overlap, under the null hypothesis of independence of the input spike trains.
has super-classes
ASSET analysis probability matrix substep c
is disjoint with
compute ASSET probability matrix (Monte Carlo method) c

compute ASSET probability matrix (Monte Carlo method)c back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ComputeASSETProbabilityMatrixMonteCarlo

ASSET analysis probability matrix substep that computes the probability matrix (PMAT) employing a Monte Carlo approach using surrogate data obtained from the input spike trains. Different than the analytical method of computation, the null hypothesis in this method does not incorporate the assumptions that the spike trains are Poisson. Spike train surrogates can be generated using distinct methods.
has super-classes
ASSET analysis probability matrix substep c
is disjoint with
compute ASSET probability matrix (analytical method) c

compute ASSET sequence of synchronous events extractionc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ComputeASSETSequenceSynchronousEventsExtraction

Last substep of the ASSET analysis method. Given the cluster matrix (CMAT), the identity of the neurons present in each bin of the repeated sequence in the identified diagonal structures is extracted (considering the input spike trains for the computation of the intersection matrix). The output of this substep is the final description of the ASSET pattern for each diagonal structure in the CMAT.
has super-classes
ASSET analysis substep c

compute autocorrelation functionc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ComputeAutocorrelationFunction

An autocorrelation analysis that computes the estimator for the autocorrelation function, i.e. the autocorrelation values of a time series input for a number of lags. The autocorrelation function shows temporal dependencies and repetitive patterns within the input data. The value of the autocorrelation at a specific lag shows how similar the values in the time series input are when separated by a number of time units equal to that lag. The autocorrelation value of 0 indicates no correlation. The autocorrelation varies between 1 and -1 (positive and negative correlation, respectively).
has super-classes
autocorrelation analysis c

compute canonical coherencec back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ComputeCanonicalCoherence

A coherence analysis that computes the canonical coherence (caCOH) according to Vidaurre et al. (2019). The computation maximizes the coherence between two inputs (e.g., distinct datasets with electroencephalogram, electromyogram or local field potential recordings). The absolute value of the coherence between the two multivariate spaces of the inputs in the frequency domain is maximized. The caCOH aims to maximize the strength of the synchronization of oscillatory signals when two multichannel datasets are present (e.g., multiple subjects). The method then finds two spatial projections maximizing the strength of synchronization.
has super-classes
coherence analysis c

compute coefficient of variationc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ComputeCV

A dispersion statistical analysis that computes the coefficient of variation (CV). The CV is the ratio of the standard deviation to the mean. It is useful to compare different inputs, as the measure is unitless and indicates the relative variability in the input data.
has super-classes
dispersion statistical analysis c

compute coefficient of variation of the interspike intervalsc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ComputeCVInterspikeIntervals

An interspike interval variability analysis that computes the coefficient of variation (CV) of the interspike intervals (ISIs). The CV is computed as the ratio of the standard deviation of the ISIs to their mean.
has super-classes
interspike interval variability analysis c

compute coherencec back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ComputeCoherence

A coherence analysis that computes the coherence value between two inputs. It is obtained from the magnitude of the complex-valued cross power spectral density obtained for the two inputs normalized by their power spectral density (i.e., auto spectral density). Several frequency decomposition approaches can be used to obtain the cross and auto power spectral densities from the inputs. The computation can return the magnitude coherence (i.e., by taking the absolute value of the cross power spectral density and normalizing by the square root of the product of the two auto spectral densities) or the magnitude squared coherence (i.e., by computing the squared magnitude of the cross power spectral density and normalizing by the product of the two auto spectral densities).
has super-classes
coherence analysis c
has sub-classes
compute coherence (Carter method) c, compute coherence (Rosenberg method) c, compute coherence (Welch method) c, compute coherence (multitaper method) c

compute coherence (Carter method)c back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ComputeCoherenceCarter

A coherence analysis that computes the coherence from two inputs according to Carter (1987). The computation produces the magnitude squared coherence.
has super-classes
compute coherence c
is disjoint with
compute coherence (multitaper method) c, compute coherence (Rosenberg method) c, compute coherence (Welch method) c

compute coherence (multitaper method)c back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ComputeCoherenceMultitaper

A coherence analysis where the coherence value between the two inputs is computed using cross and auto power spectral densities obtained using a multitaper approach according to Thomson (1982).
has super-classes
compute coherence c
is disjoint with
compute coherence (Carter method) c, compute coherence (Rosenberg method) c, compute coherence (Welch method) c

compute coherence (Rosenberg method)c back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ComputeCoherenceRosenberg

A coherence analysis that computes the coherence from two inputs according to Rosenberg et al. (1989). The method is described for point processes (i.e., spike trains). The computation produces the magnitude squared coherence.
has super-classes
compute coherence c
is disjoint with
compute coherence (Carter method) c, compute coherence (multitaper method) c, compute coherence (Welch method) c

compute coherence (Welch method)c back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ComputeCoherenceWelch

A coherence analysis where the coherence value between the two inputs is obtained from cross and auto power spectral densities obtained using the method described by Welch (1967). For the computation, the inputs are divided into multiple overlapping segments, and the overall cross and power spectral densities for computing the coherence are obtained from averaging the single-segment estimates.
has super-classes
compute coherence c
is disjoint with
compute coherence (Carter method) c, compute coherence (multitaper method) c, compute coherence (Rosenberg method) c

compute coherencyc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ComputeCoherency

A coherency analysis that computes the coherency for two inputs. It is computed from the complex-valued cross power spectral density obtained for the two inputs normalized by the square root of the product of their power spectral densities (i.e., auto spectral densities). For each frequency, the magnitude of the complex-valued coherency describes the strength of the association and the angle describes the phase lag between the two inputs. If the inputs contain multi-epoch data (e.g., multiple trial repetitions), the cross and auto spectral densities are averaged across epochs.
has super-classes
coherency analysis c

compute complexity distributionc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ComputeComplexityDistribution

A spike train synchrony analysis that computes the complexity distribution across a set of input spike trains that typically contain the activity of different neurons. In a neuronal population, the complexity represents the total number of neurons that were spiking within a discrete time interval. For the computation, the binarized population histogram (i.e., a spike time histogram computed across spike trains, where each bin will have the count of spike trains that had at least one spike within the bin interval) is obtained using a bin size specified as a parameter. The value at each bin is the complexity. The complexity distribution is obtained by finding the frequency of each complexity value (complexity histogram) and corresponding probability density function (PDF). The complexity PDF describes the likelihood of different complexity values occurring within the neuronal population.
has super-classes
spike train synchrony analysis c

compute confidence interval (bootstrap resampling)c back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ComputeConfidenceIntervalBootstrap

A confidence interval resampling analysis that computes the confidence interval using bootstrapping techniques to create many simulated samples (bootstrap samples). It involves repeatedly sampling, with replacement, from the input (observed) data. The total number of bootstrap samples is defined as a parameter. Bootstrapping makes minimal assumptions about the underlying distribution of the data, making it especially useful for small samples or when the data do not meet the assumptions of traditional parametric (non-resampling) methods.
has super-classes
confidence interval with resampling analysis c
is disjoint with
compute confidence interval (jackknife resampling) c

compute confidence interval (jackknife resampling)c back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ComputeConfidenceIntervalJackknife

A confidence interval resampling analysis that computes the confidence interval using jackknife techniques. It involves systematically leaving out one observation at a time from the input (sample) set and calculating the statistic of interest on each of these "leave-one-out" samples. The confidence interval is computed based on statistics obtained from those jackknife samples. Jackknife is useful for small samples or when the data do not meet the assumptions of traditional parametric (non-resampling) methods.
has super-classes
confidence interval with resampling analysis c
is disjoint with
compute confidence interval (bootstrap resampling) c

compute confidence interval (non-resampling)c back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ComputeConfidenceIntervalNonResampling

A confidence interval statistical analysis that computes the confidence interval assuming a statistical distribution of the input data, and uses measures of central tendency and dispersion obtained from the data points in the input(s) (e.g., mean and standard error of the mean when assuming a normal distribution). The computation relies on theoretical distributions and established statistical formulas.
has super-classes
confidence interval statistical analysis c
is disjoint with
confidence interval with resampling analysis c

compute continuous wavelet transformc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ComputeContinuousWaveletTransform

A wavelet transform analysis that convolves the input time series with scaled and translated versions of the mother wavelet. The scale parameter can be non-dyadic (i.e., can take values that are not powers of 2). The mother wavelet used is passed as a parameter, and several types can be used (e.g., Morlet, Mexican hat, Hermitian, Meyer, Poisson). The continuous wavelet transform (CWT) is ideal for analyzing non-stationary signals, with transient behavior, rapidly changing frequencies or slowly varying changes. It is comparable to the short-time Fourier transform (STFT).
has super-classes
wavelet transform analysis c
is disjoint with
compute Morlet wavelet transform c

compute corrected imaginary phase locking valuec back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ComputeCorrectedImaginaryPhaseLockingValue

A phase locking value analysis that computes the corrected imaginary phase locking value (ciPLV), following the implementation from Bruña & Maestú (2018). It re-formulates the original phase locking value (PLV) for computational efficiency. The computation uses the imaginary part of the PLV, to make the metric insensitive to zero lag synchronizations (that can be the result of volume conduction).
has super-classes
phase locking value analysis c

compute covariancec back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ComputeCovariance

A covariance analysis that computes the values of covariance in pairwise input data.
has super-classes
covariance analysis c
has sub-classes
compute regularized covariance c

compute cross power spectral density (Morlet wavelet method)c back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ComputeCrossPowerSpectralDensityMorletWavelet

A cross power spectral density analysis that uses the Morlet wavelet transform to obtain the frequency information of the two inputs used to compute the cross power spectral density.
has super-classes
cross power spectral density analysis c
is disjoint with
compute cross power spectral density (multitaper method) c, compute cross power spectral density (periodogram method) c, compute cross power spectral density (Welch method) c

compute cross power spectral density (multitaper method)c back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ComputeCrossPowerSpectralDensityMultitaper

A cross power spectral density (CPSD) analysis that uses a multitaper approach to compute the CPSD, according to Thomson (1982). The multitaper method uses discrete prolate spheroidal functions (DPSS, also known as Slepian sequences) as tapers applied to the inputs. For the computation, a CPSD using the periodogram method is obtained for each pair of tapered signals. The DPSS functions are orthogonal, and therefore applying multiple DPSS tapers result in independent estimates of the CPSD. The final CPSD is obtained by averaging the CPSDs across all tapers. The multitaper method reduces variance and bias in the CPSD, and has a high frequency resolution. However, it is computationally expensive. The number of tapers used is passed as parameter, or estimated from the desired resolution.
has super-classes
cross power spectral density analysis c
is disjoint with
compute cross power spectral density (Morlet wavelet method) c, compute cross power spectral density (periodogram method) c, compute cross power spectral density (Welch method) c

compute cross power spectral density (periodogram method)c back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ComputeCrossPowerSpectralDensityPeriodogram

A cross power spectral density (CPSD) analysis that uses the periodogram method. For the computation, the discrete Fourier transform of each input is obtained and their cross-power spectrum is obtained. A window function can be applied to the inputs before the Fourier transform, to reduce spectral leakage. The final CPSD is obtained by normalizing the cross-power spectrum to the unit frequency using the equivalent noise bandwidth (a factor that depends on the coefficients of the window function and the sampling rate). If no window function is used or a window that does not attenuate the signal is used (e.g., Boxcar or rectangular window) the equivalent noise bandwidth is equal to the frequency resolution.
has super-classes
cross power spectral density analysis c
is disjoint with
compute cross power spectral density (Morlet wavelet method) c, compute cross power spectral density (multitaper method) c, compute cross power spectral density (Welch method) c

compute cross power spectral density (Welch method)c back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ComputeCrossPowerSpectralDensityWelch

A cross power spectral density (CPSD) analysis that uses the method defined by Welch (1967). For the computation, the two inputs are divided into several overlapping segments (length and overlap passed as parameters, or computed for the desired frequency resolution based on the input length and sampling frequency). A window function (e.g., Hann) is applied to each segment, and the cross power spectral density using the periodogram method is computed. The final CPSD is obtained by averaging all the periodograms with the individual CPSDs.
has super-classes
cross power spectral density analysis c
is disjoint with
compute cross power spectral density (Morlet wavelet method) c, compute cross power spectral density (multitaper method) c, compute cross power spectral density (periodogram method) c

compute cross-correlation functionc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ComputeCrossCorrelationFunction

A cross-correlation analysis that computes an estimate of the cross-correlation function, i.e. the cross-correlation values of two time series inputs for a number of lags. The cross-correlation function shows temporal dependencies of the first input series with respect to the second. The value of the cross-correlation at a specific lag shows how similar the values in the first input series are to values in the second input at time points separated by a number of time units equal to that lag. The cross-correlation value of 0 indicates no correlation. The cross-correlation varies between 1 and -1 (positive and negative correlation, respectively).
has super-classes
cross-correlation analysis c
has sub-classes
compute cross-correlation function (biased) c, compute cross-correlation function (unbiased) c

compute cross-correlation function (biased)c back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ComputeCrossCorrelationFunctionBiased

An analysis step that computes the biased estimator for the cross-correlation function. The biased estimator produces cross-correlation values that deviate from the true cross-correlation.
has super-classes
compute cross-correlation function c
is disjoint with
compute cross-correlation function (unbiased) c

compute cross-correlation function (unbiased)c back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ComputeCrossCorrelationFunctionUnbiased

An analysis step that computes the unbiased estimator for the cross-correlation function, using the formula in Stoica & Moses (2005). The unbiased estimation uses a correction for the bias due to zero-padding in the computation, applied to the normalization coefficient. Therefore, the resultant cross-correlation values are closer to the true cross-correlation.
has super-classes
compute cross-correlation function c
is disjoint with
compute cross-correlation function (biased) c

compute cross-spectrogram (short-time Fourier transform method)c back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ComputeCrossSpectrogramShortTimeFourierTransform

A spectrogram analysis that computes a cross-spectrogram using the short-time Fourier transform (STFT). The cross-spectrogram is the time-resolved description of the power of a pair of distinct inputs across the different frequency components. This can be used to investigate how common activity between the two inputs is distributed across the frequency components, and how it varies over time.
has super-classes
spectrogram analysis c
is disjoint with
compute spectrogram (Morlet wavelet method) c, compute spectrogram (multitaper method) c, compute spectrogram (short-time Fourier transform method) c

compute CuBIC analysisc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ComputeCubicAnalysis

A spike train synchrony analysis that uses the Cumulant Based Inference of higher-order Correlation (CuBIC) test described in Staude et al. (2010). CuBIC is a statistical method to detect the presence of higher order correlations in parallel spike trains from a neuronal population (i.e., correlations among three or more neurons). It is based on the analysis of the cumulants of the population spike count. The test takes a population histogram as input data (i.e., a spike train time histogram computed across spike trains with the activity of distinct neurons). A null hypothesis that the third cumulant of the data is less than or equal to the maximized third cumulant for a correlation order is iteratively tested (with increasing orders of correlation). The output is the minimum correlation order necessary to explain the value of the third cumulant calculated from the population spike count, together with the p-values of the hypothesis tests performed.
has super-classes
spike train synchrony analysis c

compute current source density (inverse method)c back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ComputeCurrentSourceDensityICSD

A current source density (CSD) analysis that uses the inverse current source density (iCSD) estimation method described by Pettersen et al. (2006). The iCSD is based on the inversion of the electrostatic forward solution and can be applied to data obtained from electrodes with multiple configurations. The method can handle cases with spatially confined cortical activity and spatially varying extracellular conductivity. Three options for CSD estimation using the iCSD exist. The CSD is assumed to have cylindrical symmetry and follows one of three possible assumptions: 1. is localized in infinitely thin discs; 2. is step-wise constant; 3. is continuous and smoothly varying (using cubic splines) in the vertical direction.
has super-classes
current source density analysis c
is disjoint with
compute current source density (kernel method) c, compute current source density (standard method) c

compute current source density (kernel method)c back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ComputeCurrentSourceDensityKCSD

A current source density (CSD) analysis that uses kernel methods to compute the CSD (kCSD), described by Potworowski et al. (2012). kCSD is non parametric and can estimate the CSD using signals recorded from arbitrarily distributed electrodes, as the assumption of regular electrode placement is not necessary. The method can handle 1D, 2D or 3D electrode configurations. The kCSD can also estimate CSD at any location, as it is not limited to the electrode positions, and uses cross-validation to ensure no overfitting.
has super-classes
current source density analysis c
is disjoint with
compute current source density (inverse method) c, compute current source density (standard method) c

compute current source density (standard method)c back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ComputeCurrentSourceDensityStandard

A current source density (CSD) analysis that uses a double spatial derivative of the recorded extracellular potentials to compute the CSD. The original method by Freeman & Nicholson (1975) assumes homogeneous cortical in-plane activity, constant extracellular electrical conductivity and equidistant electrode contacts, and can only predict the CSD at interior electrode positions. Vaknin et al. (1988) suggested a procedure to obtain the CSD for the first and last electrodes by copying the outmost recordings, therefore extending the grid beyond the electrode contacts. This is based on the assumption that the potential varies negligibly above the first and below the last electrode.
has super-classes
current source density analysis c
is disjoint with
compute current source density (inverse method) c, compute current source density (kernel method) c

compute CV2c back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ComputeCV2

An interspike interval variability analysis that computes the CV2, a measure of the intrinsic variability of a spike train that considers adjacent interspike intervals. The CV2 is more robust against fluctuations in the firing rate than the usual approach of taking the coefficient of variation of the interspike intervals of the spike train.
has super-classes
interspike interval variability analysis c

compute debiased squared weighted phase lag indexc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ComputeDebiasedSquaredWeightedPhaseLagIndex

A phase lag index (PLI) analysis that computes the debiased squared weighted PLI (WPLI) following Vinck et al. (2011). The direct PLI estimator is positively biased, especially when the sample sizes (i.e., number of trials) are small. The debiased squared WPLI is computed by computing the imaginary components of the cross spectral densities, computing the average imaginary component of the cross spectral densities, and normalizing by the computed average over the magnitudes of the imaginary component of the cross spectral densities.
has super-classes
phase lag index analysis c

compute directed phase lag indexc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ComputeDirectedPhaseLagIndex

A phase lag index (PLI) analysis that computes the directed PLI (DPLI) according to Stam & van Straaten (2012). The DPLI uses the Heaviside step function on the imaginary part of the cross power spectral density, and provides the ability to discriminate whether the first time series is leading or lagging the second. The DPLI ranges between 0 and 1. A DPLI value of 0.5 means that the first time series leads and lags the second time series equally often. A DPLI value greater than 0.5 means that the first time series leads the second more often than it lags. A value of 1 means that the first time series always leads. On the contrary, a DPLI value smaller than 0.5 means that the first time series lags the second more often than it leads. A DPLI value of zero means that the first time series always lags. The PLI can be computed from the DPLI.
has super-classes
phase lag index analysis c

compute directed transfer functionc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ComputeDirectedTransferFunction

A functional connectivity analysis that computes the directed transfer function (DTF) according to Kaminski & Blinowska (1991). DTF can estimate the direction and frequency content of the brain activity flow. The DTF measure is obtained from the spectral transfer matrix computed from multivariate time series input data. For the DTF computation, the spectral transfer matrix is obtained from a multivariate autoregressive model. The DTF estimate is obtained by using a normalization factor computed by the sum along the rows of the spectral transfer matrix. The DTF can have values in the range from 0 to 1.
has super-classes
functional connectivity analysis c

compute event-related potentialc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ComputeEventRelatedPotential

An analysis step that computes the event-related potential (ERP). The ERP is the neural activity (e.g., local field potential or electroencephalogram voltages) around a presented stimulus or spontaneous behavioral event. Usually, the event is presented/occurs repeatedly across multiple trials, obtaining multiple event-related potential waveforms that can be averaged to cancel the noise.
has super-classes
analysis step c
has sub-classes
compute evoked potential c

compute event-triggered averagec back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ComputeEventTriggeredAverage

A triggered average analysis that uses an event of interest such as an external stimulus (e.g., electrical, visual, auditory) or a spontaneous behavior (e.g., eye blink) as a trigger to average a signal. The output of the method will provide the average value of the signal around the time where each event occurred (event-triggered average).
has super-classes
triggered average analysis c

compute evoked potentialc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ComputedEvokedPotential

An analysis step that computes the evoked potential (EP). The EP is the neural activity (e.g., local field potential or electroencephalogram voltages) around a presented stimulus (e.g., auditory, visual, electrical). Usually, the stimulus is presented repeatedly across multiple trials, obtaining multiple evoked potential waveforms that can be averaged to cancel the noise. It is an event-related potential (ERP) obtained from presenting a stimulus rather than spontaneous behavioral events.
has super-classes
compute event-related potential c

compute frequency domain conditional Granger causalityc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ComputeFrequencyDomainConditionalGrangerCausality

A conditional Granger causality (GC) analysis that computes the GC measures in the frequency domain.
has super-classes
conditional Granger causality analysis c
is disjoint with
compute time domain conditional Granger causality c

compute frequency domain pairwise Granger causality (Brovelli method)c back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ComputeFrequencyDomainPairwiseGrangerCausalityBrovelli

A frequency domain pairwise Granger causality (GC) analysis that uses the parametric approach according to Brovelli et al. (2004). It uses an MVAR (multivariate autoregressive model) to obtain the coefficients used for the computation of the spectral transfer matrix needed for GC estimation according to the frequency domain GC formulation by Geweke (1982).
has super-classes
frequency domain pairwise granger causality analysis c
is disjoint with
compute frequency domain pairwise Granger causality (Dhamala method) c, compute frequency domain pairwise Granger causality (Hafner method) c, compute frequency domain pairwise Granger causality (Wen method) c

compute frequency domain pairwise Granger causality (Dhamala method)c back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ComputeFrequencyDomainPairwiseGrangerCausalityDhamala

A frequency domain pairwise Granger causality (GC) analysis that uses the non-parametric approach according to Dhamala et al. (2008). It is based on Fourier and wavelet transforms to obtain the spectral density matrix and the algorithm from Wilson (1972) for its factorization.
has super-classes
frequency domain pairwise granger causality analysis c
is disjoint with
compute frequency domain pairwise Granger causality (Brovelli method) c, compute frequency domain pairwise Granger causality (Hafner method) c, compute frequency domain pairwise Granger causality (Wen method) c

compute frequency domain pairwise Granger causality (Hafner method)c back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ComputeFrequencyDomainPairwiseGrangerCausalityHafner

A frequency domain pairwise Granger causality (GC) analysis that uses the parametric approach according to Hafner & Herwartz (2008). It uses a multivariate GARCH (generalized autoregressive conditional heteroskedasticity) model and constructs a Wald test on noncausality in variance. This is an alternative to methods based on the residuals of estimated univariate models. The Wald test has superior power properties.
has super-classes
frequency domain pairwise granger causality analysis c
is disjoint with
compute frequency domain pairwise Granger causality (Brovelli method) c, compute frequency domain pairwise Granger causality (Dhamala method) c, compute frequency domain pairwise Granger causality (Wen method) c

compute frequency domain pairwise Granger causality (Wen method)c back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ComputeFrequencyDomainPairwiseGrangerCausalityWen

A frequency domain pairwise Granger causality (GC) analysis that uses the non-parametric approach according to Wen et al. (2013). It is a multivariate framework for estimating GC based on spectral density matrix factorization. The approach requires only a single estimation of the spectral density matrix for the entire dataset (e.g., multiple time series inputs). GC for the subsets (i.e., pairs of inputs) can then be calculated by factorizing the relevant submatrix of this overall spectral density matrix.
has super-classes
frequency domain pairwise granger causality analysis c
is disjoint with
compute frequency domain pairwise Granger causality (Brovelli method) c, compute frequency domain pairwise Granger causality (Dhamala method) c, compute frequency domain pairwise Granger causality (Hafner method) c

compute imaginary coherencyc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ComputeImaginaryCoherency

A coherency analysis that computes the imaginary part of the coherency for two inputs according to Nolte et al. (2004). For the computation, the imaginary part of the complex-valued cross power spectral density obtained for the two inputs is normalized by the square root of the product of their power spectral densities (i.e., auto spectral densities). If the inputs contain multi-epoch data (e.g., multiple trial repetitions), the cross and auto spectral densities are averaged across epochs. The imaginary part of the coherency is less affected by volume conduction than the (complex) coherency.
has super-classes
coherency analysis c

compute instantaneous firing rate (interspike interval method)c back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ComputeInstantaneousFiringRateInterspikeInterval

An instantaneous firing rate analysis that computes the instantaneous firing rate by using the reciprocal of the interspike intervals.
has super-classes
instantaneous firing rate analysis c
is disjoint with
compute instantaneous firing rate (kernel density estimation method) c, compute instantaneous firing rate (local regression method) c

compute instantaneous firing rate (kernel density estimation method)c back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ComputeInstantaneousFiringRateKernelDensityEstimation

An instantaneous firing rate analysis that computes the instantaneous firing rate by convolution of spike times with a kernel function. The output of the computation is a weighted average of the spikes around the kernel.
has super-classes
instantaneous firing rate analysis c
is disjoint with
compute instantaneous firing rate (interspike interval method) c, compute instantaneous firing rate (local regression method) c

compute instantaneous firing rate (local regression method)c back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ComputeInstantaneousFiringRateLocalRegression

An instantaneous firing rate analysis that computes the instantaneous firing rate using local regression methods. The estimation procedure approximates the log of the firing rate using a low-order polynomial within a moving window (local neighborhood).
has super-classes
instantaneous firing rate analysis c
is disjoint with
compute instantaneous firing rate (interspike interval method) c, compute instantaneous firing rate (kernel density estimation method) c

compute interquartile rangec back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ComputeInterquartileRange

A dispersion statistical analysis that computes the interquartile range (IQR). The IQR is the difference between the 75th and 25th percentiles (i.e., the range within which the central 50% of the data points lie).
has super-classes
dispersion statistical analysis c

compute interspike interval histogramc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ComputeInterspikeIntervalHistogram

An interspike interval analysis that computes the histogram of interspike intervals. For the computation, a time interval with fixed duration starting from zero is discretized into smaller intervals (bins). The count of input interspike intervals whose values fall into each bin is obtained. Therefore, the output contains a representation of the distribution of the interspike intervals in the input.
has super-classes
interspike interval analysis c

compute interspike intervalsc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ComputeInterspikeIntervals

An interspike interval analysis that computes the intervals between successive spikes in a spike train (interspike intervals; ISIs).
has super-classes
interspike interval analysis c

compute ISI-distancec back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ComputeISIDistance

A time-scale independent spike train distance analysis that computes the ISI-distance, described in Kreuz et al. (2007). For the computation, the discrete sequence of spike times is transformed into a continuous temporal profile with one value per sample point. The values at each time point are derived from the interspike intervals. The distance is then obtained as the temporal average of the time profile. ISI-distance is well-designed to describe similarities in the firing rate profile of the input spike trains, but it is not optimal to capture neuronal synchrony.
has super-classes
time-scale independent spike train distance analysis c

compute joint peristimulus time histogramc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ComputeJointPeristimulusTimeHistogram

A spike train synchrony analysis that computes the joint peristimulus time histogram (JPSTH) from trial-by-trial spike train inputs obtained from two different neurons, after the repeated presentation of a stimulus. The JPSTH provides a representation of the timing relationship between the firing of the two neurons in response to the stimulus. It combines the peristimulus time histograms (PSTHs) of each neuron to illustrate how their firing rates co-vary over time relative to the stimulus event. This helps in understanding the temporal correlation between the neurons. The computation can produce three outputs: * the JPSTH matrix (i.e., a matrix whose bins contain the counts of coincidences in the firing of the two neurons); * the peristimulus coincidence histogram (i.e., a histogram obtained from a cross-section along the main diagonal of the JPSTH matrix); * the cross-correlation histogram computed by summing the bins along the main and each paradiagonal of the JPSTH matrix (after normalizing by the bin length, as the paradiagonals in the JPSTH matrix are of different lengths).
has super-classes
spike train synchrony analysis c

compute local variationc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ComputeLV

An interspike interval variability analysis that computes the local variation (LV) of the interspike intervals. LV reflects the stepwise variability of a sequence of spikes, and is able to extract the spiking characteristics of individual neurons even in the presence of external modulations of the firing rate.
has super-classes
interspike interval variability analysis c

compute maximized imaginary coherencyc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ComputeMaximizedImaginaryCoherency

A coherency analysis that computes the maximized imaginary coherency (MIC) according to Ewald et al. (2012). The computation uses an eigenvalue-based optimization to find weight vectors that maximize the imaginary part of coherency computed between virtual channels derived from the input data. The weights are optimized for each frequency component. After the weights are obtained, the final MIC measure is obtained for each frequency.
has super-classes
coherency analysis c

compute meanc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ComputeMean

A central tendency statistical analysis that computes the mean of the input data. The mean is the arithmetic average of all data points.
has super-classes
central tendency statistical analysis c

compute mean firing ratec back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ComputeMeanFiringRate

A firing rate analysis that computes the mean firing rate, defined as the number of spikes in a time interval divided by the duration of the interval. The mean firing rate is the temporal average of the neuronal activity over that interval.
has super-classes
firing rate analysis c

compute mean phase vectorc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ComputeMeanPhaseVector

A phase analysis that computes the mean among two or more input phases. For the computation, the input phases are represented as vectors in the unit circle, and the mean phase vector is computed. The analysis can return the mean phase vector (i.e., angle and length), the vector angle, or the vector length.
has super-classes
phase analysis c

compute mean vector length (Canolty method)c back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ComputeMeanVectorLengthCanolty

A mean vector length (MVL) analysis that computes the mean vector length as described in Canolty et al. (2006). For the computation, phase is extracted from the low-frequency analytic signal, and amplitude is extracted from the high-frequency analytic signal. The phase angle and magnitude is used to define a complex-valued time series, and each complex value is a vector in the polar plane. Averaging all vectors yields a mean vector whose length indicates the coupling strength and whose direction indicates the phase where amplitude is strongest. Without coupling, the vectors cancel out, resulting in a short mean vector without meaningful phase direction. If phase-amplitude coupling exists, the magnitude of a subset of vectors is especially high at a specific phase or narrow phase range.
has super-classes
mean vector length analysis c
is disjoint with
compute mean vector length (Özkurt method) c

compute mean vector length (Özkurt method)c back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ComputeMeanVectorLengthOzkurt

A mean vector length (MVL) analysis that computes the MVL as described in Özkurt & Schnitzler (2011). The original MVL (Canolty et al., 2006) may be affected by factors in the input data (e.g., amplitude outliers or non-uniform distribution of phase angles). This computation estimates a direct MVL that is amplitude-normalized to obtain values in the 0 to 1 range, and that takes care of possible amplitude differences in the raw data.
has super-classes
mean vector length analysis c
is disjoint with
compute mean vector length (Canolty method) c

compute medianc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ComputeMedian

A central tendency statistical analysis that computes the median of the input data. The median is the middle value when data points are arranged in ascending order (i.e., it divides the data points in two equal halves, with 50% of the data points below it and 50% above it). If there is an even number of data points, the median is the average of the two middle values.
has super-classes
central tendency statistical analysis c

compute modulation indexc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ComputeModulationIndex

A phase-amplitude coupling (PAC) analysis that computes the modulation index (MI) as described in Tort et al. (2010). For the computation, the Hilbert transform is used to obtain the instantaneous phase from the input time series with the low-frequency oscillation, and the instantaneous amplitude from the input time series with the high-frequency oscillation. The phase of the low-frequency oscillation is discretized into bins and the amplitude of the high-frequency oscillation is averaged within each bin to create a distribution. This distribution is then compared to a uniform distribution using the Kullback-Leibler divergence, normalized by the maximum possible divergence, resulting in the MI.
has super-classes
phase-amplitude coupling analysis c

compute Morlet wavelet transformc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ComputeMorletWaveletTransform

A wavelet transform analysis based on the complex-valued Morlet wavelet. The transform can be performed either in the time domain (by convolution) or in the frequency domain (by multiplication).
has super-classes
wavelet transform analysis c
has sub-classes
compute Morlet wavelet transform (Le Van Quyen method) c, compute Morlet wavelet transform (Tallon-Baudry method) c
is disjoint with
compute continuous wavelet transform c

compute Morlet wavelet transform (Le Van Quyen method)c back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ComputeMorletWaveletTransformLeVanQuyen

A wavelet transform analysis using the Morlet wavelet where the parametrization of the mother wavelet is done according to Le Van Quyen et al. (2001). The size of the mother wavelet is determined in number of cycles to control the frequency and temporal resolutions (approximate number of oscillation cycles within a wavelet).
has super-classes
compute Morlet wavelet transform c
is disjoint with
compute Morlet wavelet transform (Tallon-Baudry method) c

compute Morlet wavelet transform (Tallon-Baudry method)c back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ComputeMorletWaveletTransformTallonBaudry

A wavelet transform analysis using the Morlet wavelet based on the methods described in Tallon-Baudry et al. (1997). The ratio of the central frequency to the spectral bandwidth is 7, with central frequencies ranging from 20 to 100 Hz in 1 Hz steps. This resulted in varying time/frequency resolution across the spectrum: time resolution increases with frequency, while frequency resolution decreases.
has super-classes
compute Morlet wavelet transform c
is disjoint with
compute Morlet wavelet transform (Le Van Quyen method) c

compute multivariate interaction measurec back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ComputeMultivariateInteractionMeasure

A functional connectivity analysis that computes the multivariate interaction measure (MIM) as defined by Ewald et al. (2012). MIM is constructed from the maximization of imaginary coherency.
has super-classes
functional connectivity analysis c

compute mutual informationc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ComputeMutualInformation

A functional connectivity analysis that computes a mutual information (MI) measure. MI is based on Shannon information theory, and quantifies the amount of information from one input that is obtained from another input. Therefore, it can be used to determine how the neuronal activity provides information about a variable (e.g., behavioral stimulus) or how the information flows between different brain regions or neurons. The MI is measured in bits.
has super-classes
functional connectivity analysis c

compute neuronal population vectorc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ComputeNeuronalPopulationVector

An analysis step that computes the neuronal population vector, used to describe the collective activity of a group of neurons. The neuronal population vector is obtained taking as inputs the multiple responses of a neuronal population in the context of distinct values of a behavioral measure (e.g., tuning curves showing the response of each individual neuron to different arm movement directions in the 2-D space). The analysis step obtains a weighted vectorial sum of the neural activities, which will result in an estimate of the behavioral measure considering the collective activity of the population (e.g., the movement direction given the neuronal activity).
has super-classes
analysis step c

compute noise correlationsc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ComputeNoiseCorrelations

A spike train correlation analysis that computes the noise correlations (NC) between two input spike trains. The NC is the Pearson's correlation coefficient of spike count responses to repeated presentations of identical stimuli, under the same behavioral conditions. The spike counts are typically measured over the time scale of a stimulus presentation or a behavioral trial, which range from a few hundred milliseconds to several seconds. NC assesses whether neurons exhibit trial-by-trial fluctuations in firing rates that are not influenced by varying sensory or behavioral conditions.
has super-classes
spike train correlation analysis c

compute noise covariancec back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ComputeNoiseCovariance

A covariance analysis that computes the noise covariance, i.e., how much two noise signals vary together. The noise data inputs can be non-subject recordings (e.g., recordings from the empty experimental room) or are obtained from periods without stimulation or meaningful experimental manipulations (e.g., prestimulus intervals). These reflect random variations or disturbances that are not part of the actual signal or data of interest.
has super-classes
covariance analysis c

compute optimal bin sizec back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ComputeOptimalBinSize

An analysis step that finds the optimal bin size considering the input data when discretizing data into smaller intervals (bins). The computation uses the formula from Scott (1979).
has super-classes
analysis step c

compute optimal kernel bandwidthc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ComputeOptimalKernelBandwidth

An analysis step that computes the optimal fixed bandwidth (width) for a Gaussian kernel used for the estimation of the firing rate using kernel density estimation. The analysis step uses the input spike train for which the firing rate will be computed, and follows the implementation by Shimazaki & Shinomoto (2010).
has super-classes
analysis step c

compute orthogonalized power envelope correlationc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ComputeOrthogonalizedPowerEnvelopeCorrelation

A functional connectivity analysis that computes the orthogonalized power envelope correlation, according to Hipp et al. (2012). This method relies on correlations between the instantaneous amplitudes of cross-region input signals (power envelopes). The instantaneous amplitudes of the two input time series are orthogonalized aiming to remove spurious correlations of signal power (e.g., due to limited spatial resolution of electrophysiological measures).
has super-classes
correlation analysis c

compute pairwise phase consistencyc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ComputePairwisePhaseConsistency

A phase analysis that computes the pairwise phase consistency (PPC) measure according to Vinck et al. (2010). The PPC quantifies the distribution of phase differences across the inputs, but is less biased by the number of observations in comparison to the phase locking value (PLV). For the computation, the phase difference (angular distance) is obtained for all pairs of observations in the input (that can be represented as vectors in the unit circle, where the angle is the relative phase). The cosine of the angular distance (an estimate of the dot product of the vectors corresponding to each element in a pair) is computed for every pair, and the PPC estimate is obtained from the average of all pairwise dot products. With phase synchronization, the distribution of the pairwise dot products is centered around an average value, while without synchronization it will be distributed across the unit circle. The PPC provides an unbiased estimate of the squared PLV.
has super-classes
phase analysis c

compute partial coherencec back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ComputePartialCoherence

A coherence analysis that computes the partial coherence, i.e., the coherence value between a pair of inputs (e.g., time series with recordings from two distinct electrode channels) after removing the influence of one or more additional inputs (e.g., the time series with the recordings from all remaining channels). The partial coherence is computed according to Rosenberg et al. (1998). The partial coherence reflects the linear association in the frequency domain (for each frequency component) between the pair of inputs of interest, removing spurious coherence caused by confounding factors such as shared inputs to the pair of interest or volume conduction.
has super-classes
coherence analysis c

compute partial directed coherencec back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ComputePartialDirectedCoherence

A coherence analysis that computes the partial directed coherence (PDC) according to Baccalá & Sameshima (2001). The PDC describes the relationships between multivariate time series inputs (direction of information flow). To compute the PDC, the multivariate partial coherences obtained from multivariate autoregressive models are decomposed. The PDC reflects a frequency-domain representation of the concept associated with Granger causality.
has super-classes
spectral analysis c

compute peristimulus time histogram (adaptive kernel smoothing)c back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ComputePeristimulusTimeHistogramAdaptiveKernelSmoothing

A peristimulus time histogram (PSTH) analysis that computes the PSTH using kernel density smoothing with an adaptive kernel width. The adaptive kernel is obtained by first computing the PSTH with a fixed-width kernel, and then modifying the kernel width in order to have a constant but time-dependent average number of spikes under the kernel (i.e., segments of the data with a high density of spikes will have a reduced kernel width).
has super-classes
peristimulus time histogram analysis c
is disjoint with
compute peristimulus time histogram (fixed kernel smoothing) c, compute peristimulus time histogram (optimal bin size) c, compute peristimulus time histogram (user-selected bin size) c

compute peristimulus time histogram (fixed kernel smoothing)c back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ComputePeristimulusTimeHistogramFixedKernelSmoothing

A peristimulus time histogram (PSTH) analysis that computes the PSTH using kernel density smoothing with a fixed kernel width specified as parameter.
has super-classes
peristimulus time histogram analysis c
is disjoint with
compute peristimulus time histogram (adaptive kernel smoothing) c, compute peristimulus time histogram (optimal bin size) c, compute peristimulus time histogram (user-selected bin size) c

compute peristimulus time histogram (optimal bin size)c back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ComputePeristimulusTimeHistogramOptimalBinSize

A peristimulus time histogram (PSTH) analysis that computes the PSTH finding the optimal bin size from the input spike train data, using the formula from Scott (1979).
has super-classes
peristimulus time histogram analysis c
is disjoint with
compute peristimulus time histogram (adaptive kernel smoothing) c, compute peristimulus time histogram (fixed kernel smoothing) c, compute peristimulus time histogram (user-selected bin size) c

compute peristimulus time histogram (user-selected bin size)c back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ComputePeristimulusTimeHistogramUserSelectedBinSize

A peristimulus time histogram (PSTH) analysis that computes the PSTH using a fixed bin size specified as a parameter.
has super-classes
peristimulus time histogram analysis c
is disjoint with
compute peristimulus time histogram (adaptive kernel smoothing) c, compute peristimulus time histogram (fixed kernel smoothing) c, compute peristimulus time histogram (optimal bin size) c

compute phase differencec back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ComputePhaseDifference

A phase analysis that computes the difference between two input phases.
has super-classes
phase analysis c

compute phase lag indexc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ComputePhaseLagIndex

A phase lag index (PLI) analysis that computes the PLI following Stam et al. (2007). The input data contains multiple repetitions of a pair of signals (e.g., time series with recordings from a pair of electrodes across multiple trials). For each repetition, the sign of the phase differences between the two time series is obtained from the imaginary part of the cross power spectral density (CPSD). The PLI value is the absolute value of the average of the signs of all repetitions. The PLI ranges between 0 and 1. A PLI of zero means that the first time series leads the second equally often (i.e., indicates either no coupling or coupling with a phase difference centered around 0 mod π, which could be from common sources such as volume conduction). A value greater than zero means an imbalance in the likelihood of the first time series to be either leading or lagging the second time series. A PLI of 1 indicates perfect phase locking, and that the first time series only leads or only lags (at a value of phase differences different from 0 mod π).
has super-classes
phase lag index analysis c

compute phase locking valuec back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ComputePhaseLockingValue

A phase locking value (PLV) analysis that computes the PLV value as originally described by Lachaux et al. (1999). The input data is a set of pairs of time series (e.g., the trial-by-trial local field potential recorded from two different electrodes). For each time series pair, the instantaneous phase is obtained (e.g., using the Hilbert transform or wavelet decomposition), and the phase difference for each time point is obtained. The PLV value is computed by averaging the complex phase differences across all pairs, obtaining one PLV value per time point. The PLV ranges from 0 to 1. A PLV of 1 indicates perfect phase locking, meaning the phase difference between the two time series is constant over time. A PLV of 0 indicates no phase locking, meaning the phase difference is randomly distributed over time.
has super-classes
phase locking value analysis c

compute phase slope indexc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ComputePhaseSlopeIndex

A functional connectivity analysis that computes the phase slope index (PSI) according to Nolte et al. (2008). The PSI is based on the slope of the phase of the cross-spectral density between two time series inputs, considering how the phase difference between two signals changes as you move from one frequency bin to the next. It is computed from the complex-valued coherency using a bandwidth specified as parameter. For the computation, the change in phase difference between neighboring frequency bins is obtained (considering the specified bandwidth) and weighted. The PSI value deviates from zero when the phase difference changes consistently across frequencies and there is substantial coherence. The sign of the PSI indicates the temporal order of the two signals (i.e., which signal is leading the other one).
has super-classes
phase analysis c

compute population histogramc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ComputePopulationHistogram

A spike train time histogram analysis that computes a histogram across two or more spike trains that contain the activity of different neurons (i.e., a neuronal population), recorded at fully-overlapping time intervals. The activity in each histogram bin reflects the combined activity of the population at that time, and the distribution of the histogram corresponds to the population activity over time.
has super-classes
spike train time histogram analysis c
is disjoint with
peristimulus time histogram analysis c

compute power spectral density (Bartlett method)c back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ComputePowerSpectralDensityBartlett

A power spectral density (PSD) analysis that uses the method defined by Bartlett (1950). For the computation, the input time is divided into non-overlapping segments, with length specified as a parameter. A periodogram is computed for each segment to obtain the single-segment PSD. The final PSD is obtained by averaging all the single-segment PSDs. A window function can be applied to each segment before computing the periodograms. The PSD obtained with the Bartlett method is less noisy than a single periodogram obtained from the entire signal, although the frequency resolution of the estimates is reduced due to segmenting. It is equivalent to the Welch method without any segment overlap.
has super-classes
power spectral density analysis c
is disjoint with
compute power spectral density (multitaper method) c, compute power spectral density (periodogram method) c, compute power spectral density (Welch method) c

compute power spectral density (multitaper method)c back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ComputePowerSpectralDensityMultitaper

A power spectral density (PSD) analysis that uses a multitaper approach to compute the PSD, according to Thomson (1982). The multitaper method uses discrete prolate spheroidal functions (DPSS, also known as Slepian sequences) as tapers applied to the input signal. For the computation, a PSD using the periodogram method is obtained for each tapered signal. The DPSS functions are orthogonal, and therefore applying multiple DPSS tapers result in independent estimates of the PSD. The final PSD is obtained by averaging the periodograms across all tapers. The multitaper method reduces variance and bias in the PSD, and has a high frequency resolution. However, it is computationally expensive. The number of tapers used is passed as parameter, or estimated from the desired resolution.
has super-classes
power spectral density analysis c
is disjoint with
compute power spectral density (Bartlett method) c, compute power spectral density (periodogram method) c, compute power spectral density (Welch method) c

compute power spectral density (periodogram method)c back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ComputePowerSpectralDensityPeriodogram

A power spectral density (PSD) analysis that uses the periodogram method. For the computation, the discrete Fourier transform is applied to the full length of the input, and the power spectrum is obtained. To reduce spectral leakage, a window function can be applied to the input before the computation of the Fourier transform (this is referred as the modified periodogram). The power spectrum is then normalized to the unit frequency, using the equivalent noise bandwidth (a factor that depends on the coefficients of the window function and the sampling rate). If no window function is used or a window that does not attenuate the signal is used (e.g., Boxcar or rectangular window) the equivalent noise bandwidth is equal to the frequency resolution. The PSD using the periodogram is computationally simple to obtain.
has super-classes
power spectral density analysis c
is disjoint with
compute power spectral density (Bartlett method) c, compute power spectral density (multitaper method) c, compute power spectral density (Welch method) c

compute power spectral density (Welch method)c back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ComputePowerSpectralDensityWelch

A power spectral density (PSD) analysis that uses the method defined by Welch (1967). For the computation, the input is divided into several overlapping segments (length and overlap passed as parameters, or computed for the desired frequency resolution based on the input length and sampling frequency). A window function (e.g., Hann) is applied to each segment, and a periodogram is computed to obtain the PSD for the segment. The final PSD is obtained by averaging all the periodograms with the single-segment PSDs. If there is no overlap between segments, this is equivalent to the Bartlett method.
has super-classes
power spectral density analysis c
is disjoint with
compute power spectral density (Bartlett method) c, compute power spectral density (multitaper method) c, compute power spectral density (periodogram method) c

compute rate change detection multiple filter testc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ComputeRateChangeDetectionMultipleFilterTest

An analysis step that uses the change point detection algorithm from Messer et al. (2014) to determine if a input spike train has constant firing rate (stationary) or has one or more points in which the firing rate decreases or increases (change point). In the latter case, the spike train is considered non-stationary. The analysis step outputs one or more change points in the case of non-stationarity.
has super-classes
analysis step c

compute rectified area under the curvec back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ComputeRectifiedAreaUnderCurve

An analysis step that computes the rectified area under the curve (RAUC). For the computation, the input signal is rectified (i.e., the absolute value is obtained) and the area under the curve is computed by integration using the composite trapezoidal rule.
has super-classes
analysis step c

compute regularized covariancec back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ComputeRegularizedCovariance

A covariance analysis where the computation of the covariance values incorporates regularization techniques to improve the numerical stability, especially if the number of samples is small.
has super-classes
compute covariance c

compute revised local variationc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ComputeLVR

An interspike interval variability analysis that computes the revised local variation (LvR) of interspike intervals. Compared to the original local variation (LV) measure, LvR has better invariance to fluctuations in the firing rate fluctuations. This is achieved by using a refractoriness constant in the computation of the measure.
has super-classes
interspike interval variability analysis c

compute short-time Fourier transformc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ComputeShortTimeFourierTransform

A time-frequency analysis that computes the short-time Fourier transform (STFT) of the input time series. The analysis divides the input time-domain signal into short segments with equal time and computes the Fourier transform for each segment. The output is a sequence of coefficients of complex sinusoids, each representing a frequency component in the input signal at a distinct time segment. Therefore, this provides the the time-localized frequency and phase information of the input. The segments can be windowed using a window function.
has super-classes
time-frequency analysis c

compute spectrogram (Morlet wavelet method)c back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ComputeSpectrogramMorletWavelet

A spectrogram analysis that uses the Morlet wavelet transform on the input to obtain the time-frequency information used to build the spectrogram.
has super-classes
spectrogram analysis c
is disjoint with
compute cross-spectrogram (short-time Fourier transform method) c, compute spectrogram (multitaper method) c, compute spectrogram (short-time Fourier transform method) c

compute spectrogram (multitaper method)c back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ComputeSpectrogramMultitaper

A spectrogram analysis that uses a multitaper approach to obtain the time-frequency information from the input and that is used to build the spectrogram.
has super-classes
spectrogram analysis c
is disjoint with
compute cross-spectrogram (short-time Fourier transform method) c, compute spectrogram (Morlet wavelet method) c, compute spectrogram (short-time Fourier transform method) c

compute spectrogram (short-time Fourier transform method)c back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ComputeSpectrogramShortTimeFourierTransform

A spectrogram analysis that uses the short-time Fourier transform (STFT) on the input to obtain the time-frequency information used to build the spectrogram.
has super-classes
spectrogram analysis c
is disjoint with
compute cross-spectrogram (short-time Fourier transform method) c, compute spectrogram (Morlet wavelet method) c, compute spectrogram (multitaper method) c

compute SPIKE distancec back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ComputeSPIKEDistance

A time-scale independent spike train distance analysis that computes the SPIKE-distance, described in Kreuz et al. (2012). For the computation, the discrete sequence of spike times is transformed in a continuous temporal profile with one value per sample point. The values at each time point are derived from the differences in the spike times of the two input spike trains. Compared to the ISI-distance, the SPIKE-distance is more sensitive to spike timing.
has super-classes
time-scale independent spike train distance analysis c

compute SPIKE synchronizationc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ComputeSPIKESynchronization

A time-scale independent spike train distance analysis that computes the SPIKE synchronization distance, described in Kreuz et al. (2015). The distance detects coincidences in the spiking activity and can quantify the degree of synchrony in the input spike trains. The metric quantifies the overall fraction of coincidences. It is zero-valued if and only if the input spike trains do not contain any coincidences. It has a value of 1 if and only if each spike in every input spike train has one matching spike in all the other spike trains.
has super-classes
time-scale independent spike train distance analysis c

compute spike time tiling coefficientc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ComputeSpikeTimeTilingCoefficient

A spike train correlation analysis that computes the spike time tiling coefficient (STTC) as described by Cutts and Eglen (2014). The STTC measures the pairwise correlation between two input spike trains, and has advantages over the related correlation index: it is not confounded by the firing rate, it distinguishes lack of correlation from anti-correlation, periods without neural activity don't add to the correlation, and it is sensitive to firing patterns. The computation is based on a synchronicity window parameter, that is used to define short time windows around each spike that are used in the computation (spike time tiling).
has super-classes
spike train correlation analysis c

compute spike train autocorrelation histogramc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ComputeSpikeTrainAutocorrelationHistogram

An autocorrelation analysis that computes the autocorrelation histogram for a input spike train. The histogram is obtained by sliding a time window that is discretized into smaller time intervals (bins) centered on a spike (corresponding to the lag zero). The spike count in each bin is obtained. Therefore, this binning process measures the number of spikes occurring at various time lags relative to the center spike. The histogram window is slidden over each spike in the input spike train, and the spike count in each bin is accumulated to produce the autocorrelation histogram output. The width of the bin interval is controlled by a parameter.
has super-classes
autocorrelation analysis c

compute spike train autocorrelation time scalec back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ComputeSpikeTrainAutocorrelationTimeScale

An analysis step that computes the autocorrelation time of a binned spike train input (spike train autocorrelation time scale). The computation follows the method described by Wieland et al. (2015).
has super-classes
analysis step c

compute spike train cross-correlation histogramc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ComputeSpikeTrainCrossCorrelationHistogram

A cross-correlation analysis that computes the cross-correlation histogram (CCH) for a pair of input spike trains. The CCH shows how often spikes in the reference spike train occur before or after spikes in the reference spike train, at distinct lag intervals. For the computation, the spike trains are aligned in time. The histogram is obtained by sliding a time window that is discretized into smaller time intervals (bins) centered on a spike of the first (reference) input spike train (corresponding to the lag zero). The spike count of the second (target) spike train input is then obtained in each bin. Therefore, this binning process measures the number of spikes in the target spike train occurring at various time lags relative to the spikes in the reference spike train. The histogram window is slidden over each spike in the reference spike train, and the spike count in each bin is accumulated to produce the CCH output. The width of the bin interval is controlled by a parameter.
has super-classes
cross-correlation analysis c
has sub-classes
compute spike train cross-correlation histogram (Eggermont method) c

compute spike train cross-correlation histogram (Eggermont method)c back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ComputeSpikeTrainCrossCorrelationHistogramEggermont

A computation of the cross-correlation histogram (CCH) for a pair of binned input spike trains using the method described in Eggermont (2010). The formula is valid for binned spike train inputs with at most one spike per bin, and returns the cross-correlation coefficient for the lags considered (range -1 to 1).
has super-classes
compute spike train cross-correlation histogram c

compute spike train Fano factorc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ComputeSpikeTrainFanoFactor

An analysis step that computes the Fano factor (FF) for a set of input spike trains. For each input spike train, the spike count is obtained. The Fano factor is defined as the ratio of the variance of the spike count to the mean spike count, across all spike trains. The Fano factor is usually computed for spike trains representing the activity of the same neuron over different trials. The value is interpreted as the higher the Fano factor value, the larger the cross-trial non-stationarity. For a stationary Poisson process, the Fano factor has value equal to 1.
has super-classes
neuronal firing regularity analysis c

compute spike train Pearson correlation coefficientc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ComputeSpikeTrainPearsonCorrelationCoefficient

A spike train correlation analysis that computes the Pearson correlation coefficient between two spike train inputs. The Pearson correlation coefficient is a real value that quantifies the linear relationship between the two spike trains. It has range [-1, 1], where 1 indicates a perfect positive linear relationship, -1 indicates a perfect negative linear relationship, and 0 indicates no linear relationship. For the computation, the input spike trains are discretized into time intervals (bins), and the spike count is obtained in each bin. The Pearson correlation coefficient is obtained by normalizing the covariance between the binned spike trains: the covariance is divided by the product of the standard deviation of each.
has super-classes
spike train correlation analysis c

compute spike train time histogramc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ComputeSpikeTrainTimeHistogram

An analysis step that computes the time histogram of a spike train. If the spike count in a bin is divided by the duration of the bin, this can be used to estimate the instantaneous firing rate at the bin interval.
has super-classes
spike train time histogram analysis c

compute spike waveform averagec back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ComputeSpikeWaveformAverage

A spike waveform analysis that computes the average across two or more spike waveform inputs. For the computation, the mean value across all inputs is obtained for each time point in the sampled spike waveform. This is frequently used to reduce noise across multiple spike waveform samples of a single neuron.
has super-classes
spike waveform analysis c

compute spike waveform signal-to-noise ratioc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ComputeSpikeWaveformSNR

A spike waveform analysis that computes the signal-to-noise ratio (SNR) for a set of input spike waveforms according to Hatsopoulos (2007). The SNR is defined as the difference in mean peak-to-trough voltage divided by twice the mean standard deviation (SD).The mean SD is obtained by averaging the SDs computed for each time point in the spike waveform.
has super-classes
spike waveform analysis c

compute spike waveform variancec back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ComputeSpikeWaveformVariance

A spike waveform analysis that computes the variance across two or more spike waveform inputs. For the computation, the value of the variance across all inputs is obtained for each time point in the spike waveform.
has super-classes
spike waveform analysis c

compute spike waveform widthc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ComputeSpikeWaveformWidth

A spike waveform analysis that computes the width of a spike waveform input. The computation takes two time points of interest (e.g, the times of the peak and the trough), and the width is the difference with respect to the time points (e.g., number of time points in between or time interval).
has super-classes
spike waveform analysis c

compute Spike-contrastc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ComputeSpikeContrast

A spike train synchrony analysis that computes the Spike-contrast measure using the method described by Ciba et al. (2018). Spike-contrast is a time-scale independent measure of spike synchrony. The input is a set of parallel spike train data recorded from a population of neurons. The algorithm is based on the temporal "contrast" (activity vs. non-activity in certain time bins). The computation outputs a single synchrony value (comparable to a spike train distance) and a synchrony curve showing the value of Spike-contrast as a function of the bin size.
has super-classes
spike train synchrony analysis c

compute spike-field coherence (Fries method)c back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ComputeSpikeFieldCoherenceFries

A spike-field coherence analysis that uses the method described in Fries et al. (2001) to compute the coherence between the spike train and the LFP. For the computation, first a spike-triggered average (STA) is obtained between the spike train and the LFP time series. Then, the power spectrum is obtained for each of the LFP segments used for the computation of the STA. These spectra are averaged to obtain the spike-triggered power spectrum. The SFC is then computed as the ratio of the power spectrum of the STA over the spike-triggered power spectrum.
has super-classes
spike-field coherence analysis c
is disjoint with
compute spike-field coherence (multitaper method) c, compute spike-field coherence (Welch method) c

compute spike-field coherence (multitaper method)c back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ComputeSpikeFieldCoherenceMultitaper

A spike-field coherence analysis that uses a multitaper approach to compute the coherence between the spike train and the LFP.
has super-classes
spike-field coherence analysis c
is disjoint with
compute spike-field coherence (Fries method) c, compute spike-field coherence (Welch method) c

compute spike-field coherence (Welch method)c back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ComputeSpikeFieldCoherenceWelch

A spike-field coherence analysis step that uses the method by Welch (1967) to compute the coherence between the spike train and the LFP.
has super-classes
spike-field coherence analysis c
is disjoint with
compute spike-field coherence (Fries method) c, compute spike-field coherence (multitaper method) c

compute spike-triggered averagec back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ComputeSpikeTriggeredAverage

A triggered average analysis that uses spike times as triggers to obtain the average of a signal around each spike (spike-triggered average).
has super-classes
triggered average analysis c
has sub-classes
compute spike-triggered local field potential average c

compute spike-triggered local field potential averagec back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ComputeSpikeTriggeredLFPAverage

A triggered average analysis that uses spike times as triggers to average the local field potential (LFP) signal. The LFP is the low-frequency component of the potential recorded within a specific region of the brain using extracellular electrodes. The output of the method will provide an estimation of the average LFP voltage around each spike, i.e., the spike-triggered average of the LFP signal.
has super-classes
compute spike-triggered average c

compute spike-triggered phasec back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ComputeSpikeTriggeredPhase

A phase analysis that computes the phase angle values of an analytic signal input (or from the analytic signal obtained from an input time series) at the time points where spikes occurred. The spike times are defined in a spike train input. The output is an array with the phase angle at each spike time in the input spike train (spike-triggered phases).
has super-classes
phase analysis c

compute standard deviationc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ComputeStandardDeviation

A dispersion statistical analysis that computes the standard deviation (SD), i.e., the square root of the variance. The SD indicates the average distance of each data point from the mean.
has super-classes
dispersion statistical analysis c

compute standard error of the meanc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ComputeStandardErrorMean

A dispersion statistical analysis that computes the standard error of the mean (SEM). The SEM is the standard deviation of the sampling distribution of the sample mean. It provides an estimate of how much the sample mean is expected to fluctuate around the true population mean. Smaller SEM values indicates that the sample mean is a more accurate estimate of the population mean. The SEM decreases as the sample size increases, as larger samples provide a more reliable estimate of the population mean.
has super-classes
dispersion statistical analysis c

compute Stockwell transformc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ComputeStockwellTransform

A time-frequency analysis that computes the Stockwell transform (S transform) of the input time series. The S transform generalizes the short-time Fourier transform (STFT) and extends the continuous wavelet transform (CWT). The main difference is that STFT uses a constant window width for all frequencies. The S transform is based on a moving and scalable localizing Gaussian window. Therefore, the window is frequency-dependent (adaptive windowing), which results in better time resolution in higher frequencies and better frequency resolution at lower frequencies. This makes the S transform more suitable to detect transient signals in high frequencies. The computation is computationally expensive, although fast algorithms are available.
has super-classes
time-frequency analysis c

compute time domain conditional Granger causalityc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ComputeTimeDomainConditionalGrangerCausality

A conditional Granger causality (GC) analysis that computes the GC measures in the time domain.
has super-classes
conditional Granger causality analysis c
is disjoint with
compute frequency domain conditional Granger causality c

compute time domain pairwise Granger causalityc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ComputeTimeDomainPairwiseGrangerCausality

A pairwise Granger causality (GC) analysis that computes measures of GC between two inputs in the time domain. The computation involves fitting two separate autoregressive (AR) models to the input data and comparing their fit to determine if one time series can predict the other. The quality of the fit is assessed by the variance of the residuals, with GC defined as the natural logarithm of the ratio of the residual variances from the two AR models. The first AR model is univariate, predicting the future values of one time series (e.g., the first) only from its past values. The second AR model is bivariate, predicting the future values of the first time series from its past values as well as the past values of the second time series. If the bivariate model reduces the variance of the residuals (ratio greater than 1), the GC value will be positive, indicating that the second time series Granger causes the first (directional GC estimate from the second to the first). The same method is used to predict the second time series from the first, yielding the directional GC estimate from the first to the second time series. The order of the AR model is defined as parameter, and optimal values can be estimated using optimization techniques.
has super-classes
pairwise Granger causality analysis c
is disjoint with
frequency domain pairwise granger causality analysis c

compute transfer entropyc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ComputeTransferEntropy

A functional connectivity analysis that computes a measure of transfer entropy (TE) between two input time series. TE measures the directional transfer of information between the time series. It extends Granger causality, and is able to detect non-linear forms of interaction.
has super-classes
functional connectivity analysis c

compute tuning curvec back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ComputeTuningCurve

A firing rate analysis that computes a tuning curve. The tuning curve describes the firing rate of a neuron as a function of a continuous attribute (e.g., orientation of a visual grating stimulus).
has super-classes
firing rate analysis c

compute unbiased squared phase lag indexc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ComputeUnbiasedSquaredPhaseLagIndex

A phase lag index (PLI) analysis that computes an unbiased estimate for the squared PLI following Vinck et al. (2011). The direct PLI estimator is positively biased, especially when the sample sizes (i.e., number of trials) are small. The unbiased squared PLI is computed by averaging all pairwise products of the signs computed across the repetitions. Pairs with identical observations are excluded. The unbiased squared PLI is less affected by small-sample size biases.
has super-classes
phase lag index analysis c

compute van Rossum distancec back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ComputeVanRossumDistance

A time-scale dependent spike train distance analysis that computes the van Rossum distance introduced in van Rossum (2001). For the computation, each spike in the input spike trains is convolved with an exponential kernel, producing continuous function representations of the input spike trains. The time scale parameter of the distance is set by the time constant of the exponential kernel. The distance is then obtained as the Euclidean distance of the convolved spike trains.
has super-classes
time-scale dependent spike train distance analysis c

compute variancec back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ComputeVariance

A dispersion statistical analysis that computes the variance, i.e., the average of the squared differences from the mean.
has super-classes
dispersion statistical analysis c

compute Victor-Purpura distancec back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ComputeVictorPurpuraDistance

A time-scale dependent spike train distance analysis that computes the Victor-Purpura distance, introduced in Victor & Purpura (1996). The metric defines the distance between two spike train inputs with respect to the minimum cost of transforming one spike train into the order considering three operations: spike insertion, spike deletion and shifting a spike by some interval. The first two operations have a fixed cost equal to 1. The latter depends on a cost per time unit parameter, which sets the time scale of the analysis.
has super-classes
time-scale dependent spike train distance analysis c

compute weighted phase lag indexc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ComputeWeightedPhaseLagIndex

A phase lag index (PLI) analysis that computes the weighted PLI (WPLI) following Vinck et al. (2011). The original PLI (Stam, 2007) is discontinuous, and small perturbations can turn phase lags into leads (and vice versa). Therefore, this hinders its capacity to detect changes in phase synchronization of small magnitude. The WPLI extends the PLI to weight the contributions of the phase leads and lags by the magnitude of the imaginary component of the cross spectral density. Therefore, these increases the power to detect changes in phase synchronization. The WPLI ranges between 0 and 1. A WPLI of zero means that there is no imbalance in the first time series leading or lagging the second (i.e., the total weight of all leading relationships is equal to the total weight of lagging relationships). A value greater than zero means an imbalance in the likelihood of leading or lagging. A WPLI of 1 indicates perfect phase locking, and that the first time series only leads or only lags.
has super-classes
phase lag index analysis c

conditional Granger causality analysisc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ConditionalGrangerCausalityAnalysis

A Granger causality (GC) analysis that computes a measure of conditional GC between the inputs. The conditional GC is the causality between two inputs (e.g., two time series) while controlling for the influence of an additional input (e.g., a third time series). This allows a more complete understanding of the causal relationships in multivariate time series data.
has super-classes
Granger causality analysis c
has sub-classes
compute frequency domain conditional Granger causality c, compute time domain conditional Granger causality c
is disjoint with
pairwise Granger causality analysis c

confidence interval statistical analysisc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ConfidenceIntervalStatisticalAnalysis

A statistical analysis that computes a confidence interval (CI). The CI is a measure providing a range of values of a parameter, derived from the input (sample) data, that is likely to contain the true value of the parameter in the population with a specified level of confidence. The level of confidence is specified as a parameter to the method. For example, a 95% confidence interval means that if the same population is sampled multiple times, approximately 95% of the intervals calculated from those samples will contain the true population parameter. Confidence intervals are used to estimate parameters such as the mean and are essential for making inferences about the population based on sample data. The output contains the upper and lower limits of the CI.
has super-classes
statistical analysis c
has sub-classes
compute confidence interval (non-resampling) c, confidence interval with resampling analysis c

confidence interval with resampling analysisc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ConfidenceIntervalResamplingAnalysis

A confidence interval statistical analysis that computes the confidence interval using techniques to generate multiple samples from the (observed) data input(s).
has super-classes
confidence interval statistical analysis c
has sub-classes
compute confidence interval (bootstrap resampling) c, compute confidence interval (jackknife resampling) c
is disjoint with
compute confidence interval (non-resampling) c

correlated spike times generationc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#CorrelatedSpikeTimesGeneration

A spike train generation where the output spike trains will have spikes that are correlated in time. These methods can be used to generate spike trains with patterns in their activity.
has super-classes
spike train generation c
has sub-classes
generate compound Poisson process c, generate single interaction process c
is disjoint with
random spike times generation c

correlation analysisc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#CorrelationAnalysis

An analysis step that computes a measure of correlation between two inputs. Correlation is a measure that quantifies the strength to which two variables change together. It is a scaled version of the covariance, and the values are restricted to the -1 to +1 interval. Between time series, it is computed in the time domain.
is equivalent to
has purpose op value correlation purpose
has super-classes
analysis step c
has sub-classes
compute orthogonalized power envelope correlation c

covariance analysisc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#CovarianceAnalysis

An analysis step used to compute a measure of covariance from two inputs. The covariance indicates the extent to which the two inputs change together. If the greater values of one variable mainly correspond with the greater values of the other variable, and the same applies to the lesser values, then the covariance is positive. Conversely, if greater values of one variable mainly correspond to the lesser values of the other variable, then the covariance is negative.
has super-classes
analysis step c
has sub-classes
compute covariance c, compute noise covariance c

cross power spectral density analysisc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#CrossPowerSpectralDensityAnalysis

A spectral density analysis that computes the cross power spectral density (CPSD) of two inputs, i.e., the distribution of their power across the different frequency components per unit frequency. The CPSD shows in the frequency domain how the two inputs are correlated in the time domain, and is equivalent to the Fourier transform of the cross-correlation function between the two signals. If the inputs are not correlated, the CPSD will be flat across the frequencies. A peak suggests that the signals are correlated at that frequency. The CPSD is often referred to as cross-spectrum.
has super-classes
spectral density analysis c
has sub-classes
compute cross power spectral density (Morlet wavelet method) c, compute cross power spectral density (Welch method) c, compute cross power spectral density (multitaper method) c, compute cross power spectral density (periodogram method) c

cross-correlation analysisc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#CrossCorrelationAnalysis

An analysis step used to compute a measure of cross-correlation, i.e., the correlation of two inputs computed for distinct time lags of the first to the second. The computation produces the cross-correlation value for every lag considered.
has super-classes
analysis step c
has sub-classes
compute cross-correlation function c, compute spike train cross-correlation histogram c

current source density analysisc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#CurrentSourceDensityAnalysis

An analysis step used to analyze extracellular electrical potentials (e.g., local field potentials or evoked potentials) recorded from multiple locations, enabling the estimation of the current sources responsible for generating these potentials. The methods can be applied to data recorded from different electrode configurations: laminar probe-like electrodes (1D methods), microelectrode array-like electrodes (2D methods) or electrodes configurations recording from a volume (e.g., multiple laminar probes or array electrodes with shanks with multiple depths; 3D methods). The output of the current source analysis provides the spatial map showing where currents are entering (sources) and leaving (sinks) the neural tissue. For each point in the map, a quantitative value indicates the magnitude of the current density at that point.
has super-classes
analysis step c
has sub-classes
compute current source density (inverse method) c, compute current source density (kernel method) c, compute current source density (standard method) c

data generationc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#DataGeneration

An analysis step that generates new data entities. This can be done using other input data as a start (e.g., generating a spike train surrogate from spike trains obtained from experimental recordings) or generate new data using algorithms that take specific parameters (e.g., generating a spike train using a probability distribution defined by specific parameters).
has super-classes
analysis step c
has sub-classes
artificial data generation c, generate Morlet wavelet c, spike train generation c, spike train surrogate generation c

data normalizationc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#DataNormalization

A data transformation that adjusts the ranges and distributions of the values in the input data. This can be used to transform data measured in distinct (i.e. not directly comparable) scales to a common (i.e. comparable) scale.
has super-classes
data transformation c
has sub-classes
apply median rescaling c, apply min-max normalization c, apply z-score transform c

data smoothingc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#DataSmoothing

A data transformation that uses statistical techniques to remove noise and fluctuations from the input data to reveal underlying trends and patterns.
has super-classes
data transformation c
has sub-classes
apply local regression and likelihood smoothing c, kernel smoothing c

data transformationc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#DataTransformation

An analysis step that takes one or more inputs and modifies the contents, such that it fits a particular purpose. The data transformation steps are frequently used during pre-processing the input data for the analyses. Usually, a data transformation will not change the main representation or quality of the inputs. For example, a digital filtering step will remove certain frequency components from a time series. However, the output will resemble the original input with respect to shape or physical units. In addition, the inputs can also be converted to other formats or representations that are needed for a particular analysis step. For example, spike trains can be discretized into small intervals (binning) or the dimensionality of the input can be reduced using principal component analysis. The data transformation is in contrast to steps that perform computations that take the input and generate a derived measure with new information. For example, when computing the mean firing rate from a spike train, a single scalar value is obtained from the spike count in the input data.
has super-classes
analysis step c
has sub-classes
apply Hilbert transform c, apply analytic signal conversion c, apply data concatenation c, apply independent component analysis c, apply interpolation c, apply padding c, apply rectification c, apply rereference c, apply spike extraction from time series c, apply spike train binarization c, apply spike train binning c, apply spike waveform peak alignment c, apply sum c, apply trial extraction c, artifact removal c, data normalization c, data smoothing c, detrending c, digital filtering c, dimensionality reduction c, frequency domain transformation c, line noise removal c, resampling c

detrendingc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#Detrending

A data transformation that removes a trend (i.e., a change in the mean over time) from an input time series.
has super-classes
data transformation c
has sub-classes
apply general linear model polynomial detrending c, apply local linear regression detrending c

digital filteringc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#DigitalFiltering

A data transformation that processes digital signal inputs (i.e., sampled time series) to attenuate or amplify specific frequency components. Digital filters can be designed using various methods, achieving distinct frequency responses and stability.
has super-classes
data transformation c
has sub-classes
finite impulse response filtering c, infinite impulse response filtering c

dimensionality reductionc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#DimensionalityReduction

A data transformation that obtains a low-dimensional (simplified) representation of the high-dimensional input data. This step retains important information in the input data while minimizing redundancy and noise.
has super-classes
data transformation c
has sub-classes
apply distance covariance analysis c, apply linear discriminant analysis c, apply neural trajectory Gaussian process factor analysis c, principal component analysis c, tensor component analysis c

directed analysisc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#DirectedAnalysis

An analysis step where the output provides information on the direction of influence in the relationships among the inputs (e.g., in the cross-correlation histogram, it is possible to analyze the timing of the spikes of the first input spike train with respect to the timing of the spikes of the second input spike train, i.e., whether they likely occur before or after).
is equivalent to
is directed dp value true
has super-classes
analysis step c
is disjoint with
non-directed analysis c

dispersion statistical analysisc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#DispersionStatisticalAnalysis

A statistical analysis that computes a measure that represents the variability in the input data. It indicated the degree to which the data points differ from the central tendency.
has super-classes
statistical analysis c
has sub-classes
compute coefficient of variation c, compute interquartile range c, compute standard deviation c, compute standard error of the mean c, compute variance c

distance analysisc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#DistanceAnalysis

is equivalent to
has purpose op value distance purpose
has super-classes
analysis step c

execute 3d-SPADE analysisc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#Execute3dSPADEAnalysis

A SPADE analysis that uses the 3d-SPADE implementation as defined in Stella et al. (2019). The 3d-SPADE analysis considers spatio-temporal patterns (i.e., not restricted to synchronous patterns), and the pattern signature used for statistical testing considers the size, number of occurrences and duration of a pattern.
has super-classes
SPADE analysis c
is disjoint with
execute non-3d-SPADE analysis c

execute ASSET analysisc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ExecuteASSETAnalysis

A neuronal activity pattern detection analysis that uses the Analysis of Sequences of Synchronous EvenTs (ASSET) method defined in Torre et al. (2016). ASSET is an automatized test to identify the sequential activations of groups of neurons that repeat in time. The identification of the repeated sequences is possible by computing the intersection matrix. The input spike data is discretized into smaller intervals (bins), and the overlap of neuronal activity at each bin pair is obtained as a matrix. When a sequence of activations exists and repeats in time, a characteristic diagonal structure appears in the matrix. The ASSET analysis provides a robust statistical test to automatically identify the diagonal structures and to provide the neurons and their activation pattern in each repeated sequence. Overall, the analysis is composed by 6 substeps: 1. Compute the intersection matrix (IMAT) from a set of input spike trains, using a specified bin size to discretize the data. 2. Obtain the probability matrix (PMAT). 3. Obtain the joint probability matrix (JMAT). 4. Extract significant entries in both PMAT and JMAT using specified thresholds, obtaining the mask matrix (MMAT). 5. Obtain the cluster matrix (CMAT) by using DBSCAN to cluster the significant entries in the MMAT to find each diagonal structure. Parameters for DBSCAN control the clustering result. A modified distance metric is used. 6. From the identified clusters (each a single diagonal structure in the IMAT), obtain the neuronal composition and the order of activation, producing the final neuronal activity patterns as output.
has super-classes
neuronal activity pattern detection analysis c

execute Cell Assembly Detection analysisc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ExecuteCellAssemblyDetectionAnalysis

A neuronal activity pattern detection analysis that uses the Cell Assembly Detection (CAD) method as defined in Russo & Durstewitz (2017). CAD allows detecting spatio-temporal spike patterns at different time scales, levels of precision, and with arbitrary internal organization. The analysis identifies patterns with different delays between the spikes (within a window determining the minimum and maximum lags), and is performed in two steps using an agglomerative clustering algorithm. First, significant pairwise correlations are identified, which is followed by the clustering procedure that progressively finds interactions of higher order. At each agglomeration step, the method can filter out patterns involving the same neurons, keeping the most significant pattern (significance pruning). In an additional pruning step, assemblies part of a larger assembly can also be eliminated (controlled by the subgroup pruning parameter). The algorithm stops when the detected assemblies reach their maximum size (determined by a parameter). The statistical test assumes independence under non-stationarity and Poisson distribution of the input spike trains.
has super-classes
neuronal activity pattern detection analysis c

execute non-3d-SPADE analysisc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ExecuteNon3dSPADEAnalysis

A SPADE analysis that uses the SPADE implementation as defined in Quaglio et al. (2017). This is the extension of SPADE to consider spatio-temporal patterns (i.e., patterns not restricted to synchronous spiking). In the non-3d SPADE analysis, the pattern signature used for statistical testing considers only the size and number of occurrences of a pattern.
has super-classes
SPADE analysis c
is disjoint with
execute 3d-SPADE analysis c

execute Unitary Event analysis (analytical method)c back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ExecuteUnitaryEventAnalysisAnalytical

A unitary event (UE) analysis that uses the analytical approach to determine the significance of the empirical coincidences in binned spike train data, as defined in Grün et al. (1999, 2002a, 2002b, and 2003). The analytical method tests if the number of empirical coincidences is consistent with the coincidence distribution resulting from independent processes. This distribution can be expressed analytically assuming that the input spike trains follow Poisson statistics. The UEs can be determined trial by trial, where the analytical expectancy is computed for each trial and then summed over all trials, or by averaging over all trials (according to Grün, 2003).
has super-classes
Unitary Event analysis c
is disjoint with
execute Unitary Event analysis (Monte Carlo method) c

execute Unitary Event analysis (Monte Carlo method)c back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ExecuteUnitaryEventAnalysisMonteCarlo

A unitary event (UE) analysis that uses a Monte Carlo approach based on spike train surrogates to determine the significance of the empirical coincidences in binned spike train data, according to Grün (2009). The Monte Carlo method does not rely on the assumption that the input spike data follows Poisson statistics. For the assessment of significance, the distribution of expected coincidences is determined by surrogates (spike train randomization) in each trial, and then summed over trials. The number of surrogates is determined by a parameter.
has super-classes
Unitary Event analysis c
is disjoint with
execute Unitary Event analysis (analytical method) c

field-field coupling analysisc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#FieldFieldCouplingAnalysis

An analysis step that computes interactions between the neural activity represented by distinct local field potential (LFP) signals (e.g., LFP obtained from different electrodes, or distinct LFP frequency bands).
is equivalent to
has purpose op value field-field coupling purpose
has super-classes
analysis step c
is disjoint with
spike-spike coupling analysis c

finite impulse response filteringc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#FiniteImpulseResponseFiltering

A digital filtering that uses a filter whose impulse response (i.e., the response of the filter to a brief input impulse) decays to zero after a finite amount of time. Therefore, the output of the filter depends on a finite number of past samples. The finite impulse response (FIR) filters are stable and can be designed such that they do not distort the phase of the signal. However, they have a higher computational cost.
has super-classes
digital filtering c
has sub-classes
apply finite impulse response filter c, apply finite impulse response filter with Kaiser window c, apply windowed-sinc filter c

firing rate analysisc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#FiringRateAnalysis

An analysis step used to compute a measure quantifying the firing rate of one or more neurons. The firing rate is the number of action potential (spikes) that a neuron fires per time unit and is defined with a unit of frequency (e.g., Hz or spikes/s).
has super-classes
analysis step c
has sub-classes
compute mean firing rate c, compute tuning curve c, instantaneous firing rate analysis c, spike train time histogram analysis c

frequency domain analysisc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#FrequencyDomainAnalysis

An analysis step that analyzes the input(s) with respect to its(their) frequency content.
is equivalent to
is frequency domain dp value true
has super-classes
analysis step c
is disjoint with
time domain analysis c

frequency domain pairwise granger causality analysisc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#FrequencyDomainPairwiseGrangerCausalityAnalysis

A pairwise Granger causality (GC) analysis that computes measures of GC between two inputs in the frequency domain. This is an extension of the GC concept in the time domain, and the measures of GC are obtained for the different frequency components of the inputs, according to Geweke (1982). The computation of the frequency-domain GC measures is based on two elements: the noise covariance matrix and the spectral transfer matrix. These can be estimated either with parametric or non-parametric methods. For the parametric estimation, an autoregressive model is fit and the Fourier transform of the autoregressive coefficients is used to obtain the spectral transfer matrix. For the non-parametric estimation, the cross-spectral density (CSD) matrix is obtained (using methods for CSD estimation such as multitapering or wavelet), and the CSD matrix is factorized to obtain the noise covariance and spectral transfer matrices.
has super-classes
pairwise Granger causality analysis c
has sub-classes
compute frequency domain pairwise Granger causality (Brovelli method) c, compute frequency domain pairwise Granger causality (Dhamala method) c, compute frequency domain pairwise Granger causality (Hafner method) c, compute frequency domain pairwise Granger causality (Wen method) c
is disjoint with
compute time domain pairwise Granger causality c

frequency domain transformationc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#FrequencyDomainTransformation

A data transformation that converts a time series input from the time to the frequency domain, i.e., reveal the different frequency components that make up the original signal.
has super-classes
data transformation c
has sub-classes
apply discrete Fourier transform c

functional connectivity analysisc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#FunctionalConnectivityAnalysis

An analysis step that computes measures of functional connectivity. Functional connectivity refers to statistical dependencies and patterns of synchronization between the neural activity that indicate the functional interactions and co-activations that are relevant for the function of the nervous system (e.g., the interactions between different brain regions). It does not imply direct physical connections.
is equivalent to
has purpose op value functional connectivity purpose
has super-classes
analysis step c
has sub-classes
Granger causality analysis c, compute directed transfer function c, compute multivariate interaction measure c, compute mutual information c, compute transfer entropy c

generate compound Poisson processc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#GenerateCompoundPoissonProcess

A correlated spike times generation that produces spike trains using a compound Poisson process (CPP) according to Staude et al. (2010). The CPP is a model for parallel and correlated processes with Poisson spiking statistics at predefined firing rates.
has super-classes
correlated spike times generation c
is disjoint with
generate single interaction process c

generate interspike interval shuffling surrogatec back to ToC or Class ToC

IRI: http://purl.org/neao/steps#GenerateISIShufflingSurrogate

A spike train surrogate generation step where the interspike intervals (ISIs) of the input spike train is randomly sorted. This preserves the ISI distribution and spike count as in the original spike train input, but destroys temporal dependencies and firing rate profile.
has super-classes
spike train surrogate generation c
is disjoint with
generate joint interspike interval dithering surrogate c, generate joint interspike interval dithering surrogate c, generate spike time randomization surrogate c, generate spike train dithering surrogate c, generate trial shifting surrogate c, generate trial shuffling surrogate c, generate uniform spike dithering surrogate c, generate uniform spike dithering surrogate with dead time c, generate window shuffling surrogate c

generate joint interspike interval dithering surrogatec back to ToC or Class ToC

IRI: http://purl.org/neao/steps#GenerateISIDitheringSurrogate

A spike train surrogate generation step where each spike is displaced according to the interspike interval (ISI) distribution sampled from the input spike train.
has super-classes
spike train surrogate generation c
is disjoint with
generate interspike interval shuffling surrogate c, generate joint interspike interval dithering surrogate c, generate spike time randomization surrogate c, generate spike train dithering surrogate c, generate trial shifting surrogate c, generate trial shuffling surrogate c, generate uniform spike dithering surrogate c, generate uniform spike dithering surrogate with dead time c, generate window shuffling surrogate c

generate joint interspike interval dithering surrogatec back to ToC or Class ToC

IRI: http://purl.org/neao/steps#GenerateJointISIDitheringSurrogate

A spike train surrogate generation step where spikes from adjacent interspike intervals (ISIs) are dithered according to the joint-ISI (JISI) probability distribution. The distribution is obtained from the input spike train by computing the JISI histogram (i.e., a two-dimensional histogram that shows the frequency of ISIs with a given duration that are immediately followed by intervals with another duration). Due to non-stationarities in the input spike train and/or its limited duration, it is difficult to accurately estimate the underlying JISI probability distribution. Therefore, a 2D-Gaussian smoothing is applied to the JISI histogram (with a variance determined by parameter). Dithering a spike according to this (smoothed) two-dimensional histogram involves moving the spike along the anti-diagonal of the JISI distribution. The dithering time is defined by a parameter.
has super-classes
spike train surrogate generation c
is disjoint with
generate joint interspike interval dithering surrogate c, generate interspike interval shuffling surrogate c, generate spike time randomization surrogate c, generate spike train dithering surrogate c, generate trial shifting surrogate c, generate trial shuffling surrogate c, generate uniform spike dithering surrogate c, generate uniform spike dithering surrogate with dead time c, generate window shuffling surrogate c

generate Morlet waveletc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#GenerateMorletWavelet

A data generation that constructs a Morlet wavelet considering the selected parameters (i.e., sampling frequency, fundamental frequency, and number of cycles per frequency). The Morlet wavelet is composed by a complex exponential multiplied by a Gaussian window. The output data contains the discrete time points and the corresponding wavelet values.
has super-classes
data generation c

generate non-stationary gamma processc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#GenerateNonStationaryGammaProcess

A random spike times generation that uses a gamma probability distribution to produce spike trains where the firing rate varies over time.
has super-classes
random spike times generation c
is disjoint with
generate non-stationary Poisson process c, generate stationary gamma process c, generate stationary inverse Gaussian process c, generate stationary log-normal process c, generate stationary Poisson process c

generate non-stationary Poisson processc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#GenerateNonStationaryPoissonProcess

A random spike times generation that uses a Poisson probability distribution to produce spike trains where the firing rate varies over time. A dead time can be specified as parameter (i.e., an interval where no spikes can occur, which is analogous to the neuronal refractory period).
has super-classes
random spike times generation c
is disjoint with
generate non-stationary gamma process c, generate stationary gamma process c, generate stationary inverse Gaussian process c, generate stationary log-normal process c, generate stationary Poisson process c

generate single interaction processc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#GenerateSingleInteractionProcess

A correlated spike times generation that produces a multidimensional Poisson single interaction process (SIP) plus independent Poisson processes, according to Kuhn et al. (2003). The Poisson SIP consists of Poisson time series that are independent except for events that are simultaneous in all of them.
has super-classes
correlated spike times generation c
is disjoint with
generate compound Poisson process c

generate spike time randomization surrogatec back to ToC or Class ToC

IRI: http://purl.org/neao/steps#GenerateSpikeTimeRandomizationSurrogate

A spike train surrogate generation step that keeps the spike count of the input spike train, but the spike times in the surrogate spike train output are randomly chosen within the duration interval of the input spike train.
has super-classes
spike train surrogate generation c
is disjoint with
generate joint interspike interval dithering surrogate c, generate interspike interval shuffling surrogate c, generate joint interspike interval dithering surrogate c, generate spike train dithering surrogate c, generate trial shifting surrogate c, generate trial shuffling surrogate c, generate uniform spike dithering surrogate c, generate uniform spike dithering surrogate with dead time c, generate window shuffling surrogate c

generate spike train dithering surrogatec back to ToC or Class ToC

IRI: http://purl.org/neao/steps#GenerateSpikeTrainDitheringSurrogate

A spike train surrogate generation step that displaces the whole input spike train by a random amount of time (independent for each surrogate generated). The amount of displacement is specified as a parameter (dither time), and occurs in a window (-dither time, dither time). This surrogate maintains the ISIs and the temporal correlations within the spike train.
has super-classes
spike train surrogate generation c
is disjoint with
generate joint interspike interval dithering surrogate c, generate interspike interval shuffling surrogate c, generate joint interspike interval dithering surrogate c, generate spike time randomization surrogate c, generate trial shifting surrogate c, generate trial shuffling surrogate c, generate uniform spike dithering surrogate c, generate uniform spike dithering surrogate with dead time c, generate window shuffling surrogate c

generate stationary gamma processc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#GenerateStationaryGammaProcess

A random spike times generation that uses a gamma probability distribution to produce spike trains with a constant firing rate.
has super-classes
random spike times generation c
is disjoint with
generate non-stationary gamma process c, generate non-stationary Poisson process c, generate stationary inverse Gaussian process c, generate stationary log-normal process c, generate stationary Poisson process c

generate stationary inverse Gaussian processc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#GenerateStationaryInverseGaussianProcess

A random spike times generation that uses a inverse Gaussian probability distribution to produce spike trains with a constant firing rate.
has super-classes
random spike times generation c
is disjoint with
generate non-stationary gamma process c, generate non-stationary Poisson process c, generate stationary gamma process c, generate stationary log-normal process c, generate stationary Poisson process c

generate stationary log-normal processc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#GenerateStationaryLogNormalProcess

A random spike times generation that uses a log-normal probability distribution to produce spike trains with a constant firing rate.
has super-classes
random spike times generation c
is disjoint with
generate non-stationary gamma process c, generate non-stationary Poisson process c, generate stationary gamma process c, generate stationary inverse Gaussian process c, generate stationary Poisson process c

generate stationary Poisson processc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#GenerateStationaryPoissonProcess

A random spike times generation that uses a Poisson probability distribution to produce spike trains with a constant firing rate. A dead time can be specified as parameter (i.e., an interval where no spikes can occur, which is analogous to the neuronal refractory period).
has super-classes
random spike times generation c
is disjoint with
generate non-stationary gamma process c, generate non-stationary Poisson process c, generate stationary gamma process c, generate stationary inverse Gaussian process c, generate stationary log-normal process c

generate trial shifting surrogatec back to ToC or Class ToC

IRI: http://purl.org/neao/steps#GenerateTrialShiftingSurrogate

A spike train surrogate generation step that shifts the entire spike train (containing the data of a single experimental trial) by an amount randomly chosen from a uniform distribution. The amount of displacement is specified as a parameter (dither time), and occurs in a window (-dither time, dither time). The input is a collection of spike trains of the same neuron, containing the spiking activity during different experimental trials. The amount of shift is independently chosen across trials and neurons. This surrogate preserves the ISI distribution and temporal correlations within the single-trial spike train.
has super-classes
spike train surrogate generation c
is disjoint with
generate joint interspike interval dithering surrogate c, generate interspike interval shuffling surrogate c, generate joint interspike interval dithering surrogate c, generate spike time randomization surrogate c, generate spike train dithering surrogate c, generate trial shuffling surrogate c, generate uniform spike dithering surrogate c, generate uniform spike dithering surrogate with dead time c, generate window shuffling surrogate c

generate trial shuffling surrogatec back to ToC or Class ToC

IRI: http://purl.org/neao/steps#GenerateTrialShufflingSurrogate

A spike train surrogate generation step where the the spike trains of single-trial activity of one of the neurons are randomly permuted, so that each trial is no longer paired with the corresponding trial of the other neuron, but with a randomly selected one. The input is a collection of spike trains with multitrial activity data of multiple neurons recorded in parallel.
has super-classes
spike train surrogate generation c
is disjoint with
generate joint interspike interval dithering surrogate c, generate interspike interval shuffling surrogate c, generate joint interspike interval dithering surrogate c, generate spike time randomization surrogate c, generate spike train dithering surrogate c, generate trial shifting surrogate c, generate uniform spike dithering surrogate c, generate uniform spike dithering surrogate with dead time c, generate window shuffling surrogate c

generate uniform spike dithering surrogatec back to ToC or Class ToC

IRI: http://purl.org/neao/steps#GenerateUniformSpikeDitheringSurrogate

A spike train surrogate generation step that displaces each spike time in the input spike train according to a uniform distribution centered on the spike. The displacement occurs in a time window around the spike defined by a parameter (dither time).
has super-classes
spike train surrogate generation c
is disjoint with
generate joint interspike interval dithering surrogate c, generate interspike interval shuffling surrogate c, generate joint interspike interval dithering surrogate c, generate spike time randomization surrogate c, generate spike train dithering surrogate c, generate trial shifting surrogate c, generate trial shuffling surrogate c, generate uniform spike dithering surrogate with dead time c, generate window shuffling surrogate c

generate uniform spike dithering surrogate with dead timec back to ToC or Class ToC

IRI: http://purl.org/neao/steps#GenerateUniformSpikeDitheringSurrogateWithDeadTime

A spike train surrogate generation step that displaces each spike time in the input spike train according to a uniform distribution centered on the spike. The displacement occurs in a time window around the spike defined by a parameter (dither time) that is constrained to the intervals between adjacent spikes to avoid two spikes closer than a dead time (specified by parameter). This mimics the refractory period behavior of neurons, where the neuron cannot fire additional spikes for a short interval after one spike.
has super-classes
spike train surrogate generation c
is disjoint with
generate joint interspike interval dithering surrogate c, generate interspike interval shuffling surrogate c, generate joint interspike interval dithering surrogate c, generate spike time randomization surrogate c, generate spike train dithering surrogate c, generate trial shifting surrogate c, generate trial shuffling surrogate c, generate uniform spike dithering surrogate c, generate window shuffling surrogate c

generate window shuffling surrogatec back to ToC or Class ToC

IRI: http://purl.org/neao/steps#GenerateWindowShufflingSurrogate

A spike train surrogate generation step that shuffles the entries of a binned spike train within exclusive maximal displacement windows. The maximal displacement is specified by parameter, and represents the maximum number of bins that a spike can be displaced within the window.
has super-classes
spike train surrogate generation c
is disjoint with
generate joint interspike interval dithering surrogate c, generate interspike interval shuffling surrogate c, generate joint interspike interval dithering surrogate c, generate spike time randomization surrogate c, generate spike train dithering surrogate c, generate trial shifting surrogate c, generate trial shuffling surrogate c, generate uniform spike dithering surrogate c, generate uniform spike dithering surrogate with dead time c

Granger causality analysisc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#GrangerCausalityAnalysis

A functional connectivity analysis that computes measures of Granger causality (GC). GC is a statistical concept where the future values of a time series are predicted based on its past values and the past values of other time series. GC quantifies bi-directional interactions between the inputs, determining the directional influence from one input to another. For example, with two inputs, GC measures how much the first input influences the second and vice versa (directional GC measure). This provides estimates of the directed connectivity between the inputs. It is also possible to compute associated measures, such as the instantaneous GC (a measure of interdependence between the inputs not accounted by their bi-directional interactions, such as shared neural input) and the total interdependence (the sum of all directional and instantaneous interactions between the inputs). The analysis can be performed in the time or frequency domains, and can take two (bivariate) or more inputs (multivariate).
has super-classes
functional connectivity analysis c
has sub-classes
conditional Granger causality analysis c, pairwise Granger causality analysis c

infinite impulse response filteringc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#InfiniteImpulseResponseFiltering

A digital filtering that uses a filter whose impulse response (i.e., the response of the filter to a brief input impulse) persists infinitely. Therefore, the output of the filter can depend on an infinite number of past samples. The infinite impulse response (IIR) filters can become unstable and distort the phase of the signal, but have a lower computational cost.
has super-classes
digital filtering c
has sub-classes
apply Butterworth filter c, apply infinite impulse response filter c, apply notch filter c

instantaneous firing rate analysisc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#InstantaneousFiringRateAnalysis

A firing rate analysis that computes the instantaneous firing rate, which is the estimate of the firing rate at a specific point in time. The instantaneous firing rate value can be obtained by several methods.
is equivalent to
has purpose op value instantaneous firing rate purpose
has super-classes
firing rate analysis c
has sub-classes
compute instantaneous firing rate (interspike interval method) c, compute instantaneous firing rate (kernel density estimation method) c, compute instantaneous firing rate (local regression method) c

interspike interval analysisc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#InterspikeIntervalAnalysis

An analysis step that computes or analyzes the interval between successive spikes in a spike train (interspike interval; ISI).
has super-classes
analysis step c
has sub-classes
compute interspike interval histogram c, compute interspike intervals c, interspike interval variability analysis c

interspike interval variability analysisc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#InterspikeIntervalVariabilityAnalysis

An interspike interval analysis that computes a measure describing the variability of the interspike intervals. The measure can assess how regular a neuron is firing.
has super-classes
interspike interval analysis c
has sub-classes
compute CV2 c, compute coefficient of variation of the interspike intervals c, compute local variation c, compute revised local variation c

kernel smoothingc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#KernelSmoothing

A data smoothing that performs a convolution of the input data with a kernel function. This computes a weighted average of the data around the kernel. Several kernel types can be used for the smoothing (e.g., Gaussian, exponential) and the kernel shape is controlled by a width parameter.
has super-classes
data smoothing c
has sub-classes
apply adaptive kernel smoothing c, apply fixed kernel smoothing c

latent dynamics analysisc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#LatentDynamicsAnalysis

An analysis step that aims to identify underlying patterns and structures within time series or sequential data that are not directly observable. It involves modeling hidden (latent) variables that influence the observed data and their evolution over time. The analysis captures temporal dependencies and dynamics within the data, providing insights into the processes that generate the observed sequences.
is equivalent to
has purpose op value latent dynamics purpose
has super-classes
analysis step c

line noise removalc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#LineNoiseRemoval

A data transformation that removes noise induced by the power line.
has super-classes
data transformation c
has sub-classes
apply Thomson regression noise removal c, apply discrete Fourier transform noise removal c, apply notch filter noise removal c, apply spectrum interpolation noise removal c

mean vector length analysisc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#MeanVectorLengthAnalysis

A phase-amplitude coupling (PAC) analysis that computes the mean vector length (MVL) measure. The MVL is based on a mean vector obtained from a time series defined in the complex plane, where the amplitude is taken from the high-frequency oscillation input and the phase from the low-frequency oscillation input.
has super-classes
phase-amplitude coupling analysis c
has sub-classes
compute mean vector length (Canolty method) c, compute mean vector length (Özkurt method) c

model-based analysisc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ModelBasedAnalysis

An analysis step that depends on assumptions on the interactions between the inputs to perform the computations. For example, the Granger causality analysis assumes linear relationships between the input signals.
is equivalent to
is model-based dp value true
has super-classes
analysis step c

model-free analysisc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#ModelFreeAnalysis

An analysis step that does not depend on assumptions on the interactions between the inputs to perform the computations. For example, it can consider probability distributions obtained from the input data.
is equivalent to
is model-based dp value false
has super-classes
analysis step c

multivariate analysisc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#MultivariateAnalysis

An analysis step that has three or more distinct inputs considered for the computation of the output (e.g., the time series with the local field potential signals recorded from three or more electrodes, used to compute the partial directed coherence).
is equivalent to
is multivariate dp value true
has super-classes
analysis step c
is disjoint with
bivariate analysis c

neuronal activity pattern detection analysisc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#NeuronalActivityPatternDetectionAnalysis

An analysis step that aims to identify a neuronal activity pattern, i.e., spikes of a group of neurons that occur in a specific spatio-temporal configuration.
has super-classes
analysis step c
has sub-classes
SPADE analysis c, Unitary Event analysis c, execute ASSET analysis c, execute Cell Assembly Detection analysis c

neuronal firing regularity analysisc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#NeuronalFiringRegularityAnalysis

An analysis step that computes measures to assess the regularity in the firing of a neuron. Neuronal firing regularity refers to the consistency or variability in the timing of action potentials (spikes) generated by a neuron.
is equivalent to
has purpose op value neuronal firing regularity purpose
has super-classes
analysis step c
has sub-classes
compute spike train Fano factor c

non-directed analysisc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#NonDirectedAnalysis

An analysis step where the output does not provide information on the direction of influence in the relationships among the inputs (e.g., in the Pearson correlation coefficient computed between two spike trains, it is possible to know how strongly they tend to fire together. However, it is not possible to analyze the timing of the spikes of the first input spike with respect to the timing of the spikes of the second input spike train).
is equivalent to
is directed dp value false
has super-classes
analysis step c
is disjoint with
directed analysis c

pairwise Granger causality analysisc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#PairwiseGrangerCausalityAnalysis

A Granger causality (GC) analysis that computes a measure of GC between two inputs.
has super-classes
Granger causality analysis c
has sub-classes
compute time domain pairwise Granger causality c, frequency domain pairwise granger causality analysis c
is disjoint with
conditional Granger causality analysis c

peristimulus time histogram analysisc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#PeristimulusTimeHistogramAnalysis

A spike train time histogram analysis that computes the peristimulus time histogram (PSTH). PSTH is the time histogram of two or more spike trains containing repeated recordings of a single neuron around the time when an event of interest occurred. The event of interest can occur at any time point during the duration of the source spike trains. The distribution of the histogram corresponds to the distribution of the activity of the neuron with respect to the event across the repeated recordings. The event of interest can be an externally presented stimulus or a spontaneous behavioral event. If the histogram represents the neuronal activity after the stimulus presentation or event occurrence, this can be referred as post-stimulus time histogram. Conversely, if the histogram shows the activity of the neuron before the stimulus presentation or event, this can be referred as prestimulus time histogram. Finally, if the histogram considers a behavioral event (or an event that is not an externally presented stimulus), the histogram can be referred to as perievent time histogram (PETH).
has super-classes
spike train time histogram analysis c
has sub-classes
compute peristimulus time histogram (adaptive kernel smoothing) c, compute peristimulus time histogram (fixed kernel smoothing) c, compute peristimulus time histogram (optimal bin size) c, compute peristimulus time histogram (user-selected bin size) c
is disjoint with
compute population histogram c

phase analysisc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#PhaseAnalysis

An analysis step that computes measures related to the phase in the input data. Phase expresses the position of a time-varying signal relative to a fixed reference point in time. For periodic and oscillatory signals (e.g., a sine waveform), phase analysis involves determining the angle on the unit circle that corresponds to the current position within the waveform's cycle. This helps understanding the timing and synchronization of the oscillations in the input data.
has super-classes
analysis step c
has sub-classes
compute angular mean of spike phases c, compute mean phase vector c, compute pairwise phase consistency c, compute phase difference c, compute phase slope index c, compute spike-triggered phase c, phase lag index analysis c, phase locking value analysis c, phase-amplitude coupling analysis c

phase lag index analysisc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#PhaseLagIndexAnalysis

A phase analysis that computes the phase lag index (PLI). The PLI measures the asymmetry of the distribution of the phase differences between two input time series, i.e., if there is an imbalance in the likelihood of the first time series leading or lagging the second time series. It is designed to be invariant to common sources, such as volume conduction and/or active reference electrodes.
has super-classes
phase analysis c
has sub-classes
compute debiased squared weighted phase lag index c, compute directed phase lag index c, compute phase lag index c, compute unbiased squared phase lag index c, compute weighted phase lag index c

phase locking value analysisc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#PhaseLockingValueAnalysis

A phase analysis that computes the phase locking value (PLV). The PLV quantifies the consistency of the phase difference between two input time series across time (e.g., multiple experimental trials).
has super-classes
phase analysis c
has sub-classes
compute corrected imaginary phase locking value c, compute phase locking value c

phase-amplitude coupling analysisc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#PhaseAmplitudeCouplingAnalysis

A phase analysis that computes measures describing how the phase of a low-frequency oscillation modulates the amplitude of a high-frequency oscillation. Phase-amplitude coupling (PAC) can be used to investigate interactions between different frequency bands in the neural activity.
has super-classes
phase analysis c
has sub-classes
compute modulation index c, mean vector length analysis c

power spectral density analysisc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#PowerSpectralDensityAnalysis

A spectral density analysis that computes the power spectral density of an input, i.e., the distribution of power across the different frequency components of the input signal per unit frequency. It is equivalent to the Fourier transform of the autocorrelation function of the input signal. The computed power spectral density values can be corrected depending on the analysis returning the two-sided (i.e., with negative frequencies) or one-sided (i.e., positive frequencies only) PSD. The PSD is often referred to as spectrum.
has super-classes
spectral density analysis c
has sub-classes
compute power spectral density (Bartlett method) c, compute power spectral density (Welch method) c, compute power spectral density (multitaper method) c, compute power spectral density (periodogram method) c

principal component analysisc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#PrincipalComponentAnalysis

A dimensionality reduction that reduces the dimensionality of the input data represented as a matrix with numerous rows and columns. It transforms the data into a set of principal components (PCs) that capture the maximum variance in the input. Each PC is a linear combination of the original variables and serves as a new axis in a lower-dimensional space. The PCs are orthogonal to each other, meaning they capture independent aspects of the input data's variability.
has super-classes
dimensionality reduction c
has sub-classes
apply demixed principal component analysis c, apply probabilistic principal component analysis c, apply standard principal component analysis c

random spike times generationc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#RandomSpikeTimesGeneration

A spike train generation where the output spike trains will have random spike times, taken from a specific probability distribution. The generation process can produce spike trains where the firing rate is constant (stationary) or varies (non-stationary) over time.
has super-classes
spike train generation c
has sub-classes
generate non-stationary Poisson process c, generate non-stationary gamma process c, generate stationary Poisson process c, generate stationary gamma process c, generate stationary inverse Gaussian process c, generate stationary log-normal process c
is disjoint with
correlated spike times generation c

resamplingc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#Resampling

A data transformation that changes the number of samples in the input data.
has super-classes
data transformation c
has sub-classes
apply downsampling c, apply upsampling c

SPADE analysisc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#SPADEAnalysis

A neuronal activity pattern detection analysis that uses the Spatio-temporal PAttern Detection and Evaluation (SPADE) method. The SPADE analysis takes a set of parallel spike trains as input, and returns significant spatio-temporal neuronal activity (spike) patterns. The SPADE method consists of three substeps: 1. Detect all putative patterns in the input data using the frequent item set mining (FIM) algorithm. This step requires the discretization of the input spike train data (binning). The bin size determines the temporal resolution of the analysis. 2. The detected FIM patterns are evaluated for statistical significance, considering the null hypothesis of independence of the spike trains given the modulations by the firing rate. This substep is called Pattern Spectrum Filtering (PSF). For the testing, the patterns are pooled based on their signature: size and occurrence count (non-3d-SPADE) or size, occurrence count and pattern duration (3d-SPADE). The pattern spectrum collects the counts of patterns from each signature. The statistical test is done by a Monte Carlo approach, using spike train surrogates generated from the original data. The final output of this substep is the p-value spectrum, which has the same dimensions as the pattern spectrum. The p-value is computed as the ratio of surrogates containing patterns with that signature to the total number of realizations. 3. Conditional test on the significant patterns to remove patterns arising from the overlap of true pattern spikes and chance spikes (pattern set reduction; PSR).
has super-classes
neuronal activity pattern detection analysis c
has sub-classes
execute 3d-SPADE analysis c, execute non-3d-SPADE analysis c

spectral analysisc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#SpectralAnalysis

An analysis step that computes measures describing the input data with respect to its frequency contents.
has super-classes
analysis step c
has sub-classes
coherence analysis c, coherency analysis c, compute partial directed coherence c, spectral density analysis c, time-frequency analysis c

spectral density analysisc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#SpectralDensityAnalysis

A spectral analysis that computes the density of a measure of the input(s) over the frequency spectrum. Density means that the measure value (e.g., power) for each frequency component is expressed per unit frequency. For example, for an input time series with voltages recorded from an electrode (measured in V), the power for each frequency component of the signal will be in V**2, while the power density will be in V**2/Hz. Therefore, the power values are normalized per unit frequency. This normalization allows for consistent comparisons of results from analyses with different frequency resolutions, as the spectral density remains unaffected by these variations.
has super-classes
spectral analysis c
has sub-classes
cross power spectral density analysis c, power spectral density analysis c

spectrogram analysisc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#SpectrogramAnalysis

A time-frequency analysis that shows the power (or power density) of different frequency components of the input(s) as they change over time. This can be obtained for a single input (spectrogram) or for two distinct inputs (cross-spectrogram).
has super-classes
time-frequency analysis c
has sub-classes
compute cross-spectrogram (short-time Fourier transform method) c, compute spectrogram (Morlet wavelet method) c, compute spectrogram (multitaper method) c, compute spectrogram (short-time Fourier transform method) c

spike train correlation analysisc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#SpikeTrainCorrelationAnalysis

An analysis step that computes measures estimating the correlation between spike train inputs. The correlation value is a normalized measure of covariation in the input spike train data, and reflects the strength and direction of the association: positive values mean that the inputs vary in the same direction, and negative values mean that the inputs vary in opposite directions (e.g., if the activity in one spike train increases, it decreases in the other).
has super-classes
analysis step c
has sub-classes
compute noise correlations c, compute spike time tiling coefficient c, compute spike train Pearson correlation coefficient c

spike train dissimilarity analysisc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#SpikeTrainDissimilarityAnalysis

An analysis step that computes a measure comparing spike train inputs and providing an estimation of their similarity/dissimilarity. This is frequently done by computing spike train distances, which are measures that assign the notion of distance, i.e., the input spike trains are considered as elements in a space and, if similar, will be close together.
has super-classes
analysis step c
has sub-classes
time-scale dependent spike train distance analysis c, time-scale independent spike train distance analysis c

spike train generationc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#SpikeTrainGeneration

A data generation that produces one or more artificial spike trains using distinct statistical procedures to determine the spike times.
has super-classes
data generation c
has sub-classes
correlated spike times generation c, random spike times generation c

spike train surrogate generationc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#SpikeTrainSurrogateGeneration

A data generation that produces one or more spike train surrogates. A spike train surrogate is a new spike train derived from an input spike train (usually experimentally recorded). This is done using methods that alter the original spike times while trying to maintain specific statistical features of the original spike train (e.g., firing rate, interspike interval distribution). This is used to destroy fine temporal correlations in the spiking activity.
has super-classes
data generation c
has sub-classes
generate interspike interval shuffling surrogate c, generate joint interspike interval dithering surrogate c, generate joint interspike interval dithering surrogate c, generate spike time randomization surrogate c, generate spike train dithering surrogate c, generate trial shifting surrogate c, generate trial shuffling surrogate c, generate uniform spike dithering surrogate c, generate uniform spike dithering surrogate with dead time c, generate window shuffling surrogate c

spike train synchrony analysisc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#SpikeTrainSynchronyAnalysis

An analysis step to assess synchronization in two or more spike train inputs that typically represent the activity of different neurons. Spike train synchronization refers to the temporal coordination of action potentials (spikes) between neurons, and describes the degree to which their spikes tend to occur at the same time.
has super-classes
analysis step c
has sub-classes
compute CuBIC analysis c, compute Spike-contrast c, compute complexity distribution c, compute joint peristimulus time histogram c

spike train time histogram analysisc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#SpikeTrainTimeHistogramAnalysis

A firing rate analysis that computes histograms of spike train data over time. The time histogram is obtained by discretizing the duration of the spike train into distinct time intervals (bins), and obtaining the spike count inside each bin. The histogram can show one of three different measures: * the spike count at each bin (across all spike trains); * the mean spike count per bin (spike count in the bin divided by the number of spike trains); * the firing rate (mean spike count in the bin divided by bin width).
has super-classes
firing rate analysis c
has sub-classes
compute population histogram c, compute spike train time histogram c, peristimulus time histogram analysis c

spike waveform analysisc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#SpikeWaveformAnalysis

An analysis step that is used to compute measures to describe or make inferences from spike waveform input data. A spike waveform refers to the shape of an electrical signal produced by a neuron when it fires an action potential.
has super-classes
analysis step c
has sub-classes
compute spike waveform average c, compute spike waveform signal-to-noise ratio c, compute spike waveform variance c, compute spike waveform width c

spike-field coherence analysisc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#SpikeFieldCoherenceAnalysis

An analysis step that computes the spike-field coherence (SFC), which is the coherence computed between an input spike train and an input time series with the local field potential (LFP). Coherence is a measure of the association between the two inputs in the frequency domain. SFC can be used to quantify the relationship between the spiking activity of neurons and the oscillatory activity in the LFP. It represents the similarity of dynamics between the spike train and the voltage fluctuations produced by the neural activity in the local environment where the spiking activity was recorded.
has super-classes
analysis step c
has sub-classes
compute spike-field coherence (Fries method) c, compute spike-field coherence (Welch method) c, compute spike-field coherence (multitaper method) c

spike-field coupling analysisc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#SpikeFieldCouplingAnalysis

An analysis step that computes interactions between the spiking activity of neurons (individual or population) and the local field potential (LFP).
is equivalent to
has purpose op value spike-field coupling purpose
has super-classes
analysis step c
is disjoint with
spike-spike coupling analysis c

spike-spike coupling analysisc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#SpikeSpikeCouplingAnalysis

An analysis step that computes interactions between the spiking activity of one or more neurons.
is equivalent to
has purpose op value spike-spike coupling purpose
has super-classes
analysis step c
is disjoint with
field-field coupling analysis c, spike-field coupling analysis c

statistical analysisc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#StatisticalAnalysis

A generic analysis step that computes measures to summarize and make inferences about the data input. These include measures of central tendency, dispersion and confidence intervals. The subclasses represent analysis steps that are usually used for aggregation of data and description of samples (e.g., compute the mean and standard deviation of the output of trial-by-trial analyses or across subjects). All analysis steps for specific applications related to the analysis of neuroelectrophysiology data itself (e.g., analyzing interspike interval variability) are covered by separate, independent classes.
has super-classes
analysis step c
has sub-classes
central tendency statistical analysis c, confidence interval statistical analysis c, dispersion statistical analysis c

tensor component analysisc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#TensorComponentAnalysis

A dimensionality reduction that reduces the dimensionality of the input data represented as a tensor (i.e., an array with multiple dimensions). The tensor component analysis (TCA) transforms the data into a set of low-dimensional tensors that capture the maximum variance in the input (tensor components).
has super-classes
dimensionality reduction c
has sub-classes
apply canonical polyadic tensor decomposition c, apply coupled canonical polyadic tensor decomposition c, apply non-negative tensor component analysis c

time domain analysisc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#TimeDomainAnalysis

An analysis step that analyzes the input(s) with respect to time.
is equivalent to
is time domain dp value true
has super-classes
analysis step c
is disjoint with
frequency domain analysis c

time-frequency analysisc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#TimeFrequencyAnalysis

A spectral analysis that computes measures describing the frequency content of the input(s) in a time-resolved manner. It allows the analysis of how different frequency components evolve over time, which is essential for non-stationary signals whose spectral characteristics change. The joint time-frequency representation helps in identifying transient features, frequency shifts, and other dynamic behaviors in the input(s).
has super-classes
spectral analysis c
has sub-classes
compute Stockwell transform c, compute short-time Fourier transform c, spectrogram analysis c, wavelet transform analysis c

time-scale dependent spike train distance analysisc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#TimeScaleDependentSpikeTrainDistanceAnalysis

A spike train dissimilarity analysis that computes a spike train distance that depends on a parameter that determines a temporal scale in the spike trains to which the distance is sensitive. By computing the spike train distance for different time scale parameter values, it is possible to make inferences on the time scale that is discriminative in the neural activity.
has super-classes
spike train dissimilarity analysis c
has sub-classes
compute Victor-Purpura distance c, compute van Rossum distance c

time-scale independent spike train distance analysisc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#TimeScaleIndependentSpikeTrainDistanceAnalysis

A spike train dissimilarity analysis that computes a spike train distance that does not depend on a time scale parameter and that are time-scale adaptive. They can be used in scenarios where there are no previous knowledge of the relevant time scales in the input spike trains.
has super-classes
spike train dissimilarity analysis c
has sub-classes
compute ISI-distance c, compute SPIKE distance c, compute SPIKE synchronization c

triggered average analysisc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#TriggeredAverageAnalysis

An analysis step where a signal is averaged to obtain a value around a point in time representing an event of interest (i.e., a trigger). For each event time, a finite duration window of the input time series is selected around the event time. An average for each time point is then obtained across all windows.
has super-classes
analysis step c
has sub-classes
compute event-triggered average c, compute spike-triggered average c

Unitary Event analysisc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#UnitaryEventAnalysis

A neuronal activity pattern detection analysis that uses the Unitary Event (UE) method. UE is a statistical technique focused on identifying synchronous activity among neurons, known as unitary events (UEs), which occur more frequently than an expectation based solely on firing rates. The input can contain spike trains from one or more neurons and one or more trials. For the computation, the input spike train data is discretized into small time intervals (bins), and coincidences across the different spike trains are computed. The significance of the number of observed (empirical) coincidences is determined by comparing to an expected number given the firing rates of the neurons. To account for possible non-stationarities in the firing rates, the method uses a sliding temporal window over the data, whose width is specified as parameter. Therefore, a measure of significant spike synchrony (joint surprise) is obtained for each window. The statistical evaluation can be done using either analytical methods or Monte-Carlo testing with surrogate spike data. The output presents the significant coincidences (UE patterns) and the participant neurons.
has super-classes
neuronal activity pattern detection analysis c
has sub-classes
execute Unitary Event analysis (Monte Carlo method) c, execute Unitary Event analysis (analytical method) c

wavelet transform analysisc back to ToC or Class ToC

IRI: http://purl.org/neao/steps#WaveletTransformAnalysis

A time-frequency analysis that uses wavelets to obtain the time-frequency representation of the time series input. The wavelet is a rapidly decaying oscillation. The wavelet transform breaks the signal into shifted and scaled versions of the wavelet (mother wavelet), and the output contains the information on both the frequencies present in the signal and the time. The wavelets provide good localization in time and frequency, making them suitable to analyze signals with transient features. Different types of wavelets exist, with distinct properties. They can be chosen to tailor the analysis to particular purposes. The output of the wavelet transform is often referred to as scaleogram.
has super-classes
time-frequency analysis c
has sub-classes
compute Morlet wavelet transform c, compute continuous wavelet transform c

Object Properties

has purposeop back to ToC or Object Property ToC

IRI: http://purl.org/neao/steps#hasPurpose

Defines an analysis purpose for the analysis step. This property is used to group analysis steps according to their similarity.
has domain
analysis step c
has range
analysis purpose c

Data Properties

is bivariatedp back to ToC or Data Property ToC

IRI: http://purl.org/neao/steps#isBivariate

Defines if the analysis step uses two data inputs (true).
has super-properties
top data property
has domain
analysis step c
has range
boolean

is directeddp back to ToC or Data Property ToC

IRI: http://purl.org/neao/steps#isDirected

Defines if the analysis step is used to perform an analysis providing information on the direction of the association between the inputs (true) opposed to no direction information (false).
has super-properties
top data property
has domain
analysis step c
has range
boolean

is frequency domaindp back to ToC or Data Property ToC

IRI: http://purl.org/neao/steps#isFrequencyDomain

Defines if the analysis in the step is performed in the frequency domain (true).
has super-properties
top data property
has domain
analysis step c
has range
boolean

is model-baseddp back to ToC or Data Property ToC

IRI: http://purl.org/neao/steps#isModelBased

Defines if the step performs a model-based (true) or model-free analysis (false).
has super-properties
top data property
has domain
analysis step c
has range
boolean

is multivariatedp back to ToC or Data Property ToC

IRI: http://purl.org/neao/steps#isMultivariate

Defines if the analysis step uses three or more data inputs (true).
has super-properties
top data property
has domain
analysis step c
has range
boolean

is time domaindp back to ToC or Data Property ToC

IRI: http://purl.org/neao/steps#isTimeDomain

Defines if the analysis in the step is performed in the time domain (true).
has super-properties
top data property
has domain
analysis step c
has range
boolean

Legend back to ToC

c: Classes
op: Object Properties
dp: Data Properties

Acknowledgments back to ToC

This work was performed as part of the Helmholtz School for Data Science in Life, Earth and Energy (HDS-LEE) and received funding from the Helmholtz Association of German Research Centres. This project has received funding from the European Union’s Horizon 2020 Framework Programme for Research and Innovation under Specific Grant Agreement No. 945539 (Human Brain Project SGA3), the European Union’s Horizon Europe Programme under the Specific Grant Agreement No. 101147319 (EBRAINS 2.0 Project), the Ministry of Culture and Science of the State of North Rhine-Westphalia, Germany (NRW-network "iBehave", grant number: NW21-049), and the Joint Lab "Supercomputing and Modeling for the Human Brain."

The authors would like to thank Silvio Peroni for developing LODE, a Live OWL Documentation Environment, which is used for representing the Cross Referencing Section of this document and Daniel Garijo for developing Widoco, the program used to create the template used in this documentation.