-
Distance-based Chatterjee correlation: a new generalized robust measure of directed association for multivariate real and complex-valued data
Authors:
Roberto D. Pascual-Marqui,
Kieko Kochi,
Toshihiko Kinoshita
Abstract:
Building upon the Chatterjee correlation (2021: J. Am. Stat. Assoc. 116, p2009) for two real-valued variables, this study introduces a generalized measure of directed association between two vector variables, real or complex-valued, and of possibly different dimensions. The new measure is denoted as the "distance-based Chatterjee correlation", owing to the use here of the "distance transformed dat…
▽ More
Building upon the Chatterjee correlation (2021: J. Am. Stat. Assoc. 116, p2009) for two real-valued variables, this study introduces a generalized measure of directed association between two vector variables, real or complex-valued, and of possibly different dimensions. The new measure is denoted as the "distance-based Chatterjee correlation", owing to the use here of the "distance transformed data" defined in Szekely et al (2007: Ann. Statist. 35, p2769) for the distance correlation. A main property of the new measure, inherited from the original Chatterjee correlation, is its predictive and asymmetric nature: it measures how well one variable can be predicted by the other, asymmetrically. This allows for inferring the causal direction of the association, by using the method of Blobaum et al (2019: PeerJ Comput. Sci. 1, e169). Since the original Chatterjee correlation is based on ranks, it is not available for complex variables, nor for general multivariate data. The novelty of our work is the extension to multivariate real and complex-valued pairs of vectors, offering a robust measure of directed association in a completely non-parametric setting. Informally, the intuitive assumption used here is that distance correlation is mathematically equivalent to Pearson's correlation when applied to "distance transformed" data. The next logical step is to compute Chatterjee's correlation on the same "distance transformed" data, thereby extending the analysis to multivariate vectors of real and complex valued data. As a bonus, the new measure here is robust to outliers, which is not true for the distance correlation of Szekely et al. Additionally, this approach allows for inference regarding the causal direction of the association between the variables.
△ Less
Submitted 25 June, 2024; v1 submitted 24 June, 2024;
originally announced June 2024.
-
Lagged coherence: explicit and testable definition
Authors:
Roberto D. Pascual-Marqui,
Kieko Kochi,
Toshihiko Kinoshita
Abstract:
Measures of association between cortical regions based on activity signals provide useful information for studying brain functional connectivity. Difficulties occur with signals of electric neuronal activity, where an observed signal is a mixture, i.e. an instantaneous weighted average of the true, unobserved signals from all regions, due to volume conduction and low spatial resolution. This is wh…
▽ More
Measures of association between cortical regions based on activity signals provide useful information for studying brain functional connectivity. Difficulties occur with signals of electric neuronal activity, where an observed signal is a mixture, i.e. an instantaneous weighted average of the true, unobserved signals from all regions, due to volume conduction and low spatial resolution. This is why measures of lagged association are of interest, since at least theoretically, "lagged association" is of physiological origin. In contrast, the actual physiological instantaneous zero-lag association is masked and confounded by the mixing artifact. A minimum requirement for a measure of lagged association is that it must not tend to zero with an increase of strength of true instantaneous physiological association. Such biased measures cannot tell apart if a change in its value is due to a change in lagged or a change in instantaneous association. An explicit testable definition for frequency domain lagged connectivity between two multivariate time series is proposed. It is endowed with two important properties: it is invariant to non-singular linear transformations of each vector time series separately, and it is invariant to instantaneous association. As a first sanity check: in the case of two univariate time series, the new definition leads back to the bivariate lagged coherence of 2007 (eqs 25 and 26 in https://doi.org/10.48550/arXiv.0706.1776). As a second stronger sanity check: in the case of a univariate and multivariate vector time series, the new measure presented here leads back to the original multivariate lagged coherence of 2007 (eq 31 in https://doi.org/10.48550/arXiv.0711.1455), which again trivially includes the bivariate case.
△ Less
Submitted 7 January, 2024; v1 submitted 24 November, 2023;
originally announced November 2023.
-
Cortical Xi-Alpha model for resting state electric neuronal activity
Authors:
Roberto D. Pascual-Marqui,
Kieko Kochi,
Toshihiko Kinoshita
Abstract:
Power spectra of awake resting state EEG recordings in humans typically have an Alpha peak at around 10 Hz riding a decreasing background "Xi process". Normal and pathological variations may have more than one peak or none. The single channel Xi-Alpha model (Pascual-Marqui et al 1988, https://doi.org/10.3109/00207458808985730) separated these two processes, providing a low-dimensional parametric d…
▽ More
Power spectra of awake resting state EEG recordings in humans typically have an Alpha peak at around 10 Hz riding a decreasing background "Xi process". Normal and pathological variations may have more than one peak or none. The single channel Xi-Alpha model (Pascual-Marqui et al 1988, https://doi.org/10.3109/00207458808985730) separated these two processes, providing a low-dimensional parametric description of EEG spectra. Currently lacking is a generative whole cortex model for activity spectra and intracortical functional connectivity. Here we introduce the "cortical Xi-Alpha model". The cross-spectral density matrices are modeled as additive components, each one consisting of a scalar spectrum that multiplies a frequency invariant Hermitian covariance matrix. This simple "separation of variables" form offers a very rich repertoire of spatio-spectral properties, as well as diverse whole cortex functional connectivity patterns. The scalp EEG model conserves the same form, allowing simple estimation from scalp to cortex. Two independent open-access, resting state eyes open and closed EEG data sets (203 participants with 61 electrodes, and 47 participants with 26 electrodes) were used to demonstrate, test, and validate the model. Results summary: - The average dimension of cross-spectra lies between 1.7 and 2.6, indicating two processes. - Non-negative matrix factorization of population power spectra sampled at 6239 cortical voxels with only two components (identified as Xi and Alpha) explains 99% of the variance. - The median value of explained variance was 95% for the "cortical Xi-Alpha model" across all datasets and conditions. - Alpha process generators more occipital, Xi more frontal. - Xi cortical lagged connectivities are isotropic with interdistance. - Laminar recordings suggest layers 2/3 pyramidal neurons generate Xi; layers 5/6 pyramidal neurons generate Alpha.
△ Less
Submitted 29 December, 2022; v1 submitted 27 December, 2022;
originally announced December 2022.
-
On the relation between EEG microstates and cross-spectra
Authors:
Roberto D. Pascual-Marqui,
Kieko Kochi,
Toshihiko Kinoshita
Abstract:
Brain function as measured by multichannel EEG recordings can be described to a high level of accuracy by microstates, characterized as a sequence of time intervals within which the sign invariant normalized scalp electric potential field remains quasi-stable, concatenated by fast transitions. Filtering the EEG has a small effect on the spatial microstate scalp maps, but a large effect on the dyna…
▽ More
Brain function as measured by multichannel EEG recordings can be described to a high level of accuracy by microstates, characterized as a sequence of time intervals within which the sign invariant normalized scalp electric potential field remains quasi-stable, concatenated by fast transitions. Filtering the EEG has a small effect on the spatial microstate scalp maps, but a large effect on the dynamics (e.g. duration, frequency of occurrence, and transition rates). In addition, spectral power has been found to be strongly correlated with microstate dynamics. And yet, the nature of the relation between spectra and microstates remains poorly understood. Here we show that the multivariate EEG cross-spectrum contains sufficient generative information for estimating the microstate scalp maps and their dynamics, demonstrating an underlying fundamental link between the microstate model and the multivariate cross-spectrum. Empirically, based on EEG recordings from 203 participants in eyes-closed resting state, their cross-spectral matrices were computed, from which stochastic EEG was generated. No significant differences were found for the microstate model (maps and dynamics) estimated from the actual EEG and from the stochastic EEG based solely on the cross-spectra. In addition, with the aim of quantifying the spatio-cross-spectral properties of the microstate model, we introduce here the topographic likelihood spectrum, based on the Watson distribution, which provides a frequency-by-frequency account of the contribution of a normalized microstate map to the normalized EEG cross-spectrum, independent of power. The topographic likelihood spectra are distinct for the different microstate maps. In a comparison between eyes-closed and eyes open conditions, they are shown to be significantly different in frequency specific patterns.
△ Less
Submitted 5 August, 2022; v1 submitted 4 August, 2022;
originally announced August 2022.
-
Innovations orthogonalization: a solution to the major pitfalls of EEG/MEG "leakage correction"
Authors:
Roberto D. Pascual-Marqui,
Rolando J. Biscay,
Jorge Bosch-Bayard,
Pascal Faber,
Toshihiko Kinoshita,
Kieko Kochi,
Patricia Milz,
Keiichiro Nishida,
Masafumi Yoshimura
Abstract:
The problem of interest here is the study of brain functional and effective connectivity based on non-invasive EEG-MEG inverse solution time series. These signals generally have low spatial resolution, such that an estimated signal at any one site is an instantaneous linear mixture of the true, actual, unobserved signals across all cortical sites. False connectivity can result from analysis of the…
▽ More
The problem of interest here is the study of brain functional and effective connectivity based on non-invasive EEG-MEG inverse solution time series. These signals generally have low spatial resolution, such that an estimated signal at any one site is an instantaneous linear mixture of the true, actual, unobserved signals across all cortical sites. False connectivity can result from analysis of these low-resolution signals. Recent efforts toward "unmixing" have been developed, under the name of "leakage correction". One recent noteworthy approach is that by Colclough et al (2015 NeuroImage, 117:439-448), which forces the inverse solution signals to have zero cross-correlation at lag zero. One goal is to show that Colclough's method produces false human connectomes under very broad conditions. The second major goal is to develop a new solution, that appropriately "unmixes" the inverse solution signals, based on innovations orthogonalization. The new method first fits a multivariate autoregression to the inverse solution signals, giving the mixed innovations. Second, the mixed innovations are orthogonalized. Third, the mixed and orthogonalized innovations allow the estimation of the "unmixing" matrix, which is then finally used to "unmix" the inverse solution signals. It is shown that under very broad conditions, the new method produces proper human connectomes, even when the signals are not generated by an autoregressive model.
△ Less
Submitted 23 August, 2017; v1 submitted 20 August, 2017;
originally announced August 2017.
-
The cross-frequency mediation mechanism of intracortical information transactions
Authors:
RD Pascual-Marqui,
P Faber,
S Ikeda,
R Ishii,
T Kinoshita,
Y Kitaura,
K Kochi,
P Milz,
K Nishida,
M Yoshimura
Abstract:
In a seminal paper by von Stein and Sarnthein (2000), it was hypothesized that "bottom-up" information processing of "content" elicits local, high frequency (beta-gamma) oscillations, whereas "top-down" processing is "contextual", characterized by large scale integration spanning distant cortical regions, and implemented by slower frequency (theta-alpha) oscillations. This corresponds to a mechani…
▽ More
In a seminal paper by von Stein and Sarnthein (2000), it was hypothesized that "bottom-up" information processing of "content" elicits local, high frequency (beta-gamma) oscillations, whereas "top-down" processing is "contextual", characterized by large scale integration spanning distant cortical regions, and implemented by slower frequency (theta-alpha) oscillations. This corresponds to a mechanism of cortical information transactions, where synchronization of beta-gamma oscillations between distant cortical regions is mediated by widespread theta-alpha oscillations. It is the aim of this paper to express this hypothesis quantitatively, in terms of a model that will allow testing this type of information transaction mechanism. The basic methodology used here corresponds to statistical mediation analysis, originally developed by (Baron and Kenny 1986). We generalize the classical mediator model to the case of multivariate complex-valued data, consisting of the discrete Fourier transform coefficients of signals of electric neuronal activity, at different frequencies, and at different cortical locations. The "mediation effect" is quantified here in a novel way, as the product of "dual frequency RV-coupling coefficients", that were introduced in (Pascual-Marqui et al 2016, http://arxiv.longhoe.net/abs/1603.05343). Relevant statistical procedures are presented for testing the cross-frequency mediation mechanism in general, and in particular for testing the von Stein & Sarnthein hypothesis.
△ Less
Submitted 27 March, 2017; v1 submitted 22 March, 2017;
originally announced March 2017.
-
The dual frequency RV-coupling coefficient: a novel measure for quantifying cross-frequency information transactions in the brain
Authors:
RD Pascual-Marqui,
P Faber,
T Kinoshita,
Y Kitaura,
K Kochi,
P Milz,
K Nishida,
M Yoshimura
Abstract:
Identifying dynamic transactions between brain regions has become increasingly important. Measurements within and across brain structures, demonstrating the occurrence of bursts of beta/gamma oscillations only during one specific phase of each theta/alpha cycle, have motivated the need to advance beyond linear and stationary time series models. Here we offer a novel measure, namely, the "dual freq…
▽ More
Identifying dynamic transactions between brain regions has become increasingly important. Measurements within and across brain structures, demonstrating the occurrence of bursts of beta/gamma oscillations only during one specific phase of each theta/alpha cycle, have motivated the need to advance beyond linear and stationary time series models. Here we offer a novel measure, namely, the "dual frequency RV-coupling coefficient", for assessing different types of frequency-frequency interactions that subserve information flow in the brain. This is a measure of coherence between two complex-valued vectors, consisting of the set of Fourier coefficients for two different frequency bands, within or across two brain regions. RV-coupling is expressed in terms of instantaneous and lagged components. Furthermore, by using normalized Fourier coefficients (unit modulus), phase-type couplings can also be measured. The dual frequency RV-coupling coefficient is based on previous work: the second order bispectrum, i.e. the dual-frequency coherence (Thomson 1982; Haykin & Thomson 1998); the RV-coefficient (Escoufier 1973); Gorrostieta et al (2012); and Pascual-Marqui et al (2011). This paper presents the new measure, and outlines relevant statistical tests. The novel aspects of the "dual frequency RV-coupling coefficient" are: (1) it can be applied to two multivariate time series; (2) the method is not limited to single discrete frequencies, and in addition, the frequency bands are treated by means of appropriate multivariate statistical methodology; (3) the method makes use of a novel generalization of the RV-coefficient for complex-valued multivariate data; (4) real and imaginary covariance contributions to the RV-coherence are obtained, allowing the definition of a "lagged-coupling" measure that is minimally affected by the low spatial resolution of estimated cortical electric neuronal activity.
△ Less
Submitted 17 March, 2016; v1 submitted 16 March, 2016;
originally announced March 2016.
-
The resting microstate networks (RMN): cortical distributions, dynamics, and frequency specific information flow
Authors:
Roberto D Pascual-Marqui,
Dietrich Lehmann,
Pascal Faber,
Patricia Milz,
Kieko Kochi,
Masafumi Yoshimura,
Keiichiro Nishida,
Toshiaki Isotani,
Toshihiko Kinoshita
Abstract:
A brain microstate is characterized by a unique, fixed spatial distribution of electrically active neurons with time varying amplitude. It is hypothesized that a microstate implements a functional/physiological state of the brain during which specific neural computations are performed. Based on this hypothesis, brain electrical activity is modeled as a time sequence of non-overlap** microstates…
▽ More
A brain microstate is characterized by a unique, fixed spatial distribution of electrically active neurons with time varying amplitude. It is hypothesized that a microstate implements a functional/physiological state of the brain during which specific neural computations are performed. Based on this hypothesis, brain electrical activity is modeled as a time sequence of non-overlap** microstates with variable, finite durations (Lehmann and Skrandies 1980, 1984; Lehmann et al 1987). In this study, EEG recordings from 109 participants during eyes closed resting condition are modeled with four microstates. In a first part, a new confirmatory statistics method is introduced for the determination of the cortical distributions of electric neuronal activity that generate each microstate. All microstates have common posterior cingulate generators, while three microstates additionally include activity in the left occipital/parietal, right occipital/parietal, and anterior cingulate cortices. This appears to be a fragmented version of the metabolically (PET/fMRI) computed default mode network (DMN), supporting the notion that these four regions activate sequentially at high time resolution, and that slow metabolic imaging corresponds to a low-pass filtered version. In the second part of this study, the microstate amplitude time series are used as the basis for estimating the strength, directionality, and spectral characteristics (i.e., which oscillations are preferentially transmitted) of the connections that are mediated by the microstate transitions. The results show that the posterior cingulate is an important hub, sending alpha and beta oscillatory information to all other microstate generator regions. Interestingly, beyond alpha, beta oscillations are essential in the maintenance of the brain during resting state.
△ Less
Submitted 14 November, 2014; v1 submitted 7 November, 2014;
originally announced November 2014.
-
Isolated effective coherence (iCoh): causal information flow excluding indirect paths
Authors:
RD Pascual-Marqui,
RJ Biscay,
J Bosch-Bayard,
D Lehmann,
K Kochi,
N Yamada,
T Kinoshita,
N Sadato
Abstract:
A problem of great interest in real world systems, where multiple time series measurements are available, is the estimation of the intra-system causal relations. For instance, electric cortical signals are used for studying functional connectivity between brain areas, their directionality, the direct or indirect nature of the connections, and the spectral characteristics (e.g. which oscillations a…
▽ More
A problem of great interest in real world systems, where multiple time series measurements are available, is the estimation of the intra-system causal relations. For instance, electric cortical signals are used for studying functional connectivity between brain areas, their directionality, the direct or indirect nature of the connections, and the spectral characteristics (e.g. which oscillations are preferentially transmitted). The earliest spectral measure of causality was Akaike's (1968) seminal work on the noise contribution ratio, reflecting direct and indirect connections. Later, a major breakthrough was the partial directed coherence of Baccala and Sameshima (2001) for direct connections. The simple aim of this study consists of two parts: (1) To expose a major problem with the partial directed coherence, where it is shown that it is affected by irrelevant connections to such an extent that it can misrepresent the frequency response, thus defeating the main purpose for which the measure was developed, and (2) To provide a solution to this problem, namely the "isolated effective coherence", which consists of estimating the partial coherence under a multivariate auto-regressive model, followed by setting all irrelevant associations to zero, other than the particular directional association of interest. Simple, realistic, toy examples illustrate the severity of the problem with the partial directed coherence, and the solution achieved by the isolated effective coherence. For the sake of reproducible research, the software code implementing the methods discussed here (using lazarus free-pascal "www.lazarus.freepascal.org"), including the test data as text files, are freely available at: https://sites.google.com/site/pascualmarqui/home/icoh-isolated-effective-coherence
△ Less
Submitted 12 May, 2014; v1 submitted 19 February, 2014;
originally announced February 2014.
-
A measure of association between vectors based on "similarity covariance"
Authors:
Roberto D. Pascual-Marqui,
Dietrich Lehmann,
Kieko Kochi,
Toshihiko Kinoshita,
Naoto Yamada
Abstract:
The "maximum similarity correlation" definition introduced in this study is motivated by the seminal work of Szekely et al on "distance covariance" (Ann. Statist. 2007, 35: 2769-2794; Ann. Appl. Stat. 2009, 3: 1236-1265). Instead of using Euclidean distances "d" as in Szekely et al, we use "similarity", which can be defined as "exp(-d/s)", where the scaling parameter s>0 controls how rapidly the s…
▽ More
The "maximum similarity correlation" definition introduced in this study is motivated by the seminal work of Szekely et al on "distance covariance" (Ann. Statist. 2007, 35: 2769-2794; Ann. Appl. Stat. 2009, 3: 1236-1265). Instead of using Euclidean distances "d" as in Szekely et al, we use "similarity", which can be defined as "exp(-d/s)", where the scaling parameter s>0 controls how rapidly the similarity falls off with distance. Scale parameters are chosen by maximizing the similarity correlation. The motivation for using "similarity" originates in spectral clustering theory (see e.g. Ng et al 2001, Advances in Neural Information Processing Systems 14: 849-856). We show that a particular form of similarity correlation is asymptotically equivalent to distance correlation for large values of the scale parameter. Furthermore, we extend similarity correlation to coherence between complex valued vectors, including its partitioning into real and imaginary contributions. Several toy examples are used for comparing distance and similarity correlations. For instance, points on a noiseless straight line give distance and similarity correlation values equal to 1; but points on a noiseless circle produces near zero distance correlation (dCorr=0.02) while the similarity correlation is distinctly non zero (sCorr=0.36). In distinction to the distance approach, similarity gives more importance to small distances, which emphasizes the local properties of functional relations. This paper represents a preliminary empirical study, showing that the novel similarity association has some distinct practical advantages over distance based association.For the sake of reproducible research, the software code implementing all methods here (using lazarus free-pascal "www.lazarus.freepascal.org"), including all test data, are freely available at: "sites.google.com/site/pascualmarqui/home/similaritycovariance".
△ Less
Submitted 4 February, 2013; v1 submitted 17 January, 2013;
originally announced January 2013.
