Temporal kernel CCA and its application in multimodal neuronal data analysis

Felix Biebmann, Frank C. Meinecke, Arthur Gretton, Alexander Rauch, Gregor Rainer, Nikos K. Logothetis, Klaus Muller

Research output: Contribution to journalArticle

61 Citations (Scopus)

Abstract

Data recorded from multiple sources sometimes exhibit non-instantaneous couplings. For simple data sets, cross-correlograms may reveal the coupling dynamics. But when dealing with high-dimensional multivariate data there is no such measure as the crosscorrelogram. We propose a simple algorithm based on Kernel Canonical Correlation Analysis (kCCA) that computes a multivariate temporal filter which links one data modality to another one. The filters can be used to compute a multivariate extension of the crosscorrelogram, the canonical correlogram, between data sources that have different dimensionalities and temporal resolutions. The canonical correlogram reflects the coupling dynamics between the two sources. The temporal filter reveals which features in the data give rise to these couplings and when they do so. We present results from simulations and neuroscientific experiments showing that tkCCA yields easily interpretable temporal filters and correlograms. In the experiments, we simultaneously performed electrode recordings and functional magnetic resonance imaging (fMRI) in primary visual cortex of the non-human primate. While electrode recordings reflect brain activity directly, fMRI provides only an indirect view of neural activity via the Blood Oxygen Level Dependent (BOLD) response. Thus it is crucial for our understanding and the interpretation of fMRI signals in general to relate them to direct measures of neural activity acquired with electrodes. The results computed by tkCCA confirm recent models of the hemodynamic response to neural activity and allow for a more detailed analysis of neurovascular coupling dynamics.

Original languageEnglish
Pages (from-to)5-27
Number of pages23
JournalMachine Learning
Volume79
Issue number1-2
DOIs
Publication statusPublished - 2010 Jan 1
Externally publishedYes

Fingerprint

Electrodes
Hemodynamics
Brain
Blood
Experiments
Oxygen
Magnetic Resonance Imaging
Primates

ASJC Scopus subject areas

  • Artificial Intelligence
  • Software

Cite this

Biebmann, F., Meinecke, F. C., Gretton, A., Rauch, A., Rainer, G., Logothetis, N. K., & Muller, K. (2010). Temporal kernel CCA and its application in multimodal neuronal data analysis. Machine Learning, 79(1-2), 5-27. https://doi.org/10.1007/s10994-009-5153-3

Temporal kernel CCA and its application in multimodal neuronal data analysis. / Biebmann, Felix; Meinecke, Frank C.; Gretton, Arthur; Rauch, Alexander; Rainer, Gregor; Logothetis, Nikos K.; Muller, Klaus.

In: Machine Learning, Vol. 79, No. 1-2, 01.01.2010, p. 5-27.

Research output: Contribution to journalArticle

Biebmann, F, Meinecke, FC, Gretton, A, Rauch, A, Rainer, G, Logothetis, NK & Muller, K 2010, 'Temporal kernel CCA and its application in multimodal neuronal data analysis', Machine Learning, vol. 79, no. 1-2, pp. 5-27. https://doi.org/10.1007/s10994-009-5153-3
Biebmann F, Meinecke FC, Gretton A, Rauch A, Rainer G, Logothetis NK et al. Temporal kernel CCA and its application in multimodal neuronal data analysis. Machine Learning. 2010 Jan 1;79(1-2):5-27. https://doi.org/10.1007/s10994-009-5153-3
Biebmann, Felix ; Meinecke, Frank C. ; Gretton, Arthur ; Rauch, Alexander ; Rainer, Gregor ; Logothetis, Nikos K. ; Muller, Klaus. / Temporal kernel CCA and its application in multimodal neuronal data analysis. In: Machine Learning. 2010 ; Vol. 79, No. 1-2. pp. 5-27.
@article{b0c73ab566c54df4ab965fc2458cce64,
title = "Temporal kernel CCA and its application in multimodal neuronal data analysis",
abstract = "Data recorded from multiple sources sometimes exhibit non-instantaneous couplings. For simple data sets, cross-correlograms may reveal the coupling dynamics. But when dealing with high-dimensional multivariate data there is no such measure as the crosscorrelogram. We propose a simple algorithm based on Kernel Canonical Correlation Analysis (kCCA) that computes a multivariate temporal filter which links one data modality to another one. The filters can be used to compute a multivariate extension of the crosscorrelogram, the canonical correlogram, between data sources that have different dimensionalities and temporal resolutions. The canonical correlogram reflects the coupling dynamics between the two sources. The temporal filter reveals which features in the data give rise to these couplings and when they do so. We present results from simulations and neuroscientific experiments showing that tkCCA yields easily interpretable temporal filters and correlograms. In the experiments, we simultaneously performed electrode recordings and functional magnetic resonance imaging (fMRI) in primary visual cortex of the non-human primate. While electrode recordings reflect brain activity directly, fMRI provides only an indirect view of neural activity via the Blood Oxygen Level Dependent (BOLD) response. Thus it is crucial for our understanding and the interpretation of fMRI signals in general to relate them to direct measures of neural activity acquired with electrodes. The results computed by tkCCA confirm recent models of the hemodynamic response to neural activity and allow for a more detailed analysis of neurovascular coupling dynamics.",
keywords = "Canonical correlation analysis, CCA, KCCA, Neurovascular coupling, TkCCA",
author = "Felix Biebmann and Meinecke, {Frank C.} and Arthur Gretton and Alexander Rauch and Gregor Rainer and Logothetis, {Nikos K.} and Klaus Muller",
year = "2010",
month = "1",
day = "1",
doi = "10.1007/s10994-009-5153-3",
language = "English",
volume = "79",
pages = "5--27",
journal = "Machine Learning",
issn = "0885-6125",
publisher = "Springer Netherlands",
number = "1-2",

}

TY - JOUR

T1 - Temporal kernel CCA and its application in multimodal neuronal data analysis

AU - Biebmann, Felix

AU - Meinecke, Frank C.

AU - Gretton, Arthur

AU - Rauch, Alexander

AU - Rainer, Gregor

AU - Logothetis, Nikos K.

AU - Muller, Klaus

PY - 2010/1/1

Y1 - 2010/1/1

N2 - Data recorded from multiple sources sometimes exhibit non-instantaneous couplings. For simple data sets, cross-correlograms may reveal the coupling dynamics. But when dealing with high-dimensional multivariate data there is no such measure as the crosscorrelogram. We propose a simple algorithm based on Kernel Canonical Correlation Analysis (kCCA) that computes a multivariate temporal filter which links one data modality to another one. The filters can be used to compute a multivariate extension of the crosscorrelogram, the canonical correlogram, between data sources that have different dimensionalities and temporal resolutions. The canonical correlogram reflects the coupling dynamics between the two sources. The temporal filter reveals which features in the data give rise to these couplings and when they do so. We present results from simulations and neuroscientific experiments showing that tkCCA yields easily interpretable temporal filters and correlograms. In the experiments, we simultaneously performed electrode recordings and functional magnetic resonance imaging (fMRI) in primary visual cortex of the non-human primate. While electrode recordings reflect brain activity directly, fMRI provides only an indirect view of neural activity via the Blood Oxygen Level Dependent (BOLD) response. Thus it is crucial for our understanding and the interpretation of fMRI signals in general to relate them to direct measures of neural activity acquired with electrodes. The results computed by tkCCA confirm recent models of the hemodynamic response to neural activity and allow for a more detailed analysis of neurovascular coupling dynamics.

AB - Data recorded from multiple sources sometimes exhibit non-instantaneous couplings. For simple data sets, cross-correlograms may reveal the coupling dynamics. But when dealing with high-dimensional multivariate data there is no such measure as the crosscorrelogram. We propose a simple algorithm based on Kernel Canonical Correlation Analysis (kCCA) that computes a multivariate temporal filter which links one data modality to another one. The filters can be used to compute a multivariate extension of the crosscorrelogram, the canonical correlogram, between data sources that have different dimensionalities and temporal resolutions. The canonical correlogram reflects the coupling dynamics between the two sources. The temporal filter reveals which features in the data give rise to these couplings and when they do so. We present results from simulations and neuroscientific experiments showing that tkCCA yields easily interpretable temporal filters and correlograms. In the experiments, we simultaneously performed electrode recordings and functional magnetic resonance imaging (fMRI) in primary visual cortex of the non-human primate. While electrode recordings reflect brain activity directly, fMRI provides only an indirect view of neural activity via the Blood Oxygen Level Dependent (BOLD) response. Thus it is crucial for our understanding and the interpretation of fMRI signals in general to relate them to direct measures of neural activity acquired with electrodes. The results computed by tkCCA confirm recent models of the hemodynamic response to neural activity and allow for a more detailed analysis of neurovascular coupling dynamics.

KW - Canonical correlation analysis

KW - CCA

KW - KCCA

KW - Neurovascular coupling

KW - TkCCA

UR - http://www.scopus.com/inward/record.url?scp=84897570145&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84897570145&partnerID=8YFLogxK

U2 - 10.1007/s10994-009-5153-3

DO - 10.1007/s10994-009-5153-3

M3 - Article

AN - SCOPUS:84897570145

VL - 79

SP - 5

EP - 27

JO - Machine Learning

JF - Machine Learning

SN - 0885-6125

IS - 1-2

ER -