Comparison of Mass-Univariate, Unimodal and Multivariate Multimodal Analysis Methods for Neurovascular Coupling Analysis
-
1
Technische Universität Berlin, Machine Learning Group, Germany
-
2
MPI Biological Cybernetics, Department Logothetis, Germany
-
3
Technische Universität Berlin, Bernstein Center for Computational Neuroscience, Germany
In the past years multimodal brain imaging methods have yielded valuable insights into how functional
magnetic resonance imaging (fMRI) signals are related to the underlying neural activity. However,
the rapid advances in multimodal imaging technology were not accompanied by the development of
appropriate analysis methods for multimodal data. We present a multimodal analysis framework, temporal kernel Canonical Correlation Analysis (tkCCA) [1,2], and show how it can be used to analyse the spatio-temporal and time-frequency correlation structure between simultaneously measured intracortical neurophysiological recordings and high dimensional fMRI signals. Given the spectrograms of neurophysiological activity and the simultaneously recorded fMRI data we estimate a convolution linking di_erent bands of neural bandpower to an activity pattern of fMRI signals. The convolution can be interpreted as the pattern in time-frequency and time-voxel space that maximises the canonical correlation [3] between neural and haemodynamic activity.
We show results from data recorded in primary visual cortex during spontaneous activity and during visual stimulation. The analysis resulted in robust neurovascular coupling patterns across different experimental conditions. We compared the multivariate patterns with univariate coupling measures and spatial principal component analysis (SPCA) by measuring the accuracy when predicting neural activity from BOLD signals. Our _ndings suggest that the _lters estimated by tkCCA predict neural activity better than univariate methods and unimodal methods such as SPCA.
References
[1] F Bießmann, FC Meinecke, A Gretton, A Rauch, G Rainer, NK Logothetis, and KR Müller, Temporal kernel cca and its application in multimodal neuronal data analysis, Machine Learning Journal, 2009.
[2] Y Murayama, F Bießmann, FC Meinecke, KR Müller, M Augath, A Öltermann, and NK Logothetis, Relationship between neural and hemodynamic signals during spontaneous activity studied with temporal kernel cca, Magnetic Resonance Imaging, 2010.
[3] H Hotelling., Relations between two sets of variates, Biometrika, 1936
Keywords:
computational neuroscience
Conference:
Bernstein Conference on Computational Neuroscience, Berlin, Germany, 27 Sep - 1 Oct, 2010.
Presentation Type:
Presentation
Topic:
Bernstein Conference on Computational Neuroscience
Citation:
Bießmann
F,
Meinecke
FC,
Murayama
Y,
Logothetis
NK and
Mueller
KR
(2010). Comparison of Mass-Univariate, Unimodal and Multivariate Multimodal Analysis Methods for Neurovascular Coupling Analysis.
Front. Comput. Neurosci.
Conference Abstract:
Bernstein Conference on Computational Neuroscience.
doi: 10.3389/conf.fncom.2010.51.00075
Copyright:
The abstracts in this collection have not been subject to any Frontiers peer review or checks, and are not endorsed by Frontiers.
They are made available through the Frontiers publishing platform as a service to conference organizers and presenters.
The copyright in the individual abstracts is owned by the author of each abstract or his/her employer unless otherwise stated.
Each abstract, as well as the collection of abstracts, are published under a Creative Commons CC-BY 4.0 (attribution) licence (https://creativecommons.org/licenses/by/4.0/) and may thus be reproduced, translated, adapted and be the subject of derivative works provided the authors and Frontiers are attributed.
For Frontiers’ terms and conditions please see https://www.frontiersin.org/legal/terms-and-conditions.
Received:
14 Sep 2010;
Published Online:
23 Sep 2010.
*
Correspondence:
Dr. Felix Bießmann, Technische Universität Berlin, Machine Learning Group, Berlin, Germany, felix.biessmann@tu-berlin.de