Related Experiment Video
Updated: Mar 16, 2026
13:51
Cross-Modal Multivariate Pattern Analysis
Published on: November 9, 2011
20.6K
CoSMoMVPA: Multi-Modal Multivariate Pattern Analysis of Neuroimaging Data in Matlab/GNU Octave
Nikolaas N Oosterhof1, Andrew C Connolly2, James V Haxby3
1Center for Mind/Brain Sciences, University of Trento Rovereto, Italy.
Frontiers in Neuroinformatics
|August 9, 2016
Summary
CoSMoMVPA is a new toolbox for multivariate pattern (MVP) analysis of neuroimaging data. It supports both functional magnetic resonance imaging (fMRI) and magneto-/electro-encephalography (M/EEG) data, offering advanced analysis techniques and broad compatibility.
Area of Science:
- Neuroimaging
- Computational Neuroscience
- Data Analysis
Background:
- Multivariate pattern (MVP) analysis is increasingly popular for fMRI and M/EEG data.
- Existing toolboxes often lack unified support for diverse neuroimaging modalities and analysis techniques.
Purpose of the Study:
- Introduce CoSMoMVPA, a unified, lightweight toolbox for MVP analysis.
- Facilitate advanced MVPA techniques across fMRI and M/EEG data.
- Enhance accessibility and extensibility for researchers.
Main Methods:
- Developed in Matlab/GNU Octave, supporting fMRI and M/EEG data.
- Implements state-of-the-art MVPA techniques: searchlight, classification, RSA, time generalization.
- Features uniform data representation, cross-format compatibility, and generalized multiple comparison correction.
Main Results:
- CoSMoMVPA provides a uniform interface for diverse MVP measures.
- Supports advanced spatial, temporal, and spectral analyses, including cross-modal and cross-species comparisons.
- Offers modularity, ease of use for beginners, and extensibility for experts.
Conclusions:
- CoSMoMVPA offers a versatile and accessible platform for advanced neuroimaging data analysis.
- Its unified approach and broad compatibility streamline research pipelines.
- The toolbox promotes open-source collaboration and reproducible research in neuroscience.
