Multiclass fMRI data decoding and visualization using supervised self-organizing maps

Research output: Contribution to journalArticleAcademicpeer-review

5 Citations (Scopus)

Abstract

When multivariate pattern decoding is applied to fMRI studies entailing more than two experimental conditions, a most common approach is to transform the multiclass classification problem into a series of binary problems. Furthermore, for decoding analyses, classification accuracy is often the only outcome reported although the topology of activation patterns in the high-dimensional features space may provide additional insights into underlying brain representations. Here we propose to decode and visualize voxel patterns of fMRI datasets consisting of multiple conditions with a supervised variant of self-organizing maps (SSOMs). Using simulations and real fMRI data, we evaluated the performance of our SSOM-based approach. Specifically, the analysis of simulated fMRI data with varying signal-to-noise and contrast-to-noise ratio suggested that SSOMs perform better than a k-nearest-neighbor classifier for medium and large numbers of features (i.e. 250 to 1000 or more voxels) and similar to support vector machines (SVMs) for small and medium numbers of features (i.e. 100 to 600voxels). However, for a larger number of features (>800voxels), SSOMs performed worse than SVMs. When applied to a challenging 3-class fMRI classification problem with datasets collected to examine the neural representation of three human voices at individual speaker level, the SSOM-based algorithm was able to decode speaker identity from auditory cortical activation patterns. Classification performances were similar between SSOMs and other decoding algorithms; however, the ability to visualize decoding models and underlying data topology of SSOMs promotes a more comprehensive understanding of classification outcomes. We further illustrated this visualization ability of SSOMs with a re-analysis of a dataset examining the representation of visual categories in the ventral visual cortex (Haxby et al., 2001). This analysis showed that SSOMs could retrieve and visualize topography and neighborhood relations of the brain representation of eight visual categories. We conclude that SSOMs are particularly suited for decoding datasets consisting of more than two classes and are optimally combined with approaches that reduce the number of voxels used for classification (e.g. region-of-interest or searchlight approaches).
Original languageEnglish
Pages (from-to)54-66
Number of pages13
JournalNeuroimage
Volume96
Early online date12 Feb 2014
DOIs
Publication statusPublished - 1 Aug 2014

Keywords

  • fMRI
  • Decoding
  • Multiclass classification
  • Self-organizing maps
  • VENTRAL TEMPORAL CORTEX
  • SUPPORT VECTOR MACHINES
  • HUMAN AUDITORY-CORTEX
  • HUMAN VISUAL-CORTEX
  • DISTRIBUTED PATTERNS
  • HUMAN BRAIN
  • DRUG DISCOVERY
  • SINGLE-SUBJECT
  • CLASSIFICATION
  • SELECTION

Cite this

@article{79106edae9e845468f48d14a52fd00db,
title = "Multiclass fMRI data decoding and visualization using supervised self-organizing maps",
abstract = "When multivariate pattern decoding is applied to fMRI studies entailing more than two experimental conditions, a most common approach is to transform the multiclass classification problem into a series of binary problems. Furthermore, for decoding analyses, classification accuracy is often the only outcome reported although the topology of activation patterns in the high-dimensional features space may provide additional insights into underlying brain representations. Here we propose to decode and visualize voxel patterns of fMRI datasets consisting of multiple conditions with a supervised variant of self-organizing maps (SSOMs). Using simulations and real fMRI data, we evaluated the performance of our SSOM-based approach. Specifically, the analysis of simulated fMRI data with varying signal-to-noise and contrast-to-noise ratio suggested that SSOMs perform better than a k-nearest-neighbor classifier for medium and large numbers of features (i.e. 250 to 1000 or more voxels) and similar to support vector machines (SVMs) for small and medium numbers of features (i.e. 100 to 600voxels). However, for a larger number of features (>800voxels), SSOMs performed worse than SVMs. When applied to a challenging 3-class fMRI classification problem with datasets collected to examine the neural representation of three human voices at individual speaker level, the SSOM-based algorithm was able to decode speaker identity from auditory cortical activation patterns. Classification performances were similar between SSOMs and other decoding algorithms; however, the ability to visualize decoding models and underlying data topology of SSOMs promotes a more comprehensive understanding of classification outcomes. We further illustrated this visualization ability of SSOMs with a re-analysis of a dataset examining the representation of visual categories in the ventral visual cortex (Haxby et al., 2001). This analysis showed that SSOMs could retrieve and visualize topography and neighborhood relations of the brain representation of eight visual categories. We conclude that SSOMs are particularly suited for decoding datasets consisting of more than two classes and are optimally combined with approaches that reduce the number of voxels used for classification (e.g. region-of-interest or searchlight approaches).",
keywords = "fMRI, Decoding, Multiclass classification, Self-organizing maps, VENTRAL TEMPORAL CORTEX, SUPPORT VECTOR MACHINES, HUMAN AUDITORY-CORTEX, HUMAN VISUAL-CORTEX, DISTRIBUTED PATTERNS, HUMAN BRAIN, DRUG DISCOVERY, SINGLE-SUBJECT, CLASSIFICATION, SELECTION",
author = "L. Hausfeld and G. Valente and E. Formisano",
year = "2014",
month = "8",
day = "1",
doi = "10.1016/j.neuroimage.2014.02.006",
language = "English",
volume = "96",
pages = "54--66",
journal = "Neuroimage",
issn = "1053-8119",
publisher = "Elsevier Science",

}

Multiclass fMRI data decoding and visualization using supervised self-organizing maps. / Hausfeld, L.; Valente, G.; Formisano, E.

In: Neuroimage, Vol. 96, 01.08.2014, p. 54-66.

Research output: Contribution to journalArticleAcademicpeer-review

TY - JOUR

T1 - Multiclass fMRI data decoding and visualization using supervised self-organizing maps

AU - Hausfeld, L.

AU - Valente, G.

AU - Formisano, E.

PY - 2014/8/1

Y1 - 2014/8/1

N2 - When multivariate pattern decoding is applied to fMRI studies entailing more than two experimental conditions, a most common approach is to transform the multiclass classification problem into a series of binary problems. Furthermore, for decoding analyses, classification accuracy is often the only outcome reported although the topology of activation patterns in the high-dimensional features space may provide additional insights into underlying brain representations. Here we propose to decode and visualize voxel patterns of fMRI datasets consisting of multiple conditions with a supervised variant of self-organizing maps (SSOMs). Using simulations and real fMRI data, we evaluated the performance of our SSOM-based approach. Specifically, the analysis of simulated fMRI data with varying signal-to-noise and contrast-to-noise ratio suggested that SSOMs perform better than a k-nearest-neighbor classifier for medium and large numbers of features (i.e. 250 to 1000 or more voxels) and similar to support vector machines (SVMs) for small and medium numbers of features (i.e. 100 to 600voxels). However, for a larger number of features (>800voxels), SSOMs performed worse than SVMs. When applied to a challenging 3-class fMRI classification problem with datasets collected to examine the neural representation of three human voices at individual speaker level, the SSOM-based algorithm was able to decode speaker identity from auditory cortical activation patterns. Classification performances were similar between SSOMs and other decoding algorithms; however, the ability to visualize decoding models and underlying data topology of SSOMs promotes a more comprehensive understanding of classification outcomes. We further illustrated this visualization ability of SSOMs with a re-analysis of a dataset examining the representation of visual categories in the ventral visual cortex (Haxby et al., 2001). This analysis showed that SSOMs could retrieve and visualize topography and neighborhood relations of the brain representation of eight visual categories. We conclude that SSOMs are particularly suited for decoding datasets consisting of more than two classes and are optimally combined with approaches that reduce the number of voxels used for classification (e.g. region-of-interest or searchlight approaches).

AB - When multivariate pattern decoding is applied to fMRI studies entailing more than two experimental conditions, a most common approach is to transform the multiclass classification problem into a series of binary problems. Furthermore, for decoding analyses, classification accuracy is often the only outcome reported although the topology of activation patterns in the high-dimensional features space may provide additional insights into underlying brain representations. Here we propose to decode and visualize voxel patterns of fMRI datasets consisting of multiple conditions with a supervised variant of self-organizing maps (SSOMs). Using simulations and real fMRI data, we evaluated the performance of our SSOM-based approach. Specifically, the analysis of simulated fMRI data with varying signal-to-noise and contrast-to-noise ratio suggested that SSOMs perform better than a k-nearest-neighbor classifier for medium and large numbers of features (i.e. 250 to 1000 or more voxels) and similar to support vector machines (SVMs) for small and medium numbers of features (i.e. 100 to 600voxels). However, for a larger number of features (>800voxels), SSOMs performed worse than SVMs. When applied to a challenging 3-class fMRI classification problem with datasets collected to examine the neural representation of three human voices at individual speaker level, the SSOM-based algorithm was able to decode speaker identity from auditory cortical activation patterns. Classification performances were similar between SSOMs and other decoding algorithms; however, the ability to visualize decoding models and underlying data topology of SSOMs promotes a more comprehensive understanding of classification outcomes. We further illustrated this visualization ability of SSOMs with a re-analysis of a dataset examining the representation of visual categories in the ventral visual cortex (Haxby et al., 2001). This analysis showed that SSOMs could retrieve and visualize topography and neighborhood relations of the brain representation of eight visual categories. We conclude that SSOMs are particularly suited for decoding datasets consisting of more than two classes and are optimally combined with approaches that reduce the number of voxels used for classification (e.g. region-of-interest or searchlight approaches).

KW - fMRI

KW - Decoding

KW - Multiclass classification

KW - Self-organizing maps

KW - VENTRAL TEMPORAL CORTEX

KW - SUPPORT VECTOR MACHINES

KW - HUMAN AUDITORY-CORTEX

KW - HUMAN VISUAL-CORTEX

KW - DISTRIBUTED PATTERNS

KW - HUMAN BRAIN

KW - DRUG DISCOVERY

KW - SINGLE-SUBJECT

KW - CLASSIFICATION

KW - SELECTION

U2 - 10.1016/j.neuroimage.2014.02.006

DO - 10.1016/j.neuroimage.2014.02.006

M3 - Article

VL - 96

SP - 54

EP - 66

JO - Neuroimage

JF - Neuroimage

SN - 1053-8119

ER -