Automated selection of brain regions for real-time fMRI brain-computer interfaces

Research output: Contribution to journalArticleAcademicpeer-review

Abstract

OBJECTIVE: Brain-computer interfaces (BCIs) implemented with real-time functional magnetic resonance imaging (rt-fMRI) use fMRI time-courses from predefined regions of interest (ROIs). To reach best performances, localizer experiments and on-site expert supervision are required for ROI definition. To automate this step, we developed two unsupervised computational techniques based on the general linear model (GLM) and independent component analysis (ICA) of rt-fMRI data, and compared their performances on a communication BCI. Approach. 3 T fMRI data of six volunteers were re-analyzed in simulated real-time. During a localizer run, participants performed three mental tasks following visual cues. During two communication runs, a letter-spelling display guided the subjects to freely encode letters by performing one of the mental tasks with a specific timing. GLM- and ICA-based procedures were used to decode each letter, respectively using compact ROIs and whole-brain distributed spatio-temporal patterns of fMRI activity, automatically defined from subject-specific or group-level maps.

MAIN RESULTS: Letter-decoding performances were comparable to supervised methods. In combination with a similarity-based criterion, GLM- and ICA-based approaches successfully decoded more than 80% (average) of the letters. Subject-specific maps yielded optimal performances. Significance. Automated solutions for ROI selection may help accelerating the translation of rt-fMRI BCIs from research to clinical applications.

Original languageEnglish
Article number016004
Number of pages14
JournalJournal of neural engineering
Volume14
Issue number1
DOIs
Publication statusPublished - Feb 2017

Keywords

  • ICA
  • GLM
  • BCI
  • rt-fMRI
  • communication BCI
  • neurofeedback
  • ROI selection
  • INDEPENDENT COMPONENT ANALYSIS
  • BOLD HEMODYNAMIC-RESPONSES
  • NEUROFEEDBACK
  • COMMUNICATION
  • VARIABILITY
  • ALIGNMENT
  • CORTEX
  • SERIES
  • TOP

Cite this

@article{2c48c5ef5eb24c9991107a3dc12b7bd6,
title = "Automated selection of brain regions for real-time fMRI brain-computer interfaces",
abstract = "OBJECTIVE: Brain-computer interfaces (BCIs) implemented with real-time functional magnetic resonance imaging (rt-fMRI) use fMRI time-courses from predefined regions of interest (ROIs). To reach best performances, localizer experiments and on-site expert supervision are required for ROI definition. To automate this step, we developed two unsupervised computational techniques based on the general linear model (GLM) and independent component analysis (ICA) of rt-fMRI data, and compared their performances on a communication BCI. Approach. 3 T fMRI data of six volunteers were re-analyzed in simulated real-time. During a localizer run, participants performed three mental tasks following visual cues. During two communication runs, a letter-spelling display guided the subjects to freely encode letters by performing one of the mental tasks with a specific timing. GLM- and ICA-based procedures were used to decode each letter, respectively using compact ROIs and whole-brain distributed spatio-temporal patterns of fMRI activity, automatically defined from subject-specific or group-level maps.MAIN RESULTS: Letter-decoding performances were comparable to supervised methods. In combination with a similarity-based criterion, GLM- and ICA-based approaches successfully decoded more than 80{\%} (average) of the letters. Subject-specific maps yielded optimal performances. Significance. Automated solutions for ROI selection may help accelerating the translation of rt-fMRI BCIs from research to clinical applications.",
keywords = "ICA, GLM, BCI, rt-fMRI, communication BCI, neurofeedback, ROI selection, INDEPENDENT COMPONENT ANALYSIS, BOLD HEMODYNAMIC-RESPONSES, NEUROFEEDBACK, COMMUNICATION, VARIABILITY, ALIGNMENT, CORTEX, SERIES, TOP",
author = "Michael L{\"u}hrs and Bettina Sorger and Rainer Goebel and Fabrizio Esposito",
year = "2017",
month = "2",
doi = "10.1088/1741-2560/14/1/016004",
language = "English",
volume = "14",
journal = "Journal of neural engineering",
issn = "1741-2560",
publisher = "IOP Publishing Ltd.",
number = "1",

}

Automated selection of brain regions for real-time fMRI brain-computer interfaces. / Lührs, Michael; Sorger, Bettina; Goebel, Rainer; Esposito, Fabrizio.

In: Journal of neural engineering, Vol. 14, No. 1, 016004, 02.2017.

Research output: Contribution to journalArticleAcademicpeer-review

TY - JOUR

T1 - Automated selection of brain regions for real-time fMRI brain-computer interfaces

AU - Lührs, Michael

AU - Sorger, Bettina

AU - Goebel, Rainer

AU - Esposito, Fabrizio

PY - 2017/2

Y1 - 2017/2

N2 - OBJECTIVE: Brain-computer interfaces (BCIs) implemented with real-time functional magnetic resonance imaging (rt-fMRI) use fMRI time-courses from predefined regions of interest (ROIs). To reach best performances, localizer experiments and on-site expert supervision are required for ROI definition. To automate this step, we developed two unsupervised computational techniques based on the general linear model (GLM) and independent component analysis (ICA) of rt-fMRI data, and compared their performances on a communication BCI. Approach. 3 T fMRI data of six volunteers were re-analyzed in simulated real-time. During a localizer run, participants performed three mental tasks following visual cues. During two communication runs, a letter-spelling display guided the subjects to freely encode letters by performing one of the mental tasks with a specific timing. GLM- and ICA-based procedures were used to decode each letter, respectively using compact ROIs and whole-brain distributed spatio-temporal patterns of fMRI activity, automatically defined from subject-specific or group-level maps.MAIN RESULTS: Letter-decoding performances were comparable to supervised methods. In combination with a similarity-based criterion, GLM- and ICA-based approaches successfully decoded more than 80% (average) of the letters. Subject-specific maps yielded optimal performances. Significance. Automated solutions for ROI selection may help accelerating the translation of rt-fMRI BCIs from research to clinical applications.

AB - OBJECTIVE: Brain-computer interfaces (BCIs) implemented with real-time functional magnetic resonance imaging (rt-fMRI) use fMRI time-courses from predefined regions of interest (ROIs). To reach best performances, localizer experiments and on-site expert supervision are required for ROI definition. To automate this step, we developed two unsupervised computational techniques based on the general linear model (GLM) and independent component analysis (ICA) of rt-fMRI data, and compared their performances on a communication BCI. Approach. 3 T fMRI data of six volunteers were re-analyzed in simulated real-time. During a localizer run, participants performed three mental tasks following visual cues. During two communication runs, a letter-spelling display guided the subjects to freely encode letters by performing one of the mental tasks with a specific timing. GLM- and ICA-based procedures were used to decode each letter, respectively using compact ROIs and whole-brain distributed spatio-temporal patterns of fMRI activity, automatically defined from subject-specific or group-level maps.MAIN RESULTS: Letter-decoding performances were comparable to supervised methods. In combination with a similarity-based criterion, GLM- and ICA-based approaches successfully decoded more than 80% (average) of the letters. Subject-specific maps yielded optimal performances. Significance. Automated solutions for ROI selection may help accelerating the translation of rt-fMRI BCIs from research to clinical applications.

KW - ICA

KW - GLM

KW - BCI

KW - rt-fMRI

KW - communication BCI

KW - neurofeedback

KW - ROI selection

KW - INDEPENDENT COMPONENT ANALYSIS

KW - BOLD HEMODYNAMIC-RESPONSES

KW - NEUROFEEDBACK

KW - COMMUNICATION

KW - VARIABILITY

KW - ALIGNMENT

KW - CORTEX

KW - SERIES

KW - TOP

U2 - 10.1088/1741-2560/14/1/016004

DO - 10.1088/1741-2560/14/1/016004

M3 - Article

C2 - 27900950

VL - 14

JO - Journal of neural engineering

JF - Journal of neural engineering

SN - 1741-2560

IS - 1

M1 - 016004

ER -