Classifying the mental representation of word meaning in children with Multivariate Pattern Analysis of fNIRS

Annu Int Conf IEEE Eng Med Biol Soc. 2018 Jul;2018:295-298. doi: 10.1109/EMBC.2018.8512209.


This study presents the implementation of a within-subject neural decoder, based on Support Vector Machines, and its application for the classification of distributed patterns of hemodynamic activation, measured with Functional Near Infrared Spectroscopy (fNIRS) on children, in response to meaningful and meaningless auditory stimuli. Classification accuracy nominally exceeds chance level for the majority of the participants, but fails to reach statistical significance. Future work should investigate whether individual differences in classification accuracy may relate to other characteristics of the children, such as their cognitive, speech or reading abilities.

Publication types

  • Research Support, Non-U.S. Gov't

MeSH terms

  • Child
  • Hemodynamics
  • Humans
  • Multivariate Analysis
  • Spectroscopy, Near-Infrared*
  • Support Vector Machine