Categorization system of language speakers by utilizing fMRI data during language comprehension

Abstract

Recently, functional brain imaging techniques such as fMRI have been used to measure brain activities during language comprehension. We plan to apply such fMRI data to assess second language proficiency. For the purpose, first of all, we tested whether we can categorize subjects’ fMRI data during word comprehension into the first and second language speakers by utilizing a machine learning method. By utilizing support vector machine, we could categorize fMRI data into the first and second language speakers in 100% accuracy (Leave-one-out cross validation). This accuracy was a statistically significant (x2 test: p < 0.001). Our result indicates that fMRI data can apply a categorization of language speakers. In response of the results, we expect that we can apply such a categorization system to the categorization or assessment of subjects’ fMRI data into high and low level second language speakers’ data.


Back to Thursday Posters