ID: 217492 Applications of Machine Learning on sEEG Recording to Evaluate Human Numerical Cognition

Abstract

The current neuroanatomical understanding of human numerical cognition is based primarily on noninvasive neuroimaging studies with limited contributions from intracranial EEG analyses[1]. We previously used a linear support vector machine (SVM) to accurately classify stereo-electroencephalography (sEEG) spectrograms and identify novel oscillatory patterns during keypress movements[2]. We hypothesize that this method can be used to analyze other cognitive domains such as numerical cognition.