Analysis of Imagined Speech Characteristics using Phase-based Connectivity Measures

Abstract

The Aim of this research is to access connectivity in cognitive functioning of brain networks during Imagination of speech prompts from electroencephalography(EEG) signals. Phase based connectivity analysis is performed to identify dominant neurophysiological dynamics of speech imagery paradigm. The connectivity analysis is performed on Publicly available EEG based Imagined speech ’Kara one’ dataset. The Connectivity matrices are generated using Interstice phase clustering(ISPC) over trials and Phase lag index(PLI) to quantify strength of connectivity between different regions of brains at different time and frequency points. For qualitative inspection, the thresholding of connectivity metrices is performed using a threshold of one standard deviation above the median connectivity. Topological Maps and time frequency plots are analysed for both ISPC-trials and PLI connectivity matrix to trace connectivity between neural networks at various time-frequency points.