Md Sultan Mahmud

The paper uses Bayesian nonparametrics and machine learning to analyze neural microstates during speech perception, finding that specific microstates and brain regions predict speech categorization accuracy and perceptual gradiency.

The paper uses machine learning to decode speech categorization from EEG data, finding that speech categories are decoded with high accuracy as early as 120 ms.

The paper shows induced gamma-band oscillations better decode speech categories than evoked beta-band activity, suggesting induced rhythms dominate speech categorization.