Decoding Language from Neural Activity
The ability to correctly discriminate words is crucial for successful social interaction. To comprehend auditory and visual content, the brain must decipher varied types of perceptual information in real time. A long-standing controversy persists in psycholinguistic research regarding the way words and phonemes are coded in human brain.
My work revealed neural mechanisms underlying discrimination of words and phonemes by using machine learning (Naïve-Bayes models, support vector machines) to decode linguistic building blocks from neural responses.
I used a variety of neural information, from the single cell level (single unit recordings) to the network level (fMRI recordings).