Functional Connectivity Analysis on EEG Signals: A Machine Learning Approach
Information processing in the brain involves coordination of neuronal populations (a synchronous action). Detecting and quantifying the interactions between these neuronal populations can lead to important insights into the dynamic networks that underlie human brain function. We use the concept of synchrony measurement between brain areas as a conceptual idea for finding biomarkers.
To measure synchrony and asynchrony between brain areas we develop a multivariate analysis on EEG signals that is able to isolate the most important cross participant connectivity paths that differentiate between measured groups.
We applied this method both to Dyslexia (wherein the asynchrony between certain areas may be explanatory) and to classifying Neuropathic Pain subjects where finding such biomarker can (i) provide insights into the pathophysiology of painful sensory disorders, and (ii) create an objective measure of pain level.