A Novel Framework Using Brain Computer Interfacing & EEG Microstates To Characterize Cognitive Functionality

Abstract

The rapid advancements in the field of machine learning and artificial intelligence has led to the emergence of technologies like the Brain Computer Interface (BCI), which has revolutionized rehabilitation protocols. However, given the neural basis of BCIs and the dependence of its performance on cognitive factors, BCIs may be used to characterize the functional capacity of the user. A resting state segment can also be considered for characterization of the functional network integrity, creating a two part framework that probes the functional networks and their cognitive manifestations. This thesis explores such a two part framework using a simultaneous EEG-fMRI setup on a healthy population. The BCI accuracies for all subjects increased over the course of the scan and is thought to be due to learning processes on the subject's part. Since such learning processes require cognitive faculties such as attention and working memory, these factors might modulate the BCI performance profile, making it a potential metric for the integrity of such cognitive factors. The resting state analysis identified four EEG Microstates that have been previously found to be associated with verbal, visual, saliency and attention reorientation tasks. The proportion of each microstate that composed the corresponding fMRI resting state networks (RSN) were identified, opening up the potential for predicting fMRI-based RSN information, from EEG microstates alone. The developed protocol can be used to diagnose potential conditions that negatively affect the functional capacity of the user by using the results from this study as healthy control data. This is the first known BCI based system for characterization of the user's functional integrity, opening up the possibility of using BCIs as a metric for diagnosing a neuropathology.