Development of Real-Time Neurofeedback System

Transcranial Current stimulation (TCS) has been projected as one of the most potential methods to treat neurological disorders where a small amount of current is injected at the scalp to activate specific neuronal populations. Despite its some reported success recent literatures showed that there is no definitive advantage of the method in treating neurological disorders. However, it is also concluded that the main reason behind this reason is the lack of theoretical models for its application - current practice of TCS mainly depends upon trial-and-error method - e.g., which type of current (AC or DC) is most effective for a type of particular neurological disorder, what is the optimal magnitude frequency of the stimulation current and, what is the spatial distribution of the current required to activate a neuronal population in a focussed way. This project is intended to first answer these questions and then implement it as a real-time system. The approach to solve the problem is to do large scale electromagnetic simulation of current propagation through different layers of head (considering the current is injected at the scalp) for generating realistic data that could be used in a machine learning framework to develop appropriate stimulation model in terms of nature of stimulation current and its spatial distribution.
The specific aim of the project is: to develop machine learning based real-time TCS stimulation system that could be used for focussed stimulation of specific neuronal populations. This aim will be achieved through the following specific objectives:

1. Generate data from first principle electromagnetic wave propagation simulation describing stimulation current propagation through the different layers of head and estimating the resulting electric field at a particular neural mass location for DC and AC current (with varying frequency);
2. Generate data from first principle electromagnetic wave propagation simulation describing stimulation current propagation through the different layers of head and the resulting electric field at a particular neural mass location for different spatial distribution of stimulation electrodes over the entire head;
3. Use these dataset under machine learning framework (using different machine learning approaches) to develop optimal TCS model in terms of the nature of stimulation current and spatial distribution of the electrodes over the head;
4. Verification of the model under a simulation environment;
5. Evaluating the real-time operation of the system by implementing the stimulation model on an FPGA platform.