Machine Learning Approach for Diagnosis of Schizophrenia Using EEG Signals

Schizophrenia is a serious chronic illness that meddles with the thought process affecting how a person feels and behaves often leading to hallucinations and delusions. Its tedious procedures of tests and complicated analysis makes it quite challenging for the doctors to diagnose and treat. Most popular among these tests are the electroencephalogram (EEG) signal data of the patient capturing the activity of the brain and analysing them. This data is abundant in nature, usually involving correlated data of different parts of the brain. Analysing such convoluted data can be made easier with powerful deep learning and machine learning models through user-friendly visualisation to understand the dependencies among the data to better solve the problem. Our work aims to incorporate various machine learning models to decipher the complex EEG data to identify and learn the patterns of brain activity of schizophrenia and non-schizophrenia. The outcome of our work is an end-to-end model that can be used in hospitals to assist them in their preliminary screening and prediction of a patient having schizophrenia using the EEG components data and the demographics data.