EEG-based Bipolar Disorder Deduction using Machine Learning

This review aims to critically assess the different approaches of employing EEG signals and machine learning algorithms for the diagnosis of bipolar disorder. Data acquisition is done by EEG, while signal preprocessing, feature extraction and feature selection, and classification are done by machine learning algorithms. The performance of different machine learning approaches in classifying bipolar disorder was evaluated by reviewing 40 studies published between2018 and2022. More specifically, it analysed the most promising classifiers, namely Support Vector Machine, Random Forest, Logistic Regression, and Naive Bayes, in the successful classification of bipolar disorder with an accuracy ranging from 64.80% to 93%. The study methodology involved the acquisition of EEG signals by the OPENBCI software and implementation in Python. Anticipated results are expected to show statistically significant large amplitude differences in EEG signals across healthy controls and bipolar disorder patients during mania or depression phase, especially in the delta and theta frequency bands. The objective of the study is to develop an early diagnosis of bipolar disorder with EEG using a non-invasive approach.