EEG-Based Cognitive Disorder Classification Using Frequency Band Features and Hybrid ML - CBR Approach

Millions of people worldwide are impacted by mental health illnesses, which have a major negative influence on both economic productivity and quality of life. Nearly one billion people worldwide suffer from mental illnesses, with anxiety and depression being the main causes of impairment, according to the WHO. An estimated 970 million individuals suffered from a mental health illness in 2019. The COVID-19 pandemic made matters worse by raising the prevalence of major depressive disorder by 28% and anxiety disorders by 26%, respectively. Younger people and women were disproportionately affected, with suicide being the second most common cause of death for those aged 15 to 29. Early diagnosis is essential for neurological disorders like Alzheimer's disease, epilepsy, and cognitive impairments since they present growing healthcare costs. A non-invasive neuroimaging method called electroencephalography (EEG) offers important information about cognitive health. Conventional approaches depend on professional interpretation, which can be laborious and subjective. Using Case-Based Reasoning (CBR) to classify cognitive illnesses, this research maps signals into the Alpha, Beta, Gamma Delta and Theta frequency bands to identify patterns linked to severe mental illness, Alzheimer's disease, epilepsy, and sleep disturbances. In order to improve diagnoses iteratively, CBR retrieves previous cases and compares them with fresh examples. Using machine learning reduces errors, improves accuracy, and aids in clinical decision-making. Early findings show that CBR can be used to identify cognitive problems early on, providing doctors with a scalable and effective tool. The cognitive disorders are classified using various machine learning models. Logistic regression shows 93% accuracy and CBR identifies case with 68% accuracy. Results analysis predicts the need to add more cases for analysis to increase accuracy.