An In-Depth Examination of Machine Learning Approaches for Detecting Epileptic Seizures

Epilepsy stands as an inevitable and significant chronic neurological condition impacting the human brain. It is defining characteristic lies in recurrent, disruptive seizures, characterized by synchronous abnormal innervations within a neuron's population, lasting from seconds to minutes. These seizures manifest as transient episodes of complete or partial unintended body movements, often accompanied by a loss of consciousness. Given the infrequent occurrence of epileptic seizures in patients, their profound effects on physical interactions, social engagements, and emotional well-being necessitate critical consideration in diagnosis and treatment. This survey meticulously examines many research papers, providing a comprehensive analysis of diverse machine learning approaches adopted in each study. The investigation delves into the various features considered in each work, offering insights into the performance achieved. Additionally, the survey scrutinizes the highest performance levels attained across the works, highlighting the datasets employed in each study. Towards the conclusion, the survey addresses existing research deficiencies and associated challenges, offering valuable insights to researchers for advancing future endeavors in the field of convulsive fits detection.