Globally, machine learning technology benefits the healthcare industry, especially disease diagnosis, to provide a powerful diagnosis. Due to its unifying theme and amalgamation of multiple high dimensional data sources of machine learning over distinctive applications, imparting a different perspective on disease diagnosis derives actionable insights. Machine learning offers principled, automatic, and objective algorithms for high-dimensioned and complicated biomedical data.
In epilepsy disease diagnosis, Machine learning models possess the prospective to impart the early and accurate detection and prediction of epileptic seizures. Compared to conventional statistical methods, machine learning-based epilepsy detection shows its advantages in analyzing Electroencephalogram (EEG) signals with reliable classifiers, certain feature extraction, well-chosen data, and reasonable computation cost.
The computational research studies on machine learning in neurodegenerative diseases are drastically increasing and show progress in early diagnosis, prognosis, and development of novel therapies. Several literature surveys and reviews on epilepsy detection using machine learning were conducted, focusing on machine learning classifiers and features, clinical assessments, EEG signal procession, authoritative datasets, advantages, pitfalls in the current research, and recommended future scopes.
Several machine learning models have been investigated to detect epileptic seizures for epilepsy diagnosis effectively. Various feature extraction and feature selection techniques were also studied to analyze optimal features of seizure patterns.
The application of machine learning in epilepsy detection supports epilepsy diagnosis in inpatient and ambulatory monitoring, wearable devices, pre-surgical planning, prediction of medical responses, early prediction of disease conditions, and surgical management.
Some of the challenges need to be addressed for using machine learning classifiers in epileptic seizure detection, such that appropriate selection and classification of high volume with high dimension data set, precise seizure prediction on long-duration imbalance EEG datasets, rapid seizure recognition using long-hour EEG recording, proper utilization of EEG signal for classification without omitting important EEG channel and knowledge discovery to help neurologist or neurosurgeon for suggesting epilepsy category in epilepsy diagnosis.
As epilepsy diagnosis and treatment revolution, the future scope of epilepsy detection utilizing machine learning concepts instigated deep machine learning strategies for the improved outcome and epilepsy management.