Seizure Detection using Machine Learning and Monitoring through IoT Devices

Rapid and erratic electrical firing in the brain is what causes seizures. This may result in momentary anomalous behaviors, movements (such as jerky, alternately stiffened arms and legs), sensations, or unconsciousness or an altered degree of consciousness. According to a survey, approximately 50 million people around the world suffer from epilepsy, making it one of the most common neurological diseases on a global scale. Despite modern pharmacological and surgical treatment options, more than a quarter of epilepsy patients experience uncontrollable seizures. Early detection of seizures aids in the prevention of SUDEP (Sudden Unexpected Death in Epilepsy) This model proposed a machine learning approach to detect the seizure. The Random Forest Classifier algorithm was used to diagnose the seizure that was brought to our attention. A pretrained model with a machine learning and supervised classification approach is built with the datasets received from the EEG (electroencephalogram) and ECG (electrocardiogram) to detect seizures in the pre-ictal stage. The model has been trained with existing datasets, and the research shows that there are several causes of epilepsy. Even if the causes differ, the solution is the same. In this case, the model predicts the seizure and alerts the hospital management, nurses, and carers before it occurs. Which would be extremely beneficial to doctors in treating it.

Paper

Full text

PDF

Seizure Detection using Machine Learning and Monitoring through IoT Devices

Semantic Scholar · Computer Science · 2023

Abstract

Rapid and erratic electrical firing in the brain is what causes seizures. This may result in momentary anomalous behaviors, movements (such as jerky, alternately stiffened arms and legs), sensations, or unconsciousness or an altered degree of consciousness. According to a survey, approximately 50 million people around the world suffer from epilepsy, making it one of the most common neurological diseases on a global scale. Despite modern pharmacological and surgical treatment options, more than a quarter of epilepsy patients experience uncontrollable seizures. Early detection of seizures aids in the prevention of SUDEP (Sudden Unexpected Death in Epilepsy) This model proposed a machine learning approach to detect the seizure. The Random Forest Classifier algorithm was used to diagnose the seizure that was brought to our attention. A pretrained model with a machine learning and supervised classification approach is built with the datasets received from the EEG (electroencephalogram) and ECG (electrocardiogram) to detect seizures in the pre-ictal stage. The model has been trained with existing datasets, and the research shows that there are several causes of epilepsy. Even if the causes differ, the solution is the same. In this case, the model predicts the seizure and alerts the hospital management, nurses, and carers before it occurs. Which would be extremely beneficial to doctors in treating it.

Similar papers

© 2026 NYSGPT2525 LLC