Seizures Prediction using Heart Rate Variability in ECG Signal

An exploratory study was suggested to assess the efficacy of an automatic model, and this model helps to identify heart rate (HR) changes in individuals with severe epilepsy during seizures. Heart rate (HR) is one of the automatic body functions that can alter as a result of epileptic seizures. Mainly, the method proposed in the study was used to check disruptions in the heart signal, called movement artifacts, on the ECG. These disruptions in the heart signal make it harder to detect R-peaks, which are the main spikes in the ECG, and can also reduce the accuracy of seizure detection algorithms as well. Importantly, the Artefacts Detection Algorithm introduced a new feature, which uses Principal Component Analysis (PCA) to analyze electrocardiogram (ECG) signals and can be used to describe the changes within ECG data as well. The new feature is highly focused to detect seizures by analyzing changes in the morphological characteristics (shape and structure) of the electrocardiogram signal, especially those in the QRS complex, which can indicate seizure activity. One type of machine learning model used to distinguish between normal and abnormal ECG signals is the Support Vector Machine (SVM). The clustering algorithm ‘K-means’ is used to identify seizures. The normal and abnormal signals can be easily differentiated by the system. SVM, along with k-means algorithms are used to detect seizures in a more accurate manner.