Spatiotemporal analysis of interictal EEG for automated seizure detection and classification. (January 2023)

• Record Type: Journal Article
• Authors: Joshi, Rathin K.; M., Varun Kumar; Agrawal, Megha; Rao, Avinash; Mohan, Latika; Jayachandra, M.; Pandya, Hardik J.
• Is Part Of: Biomedical signal processing and control. Volume 79(2023)Part 1
• Publication Date: 2023-01
• Subjects: Epilepsy screening -- Seizure type classification -- Interictal epileptiform discharge (IED) -- EEG Spatiotemporal Analysis -- Seizure detection; Signal processing -- Periodicals; Biomedical engineering -- Periodicals; Signal Processing, Computer-Assisted -- Periodicals; Image Processing, Computer-Assisted -- Periodicals; Biomedical Engineering -- Periodicals; 610.28
• DOI: 10.1016/j.bspc.2022.104086
• Languages: English
• ISSNs: 1746-8094

Abstract

Highlights: A novel interpretation of interictal EEG for seizure detection and classification. A quick, automated interictal epileptiform discharges extraction method. Quantification of interictal EEG in terms of spikes, sharps, spike-waves. Temporal and spatial spread of neural electrical activities from scalp recorded EEG. A method validated for human subjects with ∼ 91 % accuracy. Abstract: Objective: Seizure type classification is important as therapy differs for different epilepsy subtypes. Currently, skilled neurologists classify seizures based on visual analysis. However, manual EEG inspection is time-consuming, laborious, subjective, and prone to misclassification due to artifacts and EEG variability. This work aims to address these limitations. Methods: In this work, a quick, robust, and accurate spatiotemporal analytical algorithm is developed to classify epileptic seizures. The EEG data set is sampled at 125 Hz using a Nicolet EEG system. Robust preprocessing, feature extraction, and optimal classifiers captured IEDs (Interictal Epileptiform Discharges), including spikes, sharps, slow waves, and Spike-Wave Discharges (SWD). Results: The developed classifier results are validated against clinical impressions provided by experienced epileptologists. The algorithm automatically classifies the EEG data into four types: normal, focal, generalized, and absence, with 93.18 % accuracy (n = 88). Conclusion: The results suggest a novel way to screen epileptic subjects without false positives (accuracy: 94.32 %, n = 88) and tentatively identify the seizure type. Blind validation further confirms the generalizability of the classifier (accuracy: 90.90 %, n = 11). Significance: The developed algorithm reduces the workload of neurologists for epilepsy screening and identifies seizure onset zone, temporal spread, and overall scalp distribution of epileptic activities.