Improved patient specific seizure detection during pre-surgical evaluation

• Published in: Clinical neurophysiology Vol. 122; no. 4; pp. 672 - 679
• Main Authors: Chua, Eric C.-P., Patel, Kunjan, Fitzsimons, Mary, Bleakley, Chris J.
• Format: Journal Article
• Language: English
• Published: Oxford Elsevier Ireland Ltd 01.04.2011; Elsevier
• Subjects: Adaptation; Adult; Algorithms; Artificial Intelligence; Biological and medical sciences; Data Interpretation, Statistical; Electrodiagnosis. Electric activity recording; Electroencephalography - methods; Electroencephalography - statistics & numerical data; Epilepsy, Tonic-Clonic - classification; Epilepsy, Tonic-Clonic - diagnosis; Epilepsy, Tonic-Clonic - physiopathology; False Positive Reactions; Female; Headache. Facial pains. Syncopes. Epilepsia. Intracranial hypertension. Brain oedema. Cerebral palsy; Humans; Interactive machine learning; Investigative techniques, diagnostic techniques (general aspects); Male; Medical sciences; Middle Aged; Nervous system; Nervous system (semeiology, syndromes); Neurology; Normal Distribution; Preoperative Care; Reproducibility of Results; ROC Curve; Seizure detection; Seizures - classification; Seizures - diagnosis; Subject-specific; Wavelet Analysis; Seizure detection; Subject-specific; Interactive machine learning; Adaptation; User; Threshold detection; Epilepsy; Electrophysiology; Electroencephalography; Acquisition process; Stereoelectroencephalography; Neurological disorder; Monitoring; Human; Evaluation; Nervous system diseases; Interaction; Probability; Algorithm; Tuning; Cerebral disorder; Sensitivity; Convulsion; Central nervous system disease; Detection threshold
• ISSN: 1388-2457; 1872-8952; 1872-8952
• DOI: 10.1016/j.clinph.2010.10.002

► We present a method for adapting a subject-independent seizure detection system to subject-specific ones using feedback from the EEG technologist. This improves seizure detection performance, and unlike traditional subject-specific systems, it does not require obtaining a priori data to train the system. ► The method was tested on 529 h of intracranial EEG containing 63 seizures from 15 subjects. Compared to the standard Gotman algorithm, the subject-specific scheme improved sensitivity from 53% to 78%, and decreased false positive rate from 0.23/h to 0.18/h. This is close to the performance of the subject-specific standard Gotman algorithm. ► The proposed method could potentially increase productivity of offline EEG ana1ysis. There is considerable interest in improved off-line automated seizure detection methods that will decrease the workload of EEG monitoring units. Subject-specific approaches have been demonstrated to perform better than subject-independent ones. However, for pre-surgical diagnostics, the traditional method of obtaining a priori data to train subject-specific classifiers is not practical. We present an alternative method that works by adapting the threshold of a subject-independent to a specific subject based on feedback from the user. A subject-independent quadratic discriminant classifier incorporating modified features based partially on the Gotman algorithm was first built. It was then used to derive subject-specific classifiers by determining subject-specific posterior probability thresholds via user interaction. The two schemes were tested on 529 h of intracranial EEG containing 63 seizures from 15 subjects undergoing pre-surgical evaluation. To provide comparison, the standard Gotman algorithm was implemented and optimised for this dataset by tuning the detection thresholds. Compared to the tuned Gotman algorithm, the subject-independent scheme reduced the false positive rate by 51% (0.23 to 0.11 h −1) while increasing sensitivity from 53% to 62%. The subject-specific scheme further improved sensitivity to 78%, but with a small increase in false positive rate to 0.18 h −1. The results suggest that a subject-independent classifier scheme with modified features is useful for reducing false positive rate, while subject adaptation further enhances performance by improving sensitivity. The results also suggest that the proposed subject-adapted classifier scheme approximates the performance of the subject-specific Gotman algorithm. The proposed method could potentially increase the productivity of offline EEG analysis. The approach could also be generalised to enhance the performance of other subject independent algorithms.