OBJECTIVE: To report an innovative spike detection algorithm that tailors its detection to the patient. Interictal epileptiform activity quantification was accomplished in the setting of epileptic syndromes with continuous spike and waves during slow sleep, which is a time-consuming task for the EEG analysis. METHODS: The algorithm works in three steps. Firstly, a first spike detection is made with generic parameters. Secondly, the detected spikes are used to tailor the detection algorithm to the patient; and thirdly, the resulting patient-specific detection algorithm is used to analyze individual patient with high-quality detection. Therefore, the algorithm produces a patient-specific template -hence exhibiting improved performance metrics, without the need of a priori knowledge from the experts. RESULTS: The system was first evaluated for EEG of three patients, against the scoring of three EEG experts, demonstrating similar performance. Later, it was evaluated against the spike and wave percentage evaluation of another expert for 17 additional records. The difference between the two evaluations was 4.4% on average, which is almost the same as the interexpert difference (4.7%). CONCLUSIONS: We designed a fully automated and efficient spike detection algorithm, which is liable to trim down the specialist's diagnostic time.
OBJECTIVE: To report an innovative spike detection algorithm that tailors its detection to the patient. Interictal epileptiform activity quantification was accomplished in the setting of epileptic syndromes with continuous spike and waves during slow sleep, which is a time-consuming task for the EEG analysis. METHODS: The algorithm works in three steps. Firstly, a first spike detection is made with generic parameters. Secondly, the detected spikes are used to tailor the detection algorithm to the patient; and thirdly, the resulting patient-specific detection algorithm is used to analyze individual patient with high-quality detection. Therefore, the algorithm produces a patient-specific template -hence exhibiting improved performance metrics, without the need of a priori knowledge from the experts. RESULTS: The system was first evaluated for EEG of three patients, against the scoring of three EEG experts, demonstrating similar performance. Later, it was evaluated against the spike and wave percentage evaluation of another expert for 17 additional records. The difference between the two evaluations was 4.4% on average, which is almost the same as the interexpert difference (4.7%). CONCLUSIONS: We designed a fully automated and efficient spike detection algorithm, which is liable to trim down the specialist's diagnostic time.
Authors: Jin Jing; Aline Herlopian; Ioannis Karakis; Marcus Ng; Jonathan J Halford; Alice Lam; Douglas Maus; Fonda Chan; Marjan Dolatshahi; Carlos F Muniz; Catherine Chu; Valeria Sacca; Jay Pathmanathan; WenDong Ge; Haoqi Sun; Justin Dauwels; Andrew J Cole; Daniel B Hoch; Sydney S Cash; M Brandon Westover Journal: JAMA Neurol Date: 2020-01-01 Impact factor: 18.302
Authors: Peter C Horak; Stephen Meisenhelter; Yinchen Song; Markus E Testorf; Michael J Kahana; Weston D Viles; Krzysztof A Bujarski; Andrew C Connolly; Ashlee A Robbins; Michael R Sperling; Ashwini D Sharan; Gregory A Worrell; Laura R Miller; Robert E Gross; Kathryn A Davis; David W Roberts; Bradley Lega; Sameer A Sheth; Kareem A Zaghloul; Joel M Stein; Sandhitsu R Das; Daniel S Rizzuto; Barbara C Jobst Journal: Epilepsia Date: 2016-12-09 Impact factor: 5.864