TY - JOUR
T1 - Automatic Detection and Classification of Epileptic Seizures from EEG Data
T2 - Finding Optimal Acquisition Settings and Testing Interpretable Machine Learning Approach
AU - Statsenko, Yauhen
AU - Babushkin, Vladimir
AU - Talako, Tatsiana
AU - Kurbatova, Tetiana
AU - Smetanina, Darya
AU - Simiyu, Gillian Lylian
AU - Habuza, Tetiana
AU - Ismail, Fatima
AU - Almansoori, Taleb M.
AU - Gorkom, Klaus N.V.
AU - Szólics, Miklós
AU - Hassan, Ali
AU - Ljubisavljevic, Milos
N1 - Publisher Copyright:
© 2023 by the authors.
PY - 2023/9
Y1 - 2023/9
N2 - Deep learning (DL) is emerging as a successful technique for automatic detection and differentiation of spontaneous seizures that may otherwise be missed or misclassified. Herein, we propose a system architecture based on top-performing DL models for binary and multigroup classifications with the non-overlapping window technique, which we tested on the TUSZ dataset. The system accurately detects seizure episodes (87.7% Sn, 91.16% Sp) and carefully distinguishes eight seizure types (95–100% Acc). An increase in EEG sampling rate from 50 to 250 Hz boosted model performance: the precision of seizure detection rose by 5%, and seizure differentiation by 7%. A low sampling rate is a reasonable solution for training reliable models with EEG data. Decreasing the number of EEG electrodes from 21 to 8 did not affect seizure detection but worsened seizure differentiation significantly: 98.24 ± 0.17 vs. 85.14 ± 3.14% recall. In detecting epileptic episodes, all electrodes provided equally informative input, but in seizure differentiation, their informative value varied. We improved model explainability with interpretable ML. Activation maximization highlighted the presence of EEG patterns specific to eight seizure types. Cortical projection of epileptic sources depicted differences between generalized and focal seizures. Interpretable ML techniques confirmed that our system recognizes biologically meaningful features as indicators of epileptic activity in EEG.
AB - Deep learning (DL) is emerging as a successful technique for automatic detection and differentiation of spontaneous seizures that may otherwise be missed or misclassified. Herein, we propose a system architecture based on top-performing DL models for binary and multigroup classifications with the non-overlapping window technique, which we tested on the TUSZ dataset. The system accurately detects seizure episodes (87.7% Sn, 91.16% Sp) and carefully distinguishes eight seizure types (95–100% Acc). An increase in EEG sampling rate from 50 to 250 Hz boosted model performance: the precision of seizure detection rose by 5%, and seizure differentiation by 7%. A low sampling rate is a reasonable solution for training reliable models with EEG data. Decreasing the number of EEG electrodes from 21 to 8 did not affect seizure detection but worsened seizure differentiation significantly: 98.24 ± 0.17 vs. 85.14 ± 3.14% recall. In detecting epileptic episodes, all electrodes provided equally informative input, but in seizure differentiation, their informative value varied. We improved model explainability with interpretable ML. Activation maximization highlighted the presence of EEG patterns specific to eight seizure types. Cortical projection of epileptic sources depicted differences between generalized and focal seizures. Interpretable ML techniques confirmed that our system recognizes biologically meaningful features as indicators of epileptic activity in EEG.
KW - EEG
KW - acquisition settings
KW - activation maximization
KW - deep learning
KW - epileptic seizure
KW - interpretable machine learning
KW - source reconstruction
UR - http://www.scopus.com/inward/record.url?scp=85172480214&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85172480214&partnerID=8YFLogxK
U2 - 10.3390/biomedicines11092370
DO - 10.3390/biomedicines11092370
M3 - Article
AN - SCOPUS:85172480214
SN - 2227-9059
VL - 11
JO - Biomedicines
JF - Biomedicines
IS - 9
M1 - 2370
ER -