Genetic algorithm designed for optimization of neural network architectures for intracranial EEG recordings analysis

0577393 - ÚPT 2024 RIV US eng J - Článek v odborném periodiku
Pijáčková, Kristýna - Nejedlý, Petr - Křemen, V. - Plešinger, Filip - Mívalt, F. - Lepková, K. - Pail, Martin - Jurák, Pavel - Worrell, G. A. - Brázdil, M. - Klimeš, Petr
Genetic algorithm designed for optimization of neural network architectures for intracranial EEG recordings analysis.
Journal of Neural Engineering. Roč. 20, č. 3 (2023), č. článku 036034. ISSN 1741-2560. E-ISSN 1741-2552
Grant CEP: GA MZd(CZ) NU22-08-00278; GA ČR(CZ) GA22-28784S; GA MŠMT(CZ) LX22NPO5107
Institucionální podpora: RVO:68081731
Klíčová slova: intracranial EEG * genetic algorithms * optimization * neural network * deep learning
Obor OECD: Neurosciences (including psychophysiology
Impakt faktor: 4, rok: 2022
Způsob publikování: Open access
https://iopscience.iop.org/article/10.1088/1741-2552/acdc54

Objective. The current practices of designing neural networks rely heavily on subjective judgment and heuristic steps, often dictated by the level of expertise possessed by architecture designers. To alleviate these challenges and streamline the design process, we propose an automatic method, a novel approach to enhance the optimization of neural network architectures for processing intracranial electroencephalogram (iEEG) data. Approach. We present a genetic algorithm, which optimizes neural network architecture and signal pre-processing parameters for iEEG classification. Main results. Our method improved the macro F1 score of the state-of-the-art model in two independent datasets, from St. Anne's University Hospital (Brno, Czech Republic) and Mayo Clinic (Rochester, MN, USA), from 0.9076 to 0.9673 and from 0.9222 to 0.9400 respectively. Significance. By incorporating principles of evolutionary optimization, our approach reduces the reliance on human intuition and empirical guesswork in architecture design, thus promoting more efficient and effective neural network models. The proposed method achieved significantly improved results when compared to the state-of-the-art benchmark model (McNemar's test, p MUCH LESS-THAN 0.01). The results indicate that neural network architectures designed through machine-based optimization outperform those crafted using the subjective heuristic approach of a human expert. Furthermore, we show that well-designed data preprocessing significantly affects the models' performance.
Trvalý link: https://hdl.handle.net/11104/0348045