Presentation Information
[23p-1BM-5]Quantum Stochastic Resonance-Based Reservoir Computing System for Epileptic EEG Diagnosis
〇(M2)Xiaoyu Shi1 (1.Tokyo Univ.)
Keywords:
bioengineering,nonlinear physics,reservior computing
Epilepsy is a brain disease characterized by abnormal brain activity causing seizures and sometimes loss of awareness. Thus, developing accurate high-speed seizure detection is the key to save clinician time and improve the management of epilepsy. In order to reach the goals of accuracy and high speed at the same time, I proposed a machine learning method for automatic epileptic EEG diagnosis. The method is based on a system using quantum stochastic resonance-based reservoir computing system (QSRRC). This system innovatively leverages the strong memory and nonlinearity characteristics of QSRRC to memorize and learn from epileptic EEG features. The proposed system was first benchmarked on adult epileptic diagnosis (Siena dataset), which achieved 96.2% accuracy and 96.0% sensitivity, 10.3% higher than traditional reservoir computing system. Our proposed system also outperformed traditional system in all metrices for four states of epileptic EEG signal. To demonstrate the generalizability of the proposed method, I also applied it to the diagnosis of epilepsy EEG in children (CHB-MIT dataset). The results still achieved an accuracy and sensitivity of 94.1%, which far exceeded the traditional RC method.