Seizure detection via reservoir computing in MoS₂-based charge trap memory devices.

Area of Science:

• Materials Science
• Neuroscience
• Computer Engineering

Background:

• Neurological disorders pose a significant global health challenge, necessitating advanced diagnostic tools.
• Developing low-power wearable systems for early detection of neurological conditions is crucial but hindered by computational demands of traditional methods.
• Neuromorphic computing offers a brain-inspired, low-power alternative for on-chip processing, with 2D semiconductors like MoS2 showing great potential.

Purpose of the Study:

• To develop and demonstrate a low-power neuromorphic reservoir computing system for processing electrophysiological signals.
• To utilize MoS2-based charge trap memories (CTMs) for efficient on-chip computation in biomedical applications.
• To achieve real-time detection of seizures using the proposed system.

Main Methods:

• Implementation of a neuromorphic reservoir computing system utilizing MoS2-based CTMs.
• Processing of local-field potential (LFP) signals recorded from in vitro rodent models of ictogenesis.
• Exploitation of the nonlinear integration capabilities of MoS2-based CTMs for signal processing.

Main Results:

• Successful real-time seizure detection was achieved using the MoS2-based neuromorphic system.
• The system demonstrated effective processing of electrophysiological signals through nonlinear integration.
• The study validates the potential of MoS2-based CTMs for low-power biomedical devices.

Conclusions:

• MoS2-based CTMs are highly promising for creating low-power neuromorphic computing systems for biomedical applications.
• The developed system shows significant potential for clinical diagnosis and treatment of epilepsy through early seizure detection.
• This research advances the integration of advanced materials and neuromorphic principles for next-generation wearable health monitoring devices.