A Neuromorphic System Based on Spiking-Timing Dependent Plasticity for
Evaluating Wakefulness and Anesthesia States using Intracranial EEG
Signals
Abstract
The primary method for assessing the depth of anesthesia during clinical
surgeries currently relies on physiological and behavioral cues, such as
heart rate, blood pressure, and reaction to external stimuli. These
measures might not be dependable because their relationship to patients’
levels of awareness is not firmly established. Here, we present a
neuromorphic framework comprising a Spiking Neural Network (SNN) with
simulated neurons to access conscious states during general anesthesia,
such as wakefulness and anesthesia, from intracranial
electroencephalography (iEEG) and electroencephalogram (EEG) signals.
This framework adjusts synapse weight of each neuron utilizing
Spiking-Timing Dependent Plasticity (STDP) rules. Our analysis revealed
that the proposed neuromorphic approach can access states of
consciousness during anesthesia and characterize the transition from
wakefulness to anesthetic-induced unconsciousness. We further
implemented this framework on the Field Programmable Gate Array platform
to enhance execution efficiency and address the clinical requirements
for real-time monitoring of anesthesia states. Importantly, this study
signifies an initial viable investigation using neuromorphic computing
techniques for evaluating patients’ levels of anesthesia based on iEEG
signals.