16 October 2024 in IEEE Transactions on Biomedical Circuits and Systems
Personalized brain implants have the potential to revolutionize the treatment of neurological disorders and augment cognition. Medical implants that deliver therapeutic stimulation in response to detected seizures have already been deployed for the treatment of epilepsy. These devices require low-power integrated circuits for life-long operation. This constraint impedes the integration of machine-learning driven classifiers that could improve treatment outcomes. This paper introduces BrainForest, a neuromorphic multiplier-less bit-serial weight-memory-optimized brain-state classification processor. The architecture achieves state-of-the-art energy efficiency using two layers of neuron models to implement the spectral and temporal functions needed for classification: 1) resonate-and-fire neurons are used to extract physiological signal band energy EEG biomarkers 2) leaky integrator neurons are used to build multi-timescale representations for classification. Sparse neural model firing activity is used to clock-gate device logic, thereby decreasing power consumption by 93%. An energy-optimized 1024-tree boosted decision forest performs the classification used to trigger stimulation in response to detected pathological brain states. The IC is implemented in 65nm CMOS with state-of-the-art power consumption (best case: 9.6μW, typical: 118μW), achieving a seizure sensitivity of 97.5% with a false detection rate of 2.08 per hour.
29 October 2025 in IEEE Transactions on Systems, Man, and Cybernetics: Systems
Heart rate (HR) has been shown to correlate with mental workload in operational and instructional settings, supporting its use as an objective, continuous, and unobtrusive workload index. With the growing use of virtual reality (VR) in training and education, we evaluate whether the HR–workload relationship holds in VR simulation. To this end, we analyze HR responses derived from electrocardiogram (ECG) recordings in 20 participants performing VR flight maneuvers and assess the relationship between HR, mental workload, and task performance. Our analysis reveals that mean HR increases by 2% ( $p \lt 0.05$ ) across task difficulty levels, decreases by 5% ( $p \lt 0.01$ ) across repeated runs of the same difficulty level, and increases by 3% ( $p \lt 0.01$ ) near touchdown versus the start of the flight. We find that HR correlates positively with mental workload and negatively with task performance and that HR provides insights into mental workload beyond what performance alone reveals. We also find that HR in our VR task follows expected mental workload-related trends consistent with observations from other environments (real aircraft and legacy flight simulators), and that photoplethysmography (PPG)-derived HR closely tracked ECG patterns, supporting unobtrusive monitoring in VR. Our findings suggest that HR is a readily measurable, meaningful signal for tracking mental workload in VR and could enable real-time, closed-loop personalization of training difficulty to maintain optimal engagement.