1. <sc>BrainFuseNet</sc>: Enhancing Wearable Seizure Detection Through EEG-PPG-Accelerometer Sensor Fusion and Efficient Edge Deployment
- Author
-
Ingolfsson, Thorir Mar, Wang, Xiaying, Chakraborty, Upasana, Benatti, Simone, Bernini, Adriano, Ducouret, Pauline, Ryvlin, Philippe, Beniczky, Sandor, Benini, Luca, and Cossettini, Andrea
- Abstract
This paper introduces
BrainFuseNet , a novel lightweight seizure detection network based on the sensor fusion of electroencephalography (EEG) with photoplethysmography (PPG) and accelerometer (ACC) signals, tailored for low-channel count wearable systems.BrainFuseNet utilizes the Sensitivity-Specificity Weighted Cross-Entropy (SSWCE), an innovative loss function incorporating sensitivity and specificity, to address the challenge of heavily unbalanced datasets. TheBrainFuseNet -SSWCE approach successfully detects seizure events on the CHB-MIT dataset ($93.5\%$ sample-based sensitivity), for EEG-based classification with only four channels. On the PEDESITE dataset, we demonstrate a sample-based sensitivity and false positive rate of$76.34\%$ and$60.66\%$ FP/h, respectively, when considering EEG data alone. Additionally, we demonstrate that integrating PPG signals increases the sensitivity to$1.18$ (successfully detecting$61.22\%$ seizure events) while decreasing the number of false positives to$92\%$ FP/h. Finally, when ACC data are also considered, the sensitivity increases to$1.0$ (successfully detecting$64.28\%$ seizure events) and the number of false positives drops to only$95\%$ FP/h for sample-based estimations, with less than one false alarm per day when considering event-based estimations.$0.21$ BrainFuseNet is resource-friendly and well-suited for implementation on low-power embedded platforms, and we evaluate its performance on GAP9, a state-of-the-art parallel ultra-low power (PULP) microcontroller for tiny Machine Learning applications on wearables. The implementation on GAP9 achieves an energy efficiency of GMAC/s/W, with an energy consumption per inference of only$21.43$ mJ at high performance ($0.11$ MMAC/s). The$412.54$ BrainFuseNet -SSWCE method demonstrates effective and accurate seizure detection on heavily imbalanced datasets while achieving state-of-the-art performance in the false positive rate and being well-suited for deployment on energy-constrained edge devices.- Published
- 2024
- Full Text
- View/download PDF