An Energy-Efficient Spiking Neural Network for Finger Velocity Decoding for Implantable Brain-Machine Interface
This work addresses the need for low-power algorithms in implantable brain-machine interfaces for motor rehabilitation and mobility augmentation, representing an incremental improvement in efficiency.
The paper tackled the problem of high-power consumption in implantable brain-machine interfaces by proposing a spiking neural network decoder for finger velocity decoding, achieving the same correlation coefficient as state-of-the-art ANN decoders while reducing computation operations to 6.8% and memory access to 9.4%.
Brain-machine interfaces (BMIs) are promising for motor rehabilitation and mobility augmentation. High-accuracy and low-power algorithms are required to achieve implantable BMI systems. In this paper, we propose a novel spiking neural network (SNN) decoder for implantable BMI regression tasks. The SNN is trained with enhanced spatio-temporal backpropagation to fully leverage its ability in handling temporal problems. The proposed SNN decoder achieves the same level of correlation coefficient as the state-of-the-art ANN decoder in offline finger velocity decoding tasks, while it requires only 6.8% of the computation operations and 9.4% of the memory access.