This paper presents an efficient deep learning solution for decoding motor movements from neural recordings in non-human primates. An Autoencoder Gated Recurrent Unit (AEGRU) model was adopted as the model architecture for this task. The autoencoder is only used during the training stage to achieve better generalization. Together with the preprocessing techniques, our model achieved 0.71 $R^2$ score, surpassing the baseline models in Neurobench and is ranked first for $R^2$ in the IEEE BioCAS 2024 Grand Challenge on Neural Decoding. Model pruning is also applied leading to a reduction of 41.4% of the multiply-accumulate (MAC) operations with little change in the $R^2$ score compared to the unpruned model.

翻译：本文提出了一种高效的深度学习解决方案，用于从非人灵长类动物的神经记录中解码运动动作。本研究采用自编码器门控循环单元模型作为此任务的模型架构。自编码器仅在训练阶段使用，以实现更好的泛化能力。结合预处理技术，我们的模型取得了0.71的R²分数，超越了Neurobench中的基线模型，并在IEEE BioCAS 2024神经解码大挑战赛中R²指标排名第一。模型剪枝技术亦被应用，与未剪枝模型相比，在R²分数变化极小的情况下，乘累加运算量减少了41.4%。