Leveraging Recurrent Neural Networks for Predicting Motor Movements from Primate Motor Cortex Neural Recordings
Year: 2024
Type: conference-paper
Abstract: 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 0.71R2 score, surpassing the baseline models in Neurobench and is ranked first for R2 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 R2 score compared to the unpruned model.
DOI: 10.5167/uzh-268870
Authors: Wang, Yuanxi, Wang, Zuowen, Liu, Shih-Chii
Language: English
Subfield: Cognitive Neuroscience
Field: Neuroscience
Domain: Life Sciences
Open Access status: green
OpenAlex ID: W6910877274
FWCI:
Cites:
Cited by:
Related to: