Multi-finger motion inference using M1 neural decoding

This paper proposed a novel decoding method, which is the decoding technique for inferring multi-finger motions trained by the neural activities acquired in single finger motions. We have exploited the relationship between single and multi-finger motions. Then the model for multi-finger motions was built. The proposed method decoded multi-finger motions with 100% accuracy without a priori training information of multi-finger. These results suggest that only with the neural activities on single finger motions we can infer multi-finger motions.