Decoding wrist kinematics from local field potentials of the ipsilateral primary motor and dorsal premotor cortices

Local field potentials (LFP) are valuable signals for decoding motor kinematics in brain machine interfaces (BMIs). To take full advantage of LFPs, however, more systematic investigation of the relationship between LFPs and ipsilateral limb movement is required. In this paper we investigated the decoding power of LFPs for the ipsilateral wrist movement from two monkeys performing a 2D center-out task. The results show that LFPs could predict the ipsilateral wrist position and velocity with high accuracy, which is comparable to the accuracy of decoding the contralateral kinematics. Furthermore, similar to contralateral decoding, the low (0.3-5 Hz, 5-15 Hz) and high (100-200 Hz, 200-400 Hz) frequency bands resulted in significantly better decoding performance than the medium frequency bands. These results suggest that ipsilateral LFPs could be used to build better BMIs in similar ways of using contralateral LFPs.