Decoding neural signals with a compact and interpretable convolutional neural network

AO_SCPLOWBSTRACTC_SCPLOWBrain-computer interfaces (BCIs) decode information from neural activity and send it to external devices. In recent years, we have seen an emergence of new algorithms for BCI decoding including those based on the deep-learning principles. Here we describe a compact convolutional network-based architecture for adaptive decoding of electrocorticographic (ECoG) data into finger kinematics. We also propose a theoretically justified approach to interpreting the spatial and temporal weights in the architectures that combine adaptation in both space and time, such as the one described here. In these architectures the weights are optimized not only to align with the target sources but also to tune away from the interfering ones, in both the spatial and the frequency domains. The obtained spatial and frequency patterns characterizing the neuronal populations pivotal to the specific decoding task can then be interpreted by fitting appropriate spatial and dynamical models. We first tested our solution using realistic Monte-Carlo simulations. Then, when applied to the ECoG data from Berlin BCI IV competition dataset, our architecture performed comparably to the competition winners without requiring explicit feature engineering. Moreover, using the proposed approach to the network weights interpretation we could unravel the spatial and the spectral patterns of the neuronal processes underlying the successful decoding of finger kinematics from another ECoG dataset with known sensor positions. As such, the proposed solution offers a good decoder and a tool for investigating neural mechanisms of motor control.