Movement Decoding using Spatio-Spectral Features of Cortical and Subcortical Local Field Potentials

The application of machine learning to intracranial signal analysis has the potential to revolutionize deep brain stimulation (DBS) by personalizing therapy to dynamic brain states, specific to symptoms and behaviors. Most decoding pipelines for movement decoding in the context of adaptive DBS are based on single channel frequency domain features, neglecting spatial information available in multichannel recordings. Such features are extracted either from DBS lead recordings in the subcortical target and/or from electrocorticography (ECoG). To optimize the simultaneous use of both types of signals, we developed a supervised online-compatible movement decoding pipeline based on multichannel and multiple site recordings. We found that adding spatial information to the model has the potential to improve decoding. In addition, we demonstrate movement decoding from spatio-spectral features derived from cortical and subcortical oscillations. We demonstrate between-patients variability of the spatial neural maps and its relationship to feature decoding performance. This application of spatial filters to decode movement from combined cortical and subcortical recordings is an important step in developing machine learning approaches for intelligent DBS systems.