In this episode, we speak with Zixiao Yin, recipient of a Junior Award at the MDS Congress, about his work on reconstructing deep brain activity from cortical signals.

Using more than 700 hours of simultaneous cortical and subcortical recordings across different movement disorders, brain targets, and clinical states, Yin and colleagues developed deep learning approaches to infer subcortical activity from cortical recordings. We discuss why this could matter for adaptive DBS, how cortical signals can be used to estimate STN beta activity and reconstruct deep brain signals, and whether this approach could provide an alternative when direct subcortical sensing is unreliable.