Distributed cortico-subcortical networks enable robust speech state detection from sparse intracranial recordings | Synapse