Concurrent decoding of acoustic detail and linguistic structure enables natural, intelligible speech synthesis from limited human cortical recordings, resolving a fundamental constraint in neural ...