Subthalamic nucleus neurons encode syllable sequence and phonetic characteristics during speech

Abstract

Speech is a complex behavior that can be used to study unique contributions of the basal ganglia to motor control in the human brain. Computational models suggest that the basal ganglia encodes either the phonetic content or the sequence of speech elements. To explore this question, we investigated the relationship between phoneme and sequence features of a spoken syllable triplet and the firing rate of subthalamic nucleus (STN) neurons recorded during the implantation of deep brain stimulation (DBS) electrodes in individuals with Parkinsons disease. Patients repeated aloud a random sequence of three consonant-vowel (CV) syllables in response to audio cues. Single-unit extracellular potentials were sampled from the sensorimotor STN; a total of 227 unit recordings were obtained from the left STN of 25 subjects (4 female). Of these, 113 (50%) units showed significant task-related increased firing and 53 (23%) showed decreased firing (t-test relative to inter-trial period baseline, p<0.05). Linear regression analysis revealed that both populations of STN neurons encode phoneme and sequence features of produced speech. Maximal phoneme encoding occurred at the time of phoneme production, suggesting efference copyor sensory-related processing, rather than speech motor planning (-50ms and +175ms relative to CV transition for consonant and vowel encoding, respectively). These findings demonstrate that involvement of the basal ganglia in speaking includes separate single unit representations of speech sequencing and phoneme selection in the STN.