Neural representations of cognitive control and speech in the subthalamic nucleus

The subthalamic nucleus (STN) is a key site for control of motor function in humans. Deep brain stimulation of the subthalamic nucleus (STN-DBS) can improve movements in patients with Parkinson's disease (PD). However, STN-DBS can also have adverse effects on cognition and speech. The mechanisms behind this differential effect of STN-DBS are unclear. In this work, I share results from two sets of experiments with PD patient volunteers that explored how the STN encodes (1) cognitive control and (2) speech. The first set of experiments began with a novel electrical stimulation tract tracing paradigm performed during STN-DBS implantation surgeries. Here, suprathreshold stimulation in the human STN resulted in antidromic, short-latency evoked potentials in the prefrontal cortex, a region strongly associated with cognition. Additional intracranial electrode recordings showed that STN field potentials were coherent with 4-Hz oscillations in prefrontal cortex during performance of a simple cognitive task (fixed interval timing). These data supported the notion that a monosynaptic, cortico-STN circuit—the so-called ‘hyperdirect pathway’—played a role in cognitive control during interval timing. Subsequent testing showed that postoperative STN-DBS patients performed 12-sec fixed interval timing more accurately with 4-Hz stimulation than with high-frequency stimulation (130-Hz) or no stimulation (0-Hz). These data suggest that low-frequency deep brain stimulation may alleviate cognitive deficits in Parkinson's disease patients. The second set of experiments directly compared speech- and limb movement-related modulation in individual STN neurons. Twenty-eight STN neurons were isolated from 12 intraoperative STN-DBS patients who performed alternating trials of two tasks: speech and contralateral limb movement. Neuronal activity was assessed in terms of firing rate and event-related modulation, which examined firing rate changes around task onset and offset. These measures were regressed onto selected PD clinical measures to determine if preoperative clinical metrics (e.g. scores on the unified PD rating scale part III; Parkinson’s disease duration) predicted speech- and limb movement-evoked neuronal firing patterns. Linear mixed effects modeling of response modulation index data revealed that STN cells were differentially modulated by speech and limb movement. Additionally, participants with longer PD duration trended towards higher firing rates across STN neurons. These data demonstrate that STN neurons are differentially modulated during speech versus contralateral limb movement. This observation may suggest a mechanistic basis for the often-seen disparate effects of STN-DBS on speech and limb movement.