(11) Activation of prefrontal cortico-STN hyperdirect pathway alters response inhibition in patients with Parkinson’s disease

Introduction: Deep brain stimulation (DBS) has variable effects on the ability to inhibit unwanted actions (response inhibition) in people with Parkinson’s disease (PD). This has been attributed to the location of stimulation within the subthalamic nucleus (STN) and the activation of prefrontal cortico-STN hyperdirect projections. We hypothesized that maximizing activation of prefrontal projections impairs proactive response inhibition to a greater extent than minimizing activation or sham stimulation.

Methods: We used a cued Go/NoGo task where a left/right cue is presented on screen followed by a Go signal requiring a button press. On 25% of trials, a NoGo signal is given instead which mandates withholding the planned response. We measured median response times (RT) and percent NoGo response errors. Patients participated in two sessions (both off medication) within a two week period. Each session included two DBS settings (random order) with 30 minutes wash in period as follows: session 1) clinical and sham (0.1 mA/2 Hz) and session 2) maximized (Max) and minimized (Min) activation of prefrontal cortico-STN hyperdirect projections. To determine the optimal stimulation contact for Max and Min settings, we used predictions from patient-specific biophysical DBS models. The models used driving force predictor functions to estimate percent activation of prefrontal cortico-STN hyperdirect projections in a connectomic atlas in patient (preoperative T1) space. Additional setting parameters were 2.5mA/60µs/130Hz (or amplitude was set to ~20% below persistent side effect threshold).

Results: We report results for 21 patients (3 women, age 60.3±8 years, disease duration 10.9±4.5 years). The response time did not differ between the settings (Sham 465±124ms; Clinical 420±115ms; Max 463±150ms; Min 451±133ms). NoGo response errors (Sham 1.7±8.3 IQR; Clinical 4.2±7.5 IQR; Max 10±6.6 IQR; Min 5±9.58 IQR) were significantly affected by different settings (Friedman test’s P-value = 0.05). Although errors did not differ between Max and Min, they were significantly higher in Max compared to Sham and Clinical settings (p=0.03, p=0.04 respectively; paired Wilcoxon signed rank test). The absolute percent activation of prefrontal cortico-STN hyperdirect projections did not correlate with NoGo response errors.

Conclusion: We conclude that maximizing activation of prefrontal cortico-STN hyperdirect projections led to worsening response inhibition, compared to no stimulation or clinically optimized DBS. Whether this effect is related to relative difference in percent pathway activation for individual patients is pending further study.