Increased Subthalamic Nucleus Deep Brain Stimulation Amplitude Impairs Inhibitory Control of Eye Movements in Parkinson's Disease

Abstract

Background and objectives: Bilateral subthalamic nucleus deep brain stimulation (STN DBS) in Parkinson's disease (PD) can have detrimental effects on eye movement inhibitory control. To investigate this detrimental effect of bilateral STN DBS, we examined the effects of manipulating STN DBS amplitude on inhibitory control during the antisaccade task. The prosaccade error rate during the antisaccade task, that is, directional errors, was indicative of impaired inhibitory control. We hypothesized that as stimulation amplitude increased, the prosaccade error rate would increase.

Materials and methods: Ten participants with bilateral STN DBS completed the antisaccade task on six different stimulation amplitudes (including zero amplitude) after a 12-hour overnight withdrawal from antiparkinsonian medication.

Results: We found that the prosaccade error rate increased as stimulation amplitude increased (p < 0.01). Additionally, prosaccade error rate increased as the modeled volume of tissue activated (VTA) and STN overlap decreased, but this relationship depended on stimulation amplitude (p = 0.04).

Conclusions: Our findings suggest that higher stimulation amplitude settings can be modulatory for inhibitory control. Some individual variability in the effect of stimulation amplitude can be explained by active contact location and VTA-STN overlap. Higher stimulation amplitudes are more deleterious if the active contacts fall outside of the STN resulting in a smaller VTA-STN overlap. This is clinically significant as it can inform clinical optimization of STN DBS parameters. Further studies are needed to determine stimulation amplitude effects on other aspects of cognition and whether inhibitory control deficits on the antisaccade task result in a meaningful impact on the quality of life.

Keywords: Antisaccade; Parkinson's disease; deep brain stimulation; inhibitory control; stimulation amplitude.

Figure 1:. The antisaccade task.

A) The time course of one antisaccade trial. The trial begins with a variable 2000-3000ms fixation period (grey bar + striped grey bar that indicates the variable fixation durations). Fixation is followed by a 200ms gap before the target appears for 1800ms (black bar). B) Antisaccade: After target onset (black peripheral circle), the participant will look in the opposite direction from the target to the mirror image location. C) Prosaccade Error: After target onset, the participant will look toward the target instead of looking in the opposite direction.

Figure 2:. Stimulation amplitude effects on prosaccade error.

A) Distribution of the observed prosaccade error rate at each experimental stimulation amplitude. Each boxplot contains the median (horizontal line in the box), the upper quartile (75^th percentile, top of box), the lower quartile (25^th percentile, bottom of box), whiskers, and outliers that are beyond 1.5 times the interquartile range (75^th percentile minus the 25^th percentile) from the 25^th or 75^th percentiles (circles). B) The relationship between estimated prosaccade error rate and stimulation amplitude. The plot depicts the model-estimated prosaccade error rate (solid grey line) using 3249 data points across 6 different stimulation amplitudes for 10 participants (shaded area, 95% confidence interval). The observed mean prosaccade error rate is overlaid onto the plot using the same data (black filled circles). Stimulation amplitude was in volts for 8 participants and milliamperes for 2 participants.

Figure 3:. Active electrode contacts and the STN.

The left and right STN are presented from an anterior view. Each circle represents an active contact of an electrode from 5 participants. The contacts are grouped based on whether the participant has a positive (red) or negative (blue) relationship between stimulation amplitude and prosaccade error rate. The background image is a coronal slice of an MNI template brain.

Figure 4:. VTA-STN overlap effects on prosaccade error rate.

A) The plot depicts the model-estimated prosaccade error rate associated with VTA-STN overlap at 4 different stimulation amplitudes (solid lines) from 5 participants (shaded areas, 95% confidence intervals). The observed mean prosaccade error rate for each participant at each stimulation amplitude is overlaid onto the plot using the same data (filled circles). B) The observed mean prosaccade error rate and modelled VTA-STN overlap for 5 participants at 4V stimulation amplitude. The values for the 5 participants have been grouped based on the slope of the regression between stimulation amplitude and prosaccade error rate for each participant. 3 participants had a slight negative relationship (blue circles) and 2 participants had a slight or strong positive relationship (red circles) between stimulation amplitude and prosaccade error rate.

Figure 5:. Stimulation amplitude effects on primary saccade latency.

Distribution of the observed latency at each experimental stimulation amplitude for A) correct antisaccade trials and B) prosaccade error trials. Each boxplot contains the median (horizontal line in the box), the upper quartile (75^th percentile, top of box), the lower quartile (25^th percentile, bottom of box), whiskers, and outliers that are beyond 1.5 times the interquartile range (75^th percentile minus the 25^th percentile) from the 25^th or 75^th percentiles (circles). C) The relationship between estimated latency and stimulation amplitude for correct antisaccade trials (upper dark grey band) and prosaccade error trials (lower light grey band). The plot depicts the model-estimated latency (solid grey lines) across 6 different stimulation amplitudes for 10 participants (shaded areas, 95% confidence intervals). The observed mean latencies are overlaid onto the plot using the same data (black filled circles). Stimulation amplitude was in volts for 8 participants and milliamperes for 2 participants.