Preservation of Eye Movements in Parkinson's Disease Is Stimulus- and Task-Specific

Abstract

Parkinson's disease (PD) is a neurodegenerative disease that includes motor impairments, such as tremor, bradykinesia, and postural instability. Although eye movement deficits are commonly found in saccade and pursuit tasks, preservation of oculomotor function has also been reported. Here we investigate specific task and stimulus conditions under which oculomotor function in PD is preserved. Sixteen PD patients and 18 healthy, age-matched controls completed a battery of movement tasks that included stationary or moving targets eliciting reactive or deliberate eye movements: pro-saccades, anti-saccades, visually guided pursuit, and rapid go/no-go manual interception. Compared with controls, patients demonstrated systematic impairments in tasks with stationary targets: pro-saccades were hypometric and anti-saccades were incorrectly initiated toward the cued target in ∼35% of trials compared with 14% errors in controls. In patients, task errors were linked to short latency saccades, indicating abnormalities in inhibitory control. However, patients' eye movements in response to dynamic targets were relatively preserved. PD patients were able to track and predict a disappearing moving target and make quick go/no-go decisions as accurately as controls. Patients' interceptive hand movements were slower on average but initiated earlier, indicating adaptive processes to compensate for motor slowing. We conclude that PD patients demonstrate stimulus and task dependency of oculomotor impairments, and we propose that preservation of eye and hand movement function in PD is linked to a separate functional pathway through the superior colliculus-brainstem loop that bypasses the fronto-basal ganglia network. Our results demonstrate that studying oculomotor and hand movement function in PD can support disease diagnosis and further our understanding of disease progression and dynamics.SIGNIFICANCE STATEMENT Eye movements are a promising clinical tool to aid in the diagnosis of movement disorders and to monitor disease progression. Although Parkinson's disease (PD) patients show some oculomotor abnormalities, it is not clear whether previously described eye movement impairments are task-specific. We assessed eye movements in PD under different visual (stationary vs moving targets) and movement (reactive vs deliberate) conditions. We demonstrate that PD patients are able to accurately track moving objects but make inaccurate eye movements toward stationary targets. The preservation of eye movements toward dynamic stimuli might enable patients to accurately act on the predicted motion path of the moving target. These results can inform the development of tools for the rehabilitation or maintenance of functional performance.

Keywords: Parkinson's disease; eye movements; prediction; preservation of function; saccades; smooth pursuit.

Figure 1.

Stimulus characteristics and movement requirements in a battery of oculomotor tasks.

Figure 2.

Flowchart of participant inclusion by participant group, task, and medication status.

Figure 3.

Sequence of events and eye movements in the pro-saccade task. A, Each trial started with a drift correction followed by a fixation period. Participants had to saccade to the cued target square. B, 2D eye position in pro-saccade task for a representative PD patient (purple) and control participant (green). For illustration purposes, eye and target position data were flipped to always depict the saccade target on the right. C, Main sequence (saccade velocity vs amplitude) for 2 representative patients (purple circles) and 2 control participants (green circles). Each circle represents one trial. D, Saccade latency distributions (relative frequency of binned saccade latencies) for patients and controls. E, Mean saccade amplitude as a function of saccade latency. Each dot represents the mean saccade amplitude in a 50 ms time bin across all patients (purple) and controls (green). Vertical lines indicate SE. **p < 0.01; ***p < 0.001; ranked sum test.

Figure 4.

Sequence of events and eye movements in the anti-saccade task. A, Each trial started with a drift correction followed by a fixation period. Participants had to saccade to the uncued target square. B, 2D eye position in pro-saccade task for a representative PD patient (purple) and control participant (green). For illustration purposes, eye and target position data were flipped to always depict the saccade target on the right. C, Saccade latency distributions (relative frequency of binned saccade latencies) for patients and controls. Blue bins represent changes of mind. Red bins represent direction errors. D, Task performance (percentage of saccades toward uncued location without any corrections) as a function of saccade latency. ***p < 0.001 (ranked sum test).

Figure 5.

Comparison of pro- and anti-saccade task performance. A, Relationship between the frequency of express saccades during the pro-saccade task and the error rate (saccade toward the cued target) in the anti-saccade task. Each circle represents a patient (purple) and control participant (green). Significant regression results in patient group: ***p < 0.001. B, Saccade distributions of a control participant (C57; green) and patient (P35; purple) who had a similar rate of express saccades.

Figure 6.

Sequence of events and eye movements in sinusoidal pursuit task. A, Each trial started with a drift correction followed by five cycles of sinusoidal target motion in either horizontal or vertical direction. B, 2D eye position for a horizontally moving target at a speed of 16 deg/s for a representative PD patient (purple) and control participant (green). Saturated segments represent saccades. Lighter segments represent smooth pursuit.

Figure 7.

Sequence of events and eye movements in go/no-go track-intercept task. A, Each trial started with a fixation period. Participants viewed a moving, disappearing target and had to judge whether the target would miss or pass a strike box. B, 2D eye position in track-intercept task for a representative PD patient (purple) and control participant (green). C, First catch-up saccade latency distributions (relative frequency of binned saccade latencies) for patients and controls. Red bins indicate trials in which the go/no-go decision was incorrect. D, Go/no-go decision accuracy as a function of initial catch-up saccade latency for patients (purple) and controls (green). Circles represent group mean for given saccade interval. **p < 0.01 (ranked sum test).

Figure 8.

Hand movement dynamics in track-intercept task. A, Hand movement velocity across time for individual (thin lines) patients (purple) and controls (green). Thick lines indicate group average. B, Interception timing error for patients and controls. Positive timing errors indicate that participants intercepted too early; negative timing error indicates late interceptions.