Repeatability, reliability, and stability of eye movement measurements in Parkinson's disease, cerebellar ataxia, and healthy adults.

Introduction: Eye movements have been proposed as biomarkers to track disease progression and treatment effects in neurological diseases. Before such variables are used in the clinic or in drug trials, properties such as measurement error must be documented. In this study, we assessed repeatability, reliability, and stability of fixation, smooth pursuit, and saccade measurements in patients with Parkinson's disease, cerebellar ataxia, and healthy adults. Methods: Fixation, smooth pursuit, and saccade metrics were measured in 16 patients with Parkinson's disease, 16 patients with ataxia, and 25 healthy adults with an eye tracker (BulbiCam). The same operator repeated the measurements six times over 2 days in the patient group and two times the same day in the healthy adults. Reliability, repeatability, and stability were assessed with the intraclass correlation coefficient (ICC), Bland-Altman plots with the Agreement Index, and the Stability Index, respectively. Results: Mean pupil size in the fixation test and latency, accuracy and peak velocity in the pro-saccade test were found reliable, repeatable, and stable. Mean and max fixation in the fixation test were found reliable and stable. Smooth pursuit measurements were found repeatable within patients and stable, but not reliable. Conclusion: The saccade and pupil variables may be used both on a population level and for individual patient follow-up. Mean and max fixation duration may be used on the population level but used in the clinical evaluation on individual patients they need to be repeated.