Characterization of Two-Dimensional Oculomotor Control During Goal-Directed Eye Movements in Humans
Oculomotor control is a subset of sensorimotor control that allows humans to make extremely accurate eye movements for ADL. Impairments to oculomotor control can increase the impact of sensorimotor control deficits, especially in neurodegenerative diseases such as MS. Here, a two-dimensional computational control system of saccades and smooth-pursuit eye movements was compiled from literature to systematically characterize oculomotor control in eight visually-healthy humans as a precursor to studying the relationship between oculomotor and sensorimotor control in patient populations. Subjects visually tracked a single dot on a 41 x 30.5 cm monitor in a dark room while eye positions were recorded at 60 Hz by a video based eye tracker. Data from visual tasks separately consisting of saccades and smooth-pursuit along the horizontal and/or vertical midlines were inputs to an error minimization algorithm that identified individually for each subject the parameters characterizing motor command generation and two-dimensional interactions within ocular dynamics, with bootstrap analysis quantifying the certainty of parameter estimates. Cross-correlation between target and subject gaze positions was used to identify neuronal conduction speeds for saccades and smooth-pursuit processing. A task consisting of small saccades identified the minimum position error required for saccade initiation. A final task combining saccade and smooth-pursuit movements was used to evaluate model performances. The model accounted for 96% and 98% of variability for subject saccade and smooth-pursuit eye movements, respectively. The 2-D model analysis of saccades and smooth-pursuit identified interactions between horizontal and vertical oculomotor control indicative of component stretching but did not verify the increased speed of vertical versus horizontal eye movements reported in literature. A novel interaction associated with centrifugal curvature was also identified, but the functional effects the interactions were small. Estimated latencies of saccade and smooth-pursuit processing of 242 and 107 ms, respectively, were within ranges provided by literature, while dead zone values for saccade initiation had a 97% error from values provided by literature. The quantitative framework presented in this study may be used in future studies that include MS patients, in which oculomotor control characterization may reveal differences in control strategies for goal-directed ocular movements relative to healthy individuals.