An alarming 30-45% of patients undergoing inpatient rehabilitation for Traumatic Brain Injury (TBI) may experience spatial neglect (SN), a condition that significantly impedes recovery. Individuals with SN often exhibit limited environmental awareness and a consistent gaze bias towards one side of their spatial environment. While the neural underpinnings of SN and SN recovery are well studied in stroke survivors, there is a gap in understanding these mechanisms in TBI patients, who often have distributed brain lesions due to coup-countercoup injury. In stroke rehabilitation, Smooth Pursuit Eye Exercises (SP) coupled with repetitive optokinetic stimulation were shown to enhance visual exploration. Our study aims to identify the neural mechanisms underlying the effects of SP in individuals with chronic TBI.

Fifteen TBI participants underwent 3T MRI scans during which they viewed videos of 70 randomly distributed red dots coherently moving leftward or rightward against a black background. Participants either picked a dot and followed it with their eyes, until it disappeared off the edge of the screen (SP trials) or ignored the moving dots and fixated on a central white dot (fixation trials).  Eye movements were recorded using MRI-compatible eye tracker. SN was assessed using a multi-test diagnostic method. Neglect severity was defined as the total number of tests performed below a healthy cut-off. Evidence for left and right neglect was considered separately.

Relative to fixation, SP activated bilateral oculomotor network, including superior parietal cortex, frontal and supplementary eye fields (FEF, SEF), and anterior middle frontal gyrus. Additionally, we saw greater activation in the visual and sensorimotor networks in the SP compared to fixation trials. During rightward SP, activity in the left inferior and right superior parietal lobules (IPL, SPL), and left middle and superior temporal gyri correlated with severity of right SN. Activity in the right FEF, IPL and precentral gyrus correlated with severity of left SN. Similarly, during leftward SP, activity in the right FEF, IPL, superior frontal gyrus and frontal pole correlated with severity of left SN. Participants could be grouped into 4 eye movements patterns according to average dwell time in ms across 10 horizontal sections of the screen: right preference (n=7), symmetrical (n=3), left preference (n=3), and central preference (n=1). These eye movement patterns corresponded broadly to SN scores, such that participants with left SN had a right preference. However, correspondence for the other eye movement groups did not strictly conform to SN scores, suggesting eye tracking captured unique behavioral variance.

Our data suggest that participants with chronic TBI show predictable patterns of activation in the oculomotor and sensorimotor networks during smooth pursuit eye exercises. This finding fills the current gap in identifying neuroscience-based targets for spatial neglect rehabilitation in this population.  We showed that left and right SN have separable neural correlates within contra- and ipsilateral dorsal and ventral areas typically associated with neural attention networks. Finally, we found that eye movement patterns may capture patterns of potentially diagnostic oculomotor behavior addressing SN characteristics not already identified with neuropsychological testing.