Poster P15 - Measuring Changes in Prosaccade, Antisaccade, and Memory-Guided Saccade across Three Seasons of Play in High School Football Athletes

Introduction: Encompassing 1.4 million emergency room visits and 52,000 deaths in the United States per year, traumatic brain injury (TBI), defined by subsequent functional deficit following an external force, is a major public health issue. Mild traumatic brain injuries (mTBI) are estimated to account for 80-90% of all TBI cases. However, current definitions of mTBI lack unity and diagnostics lack objectivity and granularity. Assessing subtler differences in behavior or response when employing conventional diagnostics is difficult to differentiate at lower severities (e.g. sub-concussive exposure). Current multimodal diagnostics of concussion such as SCAT5 incorporate scores on cognitive tests, balance tests, and symptom reporting but differences associated with sub-concussive exposure may be difficult to quantify and diagnose using traditional assessments.
Eye-tracking was previously employed to measure oculomotor deficits in disorders including Parkinson’s, ALS, MS, Alzheimer’s, ADHD, and epilepsy. By measuring differences in saccades or rapid eye movements in tasks, a repeatable measure of cognitive and oculomotor function can be taken and compared between different groups. This enables quantitative, objective, and granular comparison between different levels of exposure. This study incorporates eye-tracking to quantify oculomotor changes over four years of football gameplay.

Materials and

Methods: High school football athletes participated in eye-tracking assessments prior to each season. Seven subjects were assessed over four years and three seasons. Each participant performed prosaccade (n=28), anti-saccade (n=25), and memory-guided saccade (n=25) tasks recorded using an Eyelink 1000 eye-tracking system (SR Research, Ottawa, Canada) at 1000 Hz.
In prosaccade, the gaze is fixed to a stimulus. The stimulus vanishes and appears in an offset position. The eye then saccades to the offset position. In anti-saccade, though the stimulus similarly disappears and reappears in an offset position, the participant must resist a saccade towards the stimulus and saccade to an opposite offset position. In memory-guided saccade, the subject’s gaze is also fixed to a dot stimulus. A second stimulus appears in an offset position but eventually vanishes, after which both stimuli vanish. The subject must resist saccade while the second stimulus is present but saccade to the recalled offset position following disappearance of both stimuli.
Measurements taken include latency, gain, and end-gain. Latency covers the time elapsed between stimulus position change and saccade initiation. Gain encompasses the initial saccade amplitude or distance relative to stimulus offset distance. End-gain is the final saccade distance relative to stimulus offset distance. Repeated measures ANOVA and post-hoc Tukey pairwise comparisons were employed to determine when significant differences between years occurred respectively. Effect sizes (eta-squared) per metric were compared between year and participant to elucidate whether differences were produced by noise from inter-subject variability or oculomotor changes. For all analysis, statistical significance was at p< 0.05.

Results, Conclusions, and Discussions: Significant differences were present in prosaccade latency, anti-saccade gain/end-gain, and all memory-guided saccade metrics. The post-hoc comparison of years in prosaccade latency revealed a significant increase between year 1, 2, and 3, but difference became insignificant at year 4. Anti-saccade hypometric gain/end-gain both only became significantly different from year 1 at year 4. In memory-guided saccade latency over time significantly decreased between all years, but for gain/end-gain significant differences from year 1 included an increase at year 3 and decrease in years 2/4 respectively. Determining the extent of differences between years relative to inter-participant variability provides insight into relative magnitudes of oculomotor changes. Effect size comparisons reveal larger contributions from timepoint differences than participant variability only in memory-guided saccade end-gain. More participants would help differentiate whether effects seen are attributable to true oculomotor change or inter-participant variability.
Oculomotor differences were observed in eye-tracking tasks across three seasons. Similar differences exist in literature. Hypometric anti-saccade gain and end-gain were seen in patients diagnosed with Parkinson’s and ADHD but were also present in some studies on healthy controls. Prior studies show that diminished athletic performance resulting from oculomotor deficit increases risk of further injury. Granularity provided from these eye-tracking metrics may provide insights into specific oculomotor deficits incurred and their relationship to performance (e.g. reaction time). Future work may also evaluate interdependent relationships between latency, gain, and end-gain.
The oculomotor changes seen in the 7 participants may not be attributable to football induced sub-concussive loading given present limitations. While all players begin assessments as freshmen, some athletes may have prior exposure in middle school football. As football is a positional sport, different positions may incur different head impact exposure. This study compares data between different timepoints in exposure history and does not possess non-exposure controls. This study introduces future work opportunities for further comparison with both non-exposure and concussed patient populations to associate eye-tracking results with clinical outcomes. By providing measurements over 4 years of exposure, comparison with non-exposure and concussed cohorts permits evaluation of sub-concussive injury risk over time.

Acknowledgements (Optional): The authors gratefully thank the Duke University Bass Connections Program and the Cardinal Gibbons High School community for their support of this work.