Temporal Changes in Fixational Eye Movements After Concussion in Adolescents and Adults: Preliminary Findings

Abstract

Concussions often involve ocular impairment and symptoms such as convergence insufficiency, accommodative insufficiency, blurred vision, diplopia, eye strain, and pain. Current clinical assessments of ocular function and symptoms rely on subjective symptom reporting and/or involve lengthy administration time. More objective, brief assessments of ocular function following concussion are warranted. The purpose of this study was to evaluate changes in fixational eye movements (FEMs) and their association with clinical outcomes including recovery time, symptoms, cognitive and vestibular/ocular motor impairment. Thirty-three athletes (13–27 years of age; 54.5% female) within 21 days of a diagnosed concussion participated in the study. A tracking scanning laser ophthalmoscope (TSLO) evaluated FEMs metrics during fixation on a center and corner target. Participants completed symptom (Post-Concussion Symptom Scale [PCSS]), cognitive (Immediate Post-concussion Assessment and Cognitive Testing [ImPACT], and Vestibular/Ocular Motor Screening (VOMS) evaluations. All measures were administered at the initial visit and following medical clearance, which was defined as clinical recovery. Changes in FEMs were calculated using paired-samples t tests. Linear regression (LR) models were used to evaluate the association of FEMs with clinical recovery. Pearson product-moment correlations were used to evaluate the associations among FEMs and clinical outcomes. On the center task, changes across time were supported for average microsaccade amplitude (p = 0.005; Cohen's d = 0.53), peak velocity of microsaccades (p = 0.01; d = 0.48), peak acceleration of microsaccades (p = 0.02; d = 0.48), duration of microsaccade (p < 0.001; d = 0.72), and drift vertical (p = 0.017; d = -0.154). The LR model for clinical recovery was significant (R² = 0.37; p = 0.023) and retained average instantaneous drift amplitude (β = 0.547) and peak acceleration of microsaccade (β = 0.414). On the corner task, changes across time were supported for drift proportion (p = 0.03; d = 0.43). The LR model to predict clinical recovery was significant (R² = 0.85; p = 0.004) and retained average amplitude of microsaccades (β = 2.66), peak velocity of microsaccades (β = -15.11), peak acceleration of microsaccades (β = 12.56), drift horizontal (β = 7.95), drift vertical (β = 1.29), drift amplitude (β = -8.34), drift proportion (β = 0.584), instantaneous drift direction (β = -0.26), and instantaneous drift amplitude (β = 0.819). FEMs metrics were also associated with reports of nausea and performance within the domain of visual memory. The FEMs metric were also associated with PCSS, ImPACT, and VOMS clinical concussion outcomes, with the highest magnitude correlations between average saccade amplitude and VOMS symptoms of nausea and average instantaneous drift speed and ImPACT visual memory, respectively. FEMs metrics changed across time following concussion, were useful in predicting clinical recovery, and were correlated with clinical outcomes. FEMs measurements may provide objective data to augment clinical assessments and inform prognosis following this injury.

Introduction

It is estimated that between 1,600,000 and 3,800,000 sport and recreation-related concussions occur in the United States each year.¹ Concussion results in a diffuse energy crisis that leads to heterogenous symptoms including headaches, dizziness, nausea, and photo/phonosensitivity, as well as difficulties with balance and cognitive, ocular, and vestibular impairments. Between 65% and 79% of concussions involve ocular impairments^2,3 such as receded near point of convergence (NPC), accommodative insufficiency, and saccadic dysfunction. These impairments are associated with visual symptoms including blurred vision, headaches, eye strain, sleepiness, difficulties reading, and difficulties concentrating.^4
–6 This high rate of ocular impairment and symptoms following concussion is not surprising given that approximately half of the neural connections in the human brain are involved in visual function.⁷ Ocular dysfunction is also associated with neurocognitive deficits and prolonged recovery following concussion.^8,9 There is further emerging evidence that ocular therapy is effective following concussion.⁴ Taking all this together, there is a pressing need for early identification of ocular dysfunction in order to inform timely and effective interventions.

Two of the most widely used clinical tools for the purpose of screening patients for ocular function following concussion include the Vestibular/Ocular Motor Screening (VOMS) and King–Devick (K–D) test. The VOMS assesses symptom provocation through the completion of smooth pursuits, horizontal and vertical saccades, and NPC. In addition, the VOMS includes an average measurement of NPC distance. Researchers have demonstrated that the VOMS is sensitive to identifying ocular dysfunction following concussion. However, it relies on self-report of symptom provocation, potentially limiting its utility with patients who may be motivated to minimize symptomology.^10

–14 Qualitative information that can be observed during VOMS administration, such as delayed initiation of saccades, slowed speed of saccades, or under/overshooting of targets is also subjective and requires specialized training to perform reliably and accurately. Separately, the K–D test was originally designed to assess reading speed/ability, and evaluates processing speed and visual tracking performance, which require saccadic eye movements, attention, and language function. However, the K–D test requires a baseline for accurate comparison, and its interpretive utility is limited by practice effects, which can result in a high number of false positives.^15,16 Moreover, both the VOMS and K–D are designed as screening tools rather than comprehensive standalone measurements of ocular impairment.¹⁷ Consequently, there is a need for more objective measures of ocular dysfunction to identify and monitor post injury ocular function.

Efficient and objective approaches to investigate and track concussion recovery via assessment of ocular function may be explored through in-depth investigation of fixational eye movements. Fixational eye movements (FEMs), which have been used promisingly to track disease progression in neurodegenerative disorders, are involuntary movements that keep the eye in constant motion when attempting to hold gaze on a fixed target. FEMs include microsaccades, which are microscopic eye movements that occur when attempting to fixate the eyes on a stationary object.¹⁸ Another primary component of FEMs is drift, which is the slow motion that occurs between microsaccades. FEMs have been reported to be associated with cognitive functioning, including performance measures of attention and memory.¹⁹ Video-oculography techniques, which require more sophisticated equipment including goggles and computerized processing, can measure FEMs to identify visual tracking deficits and are more sensitive to impairments than subjective clinical tools. However, because of the small magnitude of FEMs, which is beyond the resolution limit of video oculography devices that track the external motion of the pupil, they are best measured using precision tools such as a tracking scanning laser ophthalmoscope (TSLO). A TSLO device can measure eye movements with precise accuracy by directly imaging retinal motion at a microscopic scale.²⁰ In our previous work, we demonstrated that FEMs were sensitive for differentiating patients with concussion from age- and sex-matched healthy, uninjured controls.²¹ However, researchers have yet to evaluate the utility of FEMs as measured with TSLO for monitoring recovery or prognosis following concussion.

The purpose of this study was to evaluate changes in FEMs following concussion, and their association with clinical outcomes. Specifically, we aimed (1) to identify within-subject changes in FEMs metrics from initial evaluation within <21 days of injury to medical clearance (i.e., clinical recovery), (2) to determine if initial FEMs metrics are useful in prognosis for recovery time, and (3) to evaluate the association of FEMs metrics with concussion clinical outcomes including symptoms, as well as cognitive and vestibular/ocular motor impairment. We hypothesized that FEMs metrics would improve from initial evaluation to clinical recovery and that worse FEMs metrics at the initial evaluation would predict longer clinical recovery. We also hypothesized that FEMs metrics would be correlated with clinical ocular motor outcomes (e.g., smooth pursuits, saccades, NPC) and visually challenging cognitive assessments (e.g., visuo-motor processing speed, reaction time).

Methods

Participants

Participants eligible for this study included patients 13–27 years of age who were within 21 days of a diagnosed concussion, who returned for follow-up FEMs measurements at their date of medical clearance (i.e., clinical recovery). These participants were part of a previously published case-control study.²¹ Patients were excluded if they had a history of other neurological and/or psychiatric conditions, balance or vestibular disorders, oculomotor impairment, or ophthalmic conditions, if they were pending litigation or workers' compensation, or if they had had three or more previous concussions. Of the 50 eligible participants from the original study, 9 (18%) did not complete the corner raster task, as the task was added after the they were enrolled into the study. Eight (16%) additional enrolled participants did not meet the data validation criterion of tracking at least 80% of the strips in the image sequences in three out of the five trials.²¹ As a result, 33/50 (66%) of the original eligible participants were included in this study. However, 5 of the remaining 33 participants only completed the corner raster task at their initial visit, leaving a total sample of 28/50 (56%) eligible participants for that task. There were no significant differences between those who were eventually included in the study and those who were excluded in terms of age, sex, Immediate Post-concussion Assessment and Cognitive Testing (ImPACT) outcomes, and Post-Concussion Symptom Scale (PCSS) outcomes. However, there was a significant difference in terms of concussion history. Those excluded had a significantly higher rate of previous history of concussion (56% vs. 19%, p = 0.021). There were also significant differences in terms of average VOMS total symptom scores at initial evaluation (64.3 vs. 34.4 p = 0.016) (Fig. 1).

FIG. 1.

Consolidated Standards of Reporting Trials (CONSORT) diagram of participating subjects.

Measures

Definition of concussion

Concussion was defined utilizing current consensus guidelines including clear mechanism of injury, presence of one or more signs of injury and/or symptoms at the time of injury, and current symptoms and/or impairment.²² For this study, concussions were diagnosed by licensed healthcare professionals with specialty training in concussion (e.g., clinical neuropsychologists, physicians) based on the abovementioned criteria. For the diagnosis, clinical examination and interview, medical and injury history, reported symptoms, and assessments of cognitive, oculomotor, and/or vestibular impairment were documented.²¹

Recovery was further defined as formal clearance from injury from a sports medicine concussion clinic. Clearances are made when individuals are determined to have fully recovered from concussion and/or no longer have any lingering symptoms from concussion. This is determined using a combination of neurocognitive testing (ImPACT), VOMS, PCSS, and clinical judgment.

VOMS

The VOMS is a screening tool used to identify vestibular and ocular motor symptoms and impairment following concussion. Participants report baseline and post-item provocation of symptoms including headache, dizziness, nausea, and fogginess on a scale of 0 to 10 following each item. Scores are then totaled after administration to provide quantitative measurements of total symptoms (i.e., total headache, total dizziness, total nausea, total fogginess, and VOMS total). The VOMS includes the following items: Smooth Pursuits (SP), Horizontal Saccades (HSAC), Vertical Saccades (VSAC), Horizontal Vestibular Ocular Reflex (HVOR), Vertical Vestibular Ocular Reflex (VVOR), Visual Motion Sensitivity (VMS), and NPC. The VOMS also includes a measure of average NPC distance (cm). The VOMS takes 5–7 min to administer and can be administered during recovery for ongoing monitoring of symptoms.

Neurocognitive testing

ImPACT is a computer-based test used to evaluate cognitive functioning and concussion symptoms. The ImPACT test consists of six subtests that comprise four cognitive composite scores including visual motor speed, reaction time, visual memory, and verbal memory. The test typically takes 20–30 min to complete, and the program itself allows for repeat testing during concussion treatment and recovery for ongoing monitoring of performance changes.

PCSS

The PCSS, which is embedded into the ImPACT tool, is a self-report survey to measure the severity of common symptoms following concussion including headache, dizziness, and sleep, emotional, and memory problems. The PCSS consists of 22 items rated on a seven-point Likert scale, ranging from 0 (none) to 6 (severe). Items are then tallied to provide a total symptom severity score ranging from 0 to 132. The PCSS typically takes 2–3 min to administer and can be administered multiple times during recovery for ongoing monitoring of symptoms.

Procedures

Participants completed the procedures as described in our previous study.²¹ Briefly, researchers recorded imaging sequences using the TSLO (C. Light Technologies, Inc., Berkeley, CA). Most participants completed two fixation tasks (participants completed three total, including a third “active” task that will be covered in a subsequent report) which required the participants to either fixate on the center or corner of the imaging raster that appeared to the participants as a red square on a dark background. Participants first completed the center task followed by the corner task. Thirty second recordings were acquired five times for each task. Experimenters actively monitored image quality during the recordings and administered an additional image sequence if image quality became poor. Fixational saccades and drifts were segmented using procedures described previously, and FEMs metrics were computed for each trial and condition, including saccade direction, amplitude, peak velocity, peak acceleration and duration, drift velocity, amplitude, and drift proportion. Other FEMs metrics included the bivariate contour ellipse area (BCEA) to evaluate spread of fixation; we also computed statistics related to blinks, including number of blinks and mean blink duration. A trained clinical researcher administered the PCSS, ImPACT, and VOMS to each participant. All metrics and clinical measures were obtained at participants' initial clinic visit within 21 days of injury and again at their date of formal medical clearance (i.e., recovery).

Statistical analysis

To evaluate our first hypothesis that FEMs metrics would improve from initial evaluation within 21 days of injury to clinical recovery, we used paired-samples t tests to evaluate changes in FEMs metrics across time. Mean differences and Cohen's effect sizes were reported for significant outcomes. To evaluate our second hypothesis that FEMs metrics would be associated with recovery time, we employed a backwards stepwise multiple linear regression to predict days to clearance utilizing all FEMs metrics as predictors. Cutoff for inclusion into the model was p < 0.1. To evaluate if FEMs metrics were associated with concussion clinical outcomes, we observed the relationships among VOMS, ImPACT, and FEMs from visit 1 using Spearman correlations. Statistical significance was set at p < 0.05 for all tests. Analyses were performed using IBM SPSS Statistics, Version 28 (IBM Corp., Armonk, NY, USA).

Results

Data from 33 of the 50 participants (66%) met validation criteria and completed both tasks. The average age of participants was 17.33 years, with a standard deviation of 3.46 years. Eighteen (54.5%) were female, nine (27%) had sustained a prior concussion, seven (21%) had a history of motion sickness, and nine (27%) had a history of migraines. Five reported a loss of consciousness (15%), and four (12%) had been previously diagnosed with attention-deficit/hyperactivity disorder (ADHD) and/or learning disability. Demographic data are reflected in Table 1.

Table 1.

Participant Demographics

Variable	Concussed (n = 33)
(y), mean ± SD	17.33 ± 3.46
Female, n (%)	18 (55)
Prior concussion, n (%)	9 (27)
History of motion sickness, n (%)	7 (21)
History of migraines, n (%)	9 (27)

SD, standard deviation.

Changes in FEMs and predictors of recovery

Center task

A summary of the results for the center task is provided in Table 2. Significant changes between subacute and clearance were supported for average microsaccade amplitude in each trial (mean difference = 0.119; p = 0.005; Cohen's d = 0.53), average peak velocity of microsaccade in each trial (mean difference = 6.44; p = 0.01; Cohen's d = 0.48), average peak acceleration of microsaccade in each trial (mean difference = 594; p = 0.01; Cohen's d = 0.48), average duration of microsaccade in each trial (mean difference = 0.003; p < 0.001; Cohen's d = 0.72), and average drift vertical in each trial (mean difference = 0.019; p = 0.017; Cohen's d = -0.154). The multiple linear regression model to predict days to recovery (Table 3) with FEMs metrics was statistically significant (R² = 0.37; p = 0.023) and retained two predictors, average instantaneous drift amplitude (β = 0.547, p = 0.009) and average peak acceleration of microsaccade (β = 0.414, p = 0.089).

Table 2.

Results of Dependent T Test Fixational Eye Movement Statistics in Concussed Participants on Center Raster Task

Task		Fixation on center raster (n = 33)
Variable	Visit 1 mean (SD)	Visit 2 mean (SD)	Mean difference	T	df	p Value
Saccades (n)	23.29 (11.10)	25.32 (13.63)	-2.02 (7.43)	-1.566	32	0.127
Mean saccade amplitude (degrees)	1.02 (.35)	0.90 (0.36)	0.119 (0.224)	3.041	32	0.005^*
Average saccade direction (radians)	0.82 (1.46)	0.95 (1.38)	-0.133 (1.096)	-0.701	32	0.489
Average microsaccade peak velocity (degrees/sec)	61.05 (20.77)	54.61 (21.23)	6.44 (13.54)	2.733	32	0.010^*
Average Microsaccade peak acceleration (degrees/sec²)	5514 (1892)	4919.55 (1929)	594 (1247)	2.739	32	0.010^*
Average microsaccade duration (sec)	0.064 (0.006)	0.061 (0.005)	0.003 (0.005)	4.109	32	<0.001^*
Mean drift horizontal (regression slope)	0.1393 (0.11)	0.153 (0.15)	-0.013 (0.087)	-0.884	32	0.383
Mean drift vertical (regression slope)	0.06 (0.04)	0.08 (0.05)	-0.019 (0.044)	-2.516	32	0.017^*
Drift amplitude (degrees)	4.2 (2.76)	4.83 (3.39)	-0.63 (2.18)	-1.673	32	0.104
Standard error vertical	0.32 (0.16)	0.35 (0.26)	0.040 (.174)	1.308	32	0.200
Standard error horizontal	0.35 (0.26)	0.32 (0.24)	0.033 (0.161)	1.176	32	0.248
Drift proportion	0.86 (0.07)	0.84 (0.07)	0.014 (0.065)	1.258	32	0.218
BCEA	1.41 (0.88)	1.19 (0.79)	0.223 (0.729)	1.762	32	0.088
Average instantaneous drift speed	0.95 (0.20)	1.00 (0.30)	-0.051 (0.271)	-1.077	32	0.289
Average instantaneous drift direction	-0.67 (1.76)	-1.09 (1.67)	0.413 (1.17)	2.020	32	0.052
Average instantaneous drift amplitude	0.002 (0.0004)	0.002 0.0006)	-0.0001 (0.0006)	-1.124	32	0.269

All statistics reflected as standard deviation (SD).

Denotes significant change over time (paired samples t test).

BCEA, bivariate contour ellipse area.

Table 3.

Predicting Days to Recovery (Center Task)

Predictor	Unstandardized coefficients		Standardized coefficients	t	p
Predictor	B	Standard error	Beta	t	p
(Constant)	-87.515	43.211		-2.002	0.059
Average microsaccade peak acceleration	0.011	0.006	0.414	1.789	0.089
Average instantaneous drift amplitude	59,790.708	20,733.480	0.547	2.884	0.009

Dependent variable: clearance days from injury.

Corner task

Drift proportion (mean difference = 0.03; p = 0.03; Cohen's d = 0.43) significantly changed over time and demonstrated medium effect sizes. There were no other significant changes between FEMs variables across time (Table 4). The multiple linear regression model to predict days to recovery using FEMs metrics at time of clearance was statistically significant (R² = 0.85; p = 0.004) and retained nine predictors including average amplitude of microsaccades (β = 2.66; p = 0.001), average peak velocity of microsaccades (β = -15.11; p = 0.001), average peak acceleration of microsaccades (β = 12.56; p = 0.001), average drift horizontal (β = 7.95; p = 0.002), average drift vertical (β = 1.29; p = 0.011), drift amplitude (β = -8.34; p = 0.003), drift proportion (β = 0.584; p = 0.013), average instantaneous drift direction (β = -0.26; p = 0.083), and average instantaneous drift amplitude (β = 0.819; p = 0.002) (Table 5).

Table 4.

Results of Dependent t Test Fixational Eye Movement Statistics in Concussed Participants On Corner Raster Task

Task		Fixation on corner raster (n = 28)
Variable	Visit 1 mean (SD)	Visit 2 mean (SD)	Mean difference	t	df	p Value
Saccades (n)	32.68 (16.61)	35.56 (15.92)	-2.88 (13.17)	-1.157	27	0.258
Mean saccade amplitude (degrees)	0.477 (0.187)	0.480 (.174)	-0.004 (0.129)	-0.150	27	0.882
Average saccade direction (radians)	1.15 (1.28)	1.19 (1.19)	-0.037 (0.341)	-0.568	27	0.575
Average microsaccade peak velocity (degrees/sec)	28.88 (10.69)	29.03 (10.56)	-0.155 (7.98)	-0.103	27	0.919
Average microsaccade peak acceleration (degrees/sec²)	2572 (972)	2572 (966)	-0.85 (714)	-0.006	27	0.995
Average microsaccade duration (sec)	0.067 (0.008)	0.069 (0.021)	-0.002 (0.022)	-0.486	27	0.631
Mean drift horizontal (regression slope)	0.128 (0.128)	0.159 (0.167)	-0.032 (0.087)	-1.947	27	0.062
Mean drift vertical (regression slope)	0.048 (0.031)	0.052 (0.040)	-0.0039 (0.038)	-0.543	27	0.592
Drift amplitude (degrees)	3.68 (3.00)	3.94 (3.53)	-0.36 (1.77)	-1.077	27	0.291
Standard error vertical	0.242 (0.141)	0.284 (0.313)	-0.042 (0.275)	-0.804	27	0.428
Standard error horizontal	0.297 (0.288)	0.348 (0.308)	-0.051 (0.359)	-0.750	27	0.459
Drift proportion	0.83 (0.067)	0.796 (0.079)	0.034 (0.078)	2.277	27	0.015^*
BCEA	0.268 (0.242)	0.28 (0.214)	-0.011 (0.199)	-0.309	27	0.380
Average instantaneous drift speed	0.888 (0.226)	0.989 (0.36)	-0.101 (0.309)	-1.737	27	0.094
Average instantaneous drift direction	-1.31 (1.25)	-1.58 (0.71)	0.272 (1.077)	1.338	27	0.192
Average instantaneous drift amplitude	0.0018 (0.00047)	0.0021 (0.00075)	-0.00022 (0.00065)	-1.786	27	0.085

All statistics reflected as standard deviation (SD).

Denotes significant change over time (paired samples t test).

BCEA, bivariate contour ellipse area.

Table 5.

Predicting Days to Recovery (Corner Task)

Predictor	Unstandardized coefficients		Standardized coefficients	t	p
Predictor	B	Standard error	Beta	t	p
(Constant)	-284.161	98.165		-2.895	0.016
Mean saccade amplitude (degrees)	857.452	188.685	2.656	4.569	0.001
Average microsaccade peak velocity (degrees/sec)	-83.963	16.403	-15.109	-5.119	<0.001
Average microsaccade peak acceleration (degrees/sec²)	0.745	0.164	12.555	4.534	0.001
Mean drift horizontal (regression slope)	2559.155	635.767	7.949	4.025	0.002
Mean drift vertical (regression slope)	1841.317	587.271	1.289	3.135	0.011
Drift amplitude (degrees)	-119.564	30.780	-8.335	-3.885	0.003
Drift proportion	331.988	109.376	0.584	3.035	0.013
Average instantaneous drift direction	-9.924	5.158	-0.261	-1.924	0.083
Average instantaneous drift amplitude	68,862.516	16,301.966	0.819	4.224	0.002

Dependent variable: clearance days from injury.

Correlations among FEMs and concussion outcomes

Center task

Average saccade amplitude correlated with VOMS HSAC nausea (r = 0.317; p = 0.041), VOMS VSAC nausea (r = 0.317; p = 0.041), VOMS NPC nausea (r = 0.317, p = 0.041), and VOMS HVOR nausea (r = 0.326; p = 0.035). Average saccade direction correlated with baseline fogginess (r = -0.342, p = 0.027). Average peak velocity correlated with VOMS HSAC nausea (r = 0.306; p = 0.048), VOMS VSAC nausea (r = 0.306; p = 0.048), NPC nausea (r = 0.306; p = 0.048), HVOR nausea (r = 0.315; p = 0.042), and VMS nausea (r = 0.307; p = 0.048). Average peak acceleration correlated with HSAC nausea (r = 0.324; p = 0.036), HS total (r = 0.309, p = 0.046), VSAC nausea (r = 0.324; p = 0.036), VSAC total (r = 0.313; p = 0.043), HVOR nausea (r = 0.332; p = 0.032), VVOR nausea (r = 0.321; p = 0.038), VMS nausea (r = 0.324; p = 0.037), and NPC nausea (r = 0.324; p = 0.036). Mean drift vertical correlated with ImPACT visual motor speed (r = 0.337; p = 0.029), ImPACT visual memory (r = 0.443; p = 0.003), and ImPACT verbal memory (r = 0.38; p = 0.013). Average instantaneous drift speed correlated with HVOR headache (r = 0.311; p = 0.045); VVOR headache (r = 0.359; p = 0.020), ImPACT visual memory (r = 0.484; p = 0.001), and VMS headache (r = 0.347; p = 0.024). Average instantaneous drift direction correlated with ImPACT impulse control (r = -0.338, p = 0.028), VMS fogginess (r = 0.321; p = 0.038), and VVOR fogginess (r = 0.319; p = 0.039). Average instantaneous drift amplitude correlated with NPC headache (r = 0.307, p = 0.048), HVOR headache (r = 0.310; p = 0.045), VVOR headache (r = 0.357; p = 0.020), VMS headache (r = 0.352; p = 0.022), and ImPACT visual memory (r = 0.484, p = 0.001). There were no significant correlations between NPC measurement and FEMs metrics.

Corner task

Average saccade amplitude correlated with ImPACT reaction time (r = 0.476; p = 0.005), total nausea (r = 0.352; p = 0.044), VMS nausea (r = 0.466; p = 0.006), and VVOR nausea (r = 0.347; p = 0.048). Average peak velocity correlated with VVOR nausea (r = 0.347; p = 0.048) and VMS nausea (r = 0.445; p = 0.010), and ImPACT reaction time (r = 0.481; p = 0.005). Average peak acceleration correlated with VMS nausea (r = 0.427; p = 0.013). Average duration of saccades correlated with baseline dizziness (r = 0.393; p = 0.024) and smooth pursuits dizziness (r = 0.420; p = 0.015). Mean drift horizontal correlated with ImPACT impulse control (r = 0.361; p = 0.039). Drift amplitude correlated with NPC average (r = 0.376; p = 0.031). Average instantaneous drift speed correlated with ImPACT visual memory (r = 0.419; p = 0.015). Average instantaneous drift direction correlated with ImPACT impulse control (r = -0.400; p = 0.021). Average instantaneous drift amplitude correlated with ImPACT visual memory (r = 0.43; p = 0.012). Average NPC measurement across trials was negatively correlated with drift amplitude (r = -0.410; p = 0.018); however, no further correlations between FEMs metrics and NPC measurements were found.

Discussion

This is the first study to examine changes in FEMs using TSLO across time from the initial clinic visit (within 21 days of injury) to the date of medical clearance (i.e., clinical recovery) following concussion. Our hypotheses that FEMs metrics would significantly change from initial evaluation to clinical recovery, and that FEMs metrics at the initial evaluation would be useful in predicting clinical recovery, were largely supported. Significant differences were more consistently observed in the center fixation task than in the corner task, but the corner task had more predictors associated with recovery duration. Further, FEMs metrics were correlated with several clinical assessment tools that measure ocular motor symptoms (e.g., nausea, headache) and impairments (e.g., cognitive test performance) at the initial evaluation, although the effect sizes trended small to medium.

Multiple FEMs metrics significantly changed from initial concussion clinic visit (i.e., <21 days post-injury) to clinical recovery visit. Specifically, changes in average saccade amplitude, average peak velocity and acceleration of saccades, average duration of saccades, and mean drift vertical when completing the center raster task were observed. Drift proportion also significantly decreased on the corner raster task. Given our lack of a control group for comparison, it is not entirely clear if these changes reflect authentic improvement and/or potential return to baseline functioning. Nonetheless, previous researchers described that drift dynamics may be associated with mental fatigue,²³ and these findings are reflective of the recovery process of ocular dysfunction while being concomitant with resolution of associated symptomology. Alternatively, there could be a reversed directionality of the effect, particularly when considering the possibility that improved mental stamina secondary to concussion recovery could impact drift performance. In any case, our findings are consistent with previous research suggesting that increased amplitude is linked to ocular dysfunction,²⁴ and the decreasing amplitude observed in the study would suggest that improvement is occurring.

Several FEMs metrics were significant in predicting recovery duration, especially for the corner raster task. Average saccade amplitude, average peak velocity of microsaccade, average peak acceleration of microsaccade, mean drift horizontal, mean drift vertical, drift amplitude, drift proportion, and average instantaneous drift amplitude were all found to be significant predictors of recovery. For the center raster task, average instantaneous drift amplitude was a significant predictor of overall recovery. The discrepancy in findings for corner versus center tasks was not surprising given findings from our previous report of significant differences between concussed and controls being observed only for the center fixation task. We speculated that this may be because of the differential effect of concussion on the neural encoding of fixation position for large versus small targets.²¹ However, more work is needed to understand the mechanisms driving changes in fixation following concussion and recovery.

Interestingly, average instantaneous drift amplitude was the only metric that maintained significance as a predictor of recovery time for both tasks. Despite this, its value did not significantly change within concussed subjects from initial visit to clearance. Immediate implications for this finding would suggest that instantaneous drift, rather than being a metric specifically impacted by concussion, might be associated with pre-existing conditions (i.e., potentially related to extraneous pre-injury variables), which is associated with a longer recovery. This implication is further supported by our group's prior work, in which there were no significant differences between concussed and controls on this metric. Prior research has established a risk factor model for concussion outcomes, including protracted recovery, which has been supported for other clinical profiles/subtypes, such as migraine and mood profiles, and this could apply to this metric and pre-injury ocular dysfunction as well.^25,26 Further research is needed to better understand its relationship with concussion.

Our results revealed several relationships between FEMs metrics and clinical measure scores and symptom burden. Our most consistent finding was the association between nausea symptom intensity on the VOMS and FEMs metrics. Specifically, nausea was associated with symptoms on saccades screening. This relationship was also present for vestibular screening items, which are commonly associated with nausea,^27,28 and highlights the complex relationship between vestibular and ocular systems. There were also medium-size relationships between neurocognitive testing and FEMs metrics. A majority of drift metrics, which were correlated with visual memory performance on ImPACT testing, further suggest that ocular weaknesses might be associated with memory performance. This finding is consistent with prior literature, which has found ocular disruption to be related to worse scores on neurocognitive testing,^29,30 with some research purporting that the effects of ocular disruption on neurocognitive testing are more impactful, particularly in the context of concussion.³¹ Surprisingly, just one metric (mean drift vertical, center task) correlated with visual motor speed. Additionally, there were just two metrics (average amplitude of microsaccade, corner task, and average peak velocity of microsaccade, corner task) that correlated with reaction time. Conjecturally speaking, the comparatively more robust findings of correlations between FEMs metrics and visual memory could be caused by task similarity (i.e., a major component of visual memory requires individuals to fixate on and memorize designs in the absence of a motor demand), whereas visual motor speed and reaction time both include strong components of motor demands to complete. Future research is required to validate this claim, although the correlations we found warrant further exploration of the relationship between ocular metrics and neurocognitive performance in the context of concussion.

In comparing current clinical measures of ocular performance (i.e., VOMS) to the current measures of FEMs metrics throughout recovery, there were no associated correlations with metrics and measurement of NPC other than a negative correlation between drift amplitude and average NPC measurement across trials on the corner raster task. It is not entirely surprising that there were not significant correlations between FEMS metrics and NPC measurements given the discrepancy between the tasks themselves and differences between the neurophysiological demands of the tasks.^32
–34 Nonetheless, this lack of correlation, in conjunction with the fact that metrics did significantly change over time, reflects that one of the primary clinical measurements of ocular disruption might not be fully detecting more subtle changes in ocular disruption during the recovery process. In other words, the objective measurement of NPC and other current measurements used in clinic could benefit from more high specialized instruments such as the TSLO, to more efficiently detect the more subtle changes in ocular performance during the concussion recovery process.

Limitations

Although this study was the first to examine FEMs across recovery following concussion, there are several limitations to the approach. The sample size of this study was small, and some participants were lost to follow-up. Future studies would benefit from examining a larger sample of concussion patients and from including a control group with multiple testing visits rather than simply two time points. Adolescents and young adults, as well as males and females, were combined for analysis in this study, thereby obfuscating potential age and sex group differences. Researchers should target specific age groups and compare the FEMs findings to determine if there are any age effects. Additionally, the inclusion criteria of being within 21 days of injury may have resulted in the inclusion of participants whose ocular impairments were already resolved. Prior research has suggested that the majority of ocular impairments may resolve within 14 days post-injury.³⁵ Future research should focus on participants in the acute (within 72 h) stage of injury and include multiple assessments of FEMs metrics across recovery. Another limitation is the preliminary nature of the study. Whereas we sought to identify the potential utility of tracking ocular metrics during recovery, other assessment tools have robust empirical support for their use in diagnosing concussion and for their relationship with other factors of recovery such as symptom burden and return to play decision making. Establishing the measurement of ocular metrics for diagnostic purposes and clinical decision making was outside the purview of this article. The authors further acknowledge that future research with a larger data set is required to help establish the utility of ocular metrics for these purposes, for replication by other clinicians, and for eventual clinical use.

Conclusion

The current findings provide preliminary evidence for improvements in FEMs metrics across recovery following concussion. These findings support the potential clinical utility of a TSLO-based assessment of FEMs to augment current tools by providing a more objective assessment to track recovery of ocular function following this injury. In addition, the FEMs metrics were correlated to several clinical measures of ocular motor function, which supports concurrent validity with these measures. However, given that most of these relationships were moderate in magnitude, FEMs may be more sensitive to the subtle effects of concussion, or simply measure different aspects of ocular function than do commonly employed clinical measures, which focus on saccadic function and binocular vision skills. Given that FEMs are involved in spatial vision, we can speculate that changes in FEMs may coincide with improvements in visual performance on everyday tasks that require fine spatial vision such as reading and driving, which is a commonly reported problem following concussion. Further, recovery of FEMs metrics may coincide with improvement in complex coordinated visuo-motor behaviors post-concussion.³⁶ Moving forward, researchers should evaluate FEMs in the context of targeted treatment approaches, and consider other factors including age, sex, and ocular history, which may influence the current findings.

Transparency, Rigor, and Reproducibility Summary

This study was not formally registered because it was a follow-up to a prior study. The analysis plan was not formally pre-registered, but the team member with primary responsibility for the analysis certifies that the analysis was pre-specified. A sample size of 50 concussed subjects was planned based on availability of data from the prior study. Actual sample size was 33 subjects and effect sizes were medium. Data analyses were performed by investigators blinded to relevant characteristics of the participants. Data were collected following completion of clinical treatment and ultimate recovery from concussion. Specific equipment used to perform acquisition of data is not publicly available; however, specific software used to perform analyses is widely available from IBM. The key inclusion criteria are established standards in the field and the primary clinical outcome measure is an emerging standard in the field. Key inclusion criteria and clinical outcomes were assessed by investigators with extensive experience in the field. The sample size and degrees of freedom reflects the number of independent measurements. Both the original measures of statistical error rates and corrected measures of statistical error rates have been accounted for in the data. No replication or external validation studies have been performed or are planned at this time, to our knowledge. De-identified data from this study are not available in a public archive. De-identified data from this study will be made available (as allowable according to institutional review board [IRB] standards) by e-mailing the corresponding author upon acceptance. Analytic code used to conduct the analyses presented in this study are not available in a public repository. They may be available by e-mailing the corresponding author as of acceptance. The authors agree to provide the full content of the manuscript on request by contacting the corresponding author.

Footnotes

Acknowledgments

The authors thank the following researchers for their work in our preceding study and for the collection and storage of data which made this study possible: Bianca T. Leonard, Gregory F. Marchetti, Min Zhang, Hope M. Reecher, Ethan S. Bensinger, Valerie Snyder, Christy Sheehy, and Cyndi Holland.

Authors' Contributions

The authors contributed as follows: Ted J. Albrecht was responsible for conceptualization, statistical analysis, and writing – original draft; Bindal Makwana Mehmel was responsible for writing – review and editing; Ethan A. Rossi was responsible for conceptualization, investigation, writing – review and editing, and funding acquisition; Alicia M. Trbovich was responsible for writing - review and editing; Shawn R. Eagle was responsible for statistical analysis and methodology; and Anthony P. Kontos was responsible for conceptualization, writing – abstract, review and editing, and supervision. Bianca T. Leonard, Gregory F. Marchetti, Min Zhang, Hope M. Reecher^, Valerie Snyder, and Cyndi L. Holland were participating investigators.

Funding Information

This research was supported by a grant from the National Institute of Neurological Disorders and Stroke (R44NS095090) to C. Light Technologies and the University of Pittsburgh and by a grant from C. Light Technologies, Inc., to EAR and by departmental startup funds from the University of Pittsburgh to EAR. This project was also supported by the University of Pittsburgh Clinical and Translational Science Institute, through grants from the National Institutes of Health (UL TR001857, KL2 TR001856, TL1 TR001858). This work was also supported by a National Institutes of Health Core Grant (P30 EY08098) to the University of Pittsburgh Department of Ophthalmology and an unrestricted grant to the University of Pittsburgh Department of Ophthalmology from Research to Prevent Blindness.

Author Disclosure Statement

No competing financial interests exist. Dr. Kontos receives book royalties from APA Books, and funding for his research through the University of Pittsburgh from the Centers for Disease Control and Prevention, Chuck Noll Foundation for Brain Injury Research, Department of Defense (CDMRP, USAMRAA, USUHS), National Football League, National Institutes of Health (NICHD, NIMH, NINDS), and private donors.

References

Langlois

, Rutland-Brown

, & Wald

. The epidemiology and impact of traumatic brain injury: a brief overview. J Head Trauma Rehabil, 2006; 21(5):375–378; doi: 10.1097/00001199-200609000-0001

Merezhinskaya

, Mallia

, Park

, et al. Visual deficits and dysfunctions associated with traumatic brain injury: a systematic review and meta-analysis. Optom Vis Sci, 2019; 96(8):542–555; doi: 10.1097/OPX.0000000000001407

Scheiman

, Grady

, Jenewein

, et al. Frequency of oculomotor disorders in adolescents 11 to 17 years of age with concussion, 4 to 12 weeks post injury. VisionRes, 2021; 183:73–80; doi: 10.1016/j.visres.2020.09.011

Santo

, Race

, Teel

. Near point of convergence deficits and treatment following concussion: a systematic review. J Sport Rehabil, 2020; 29(8):1179–1193; doi: 10.1123/jsr.2019-0428

Master

, Scheiman

, Gallaway

, et al. Vision diagnoses are common after concussion in adolescents. Clin Pediatr (Phila), 2016; 55(3):260–267; doi: 10.1177/0009922815594367

Maruta

, Spielman

, Rajashekar

, et al. Association of visual tracking metrics with post-concussion symptomatology. Front Neurol, 2018; 9:611; doi: 10.3389/fneur.2018.00611

Sussman

, Ho

, Pendharkar

, et al. Clinical evaluation of concussion: the evolving role of oculomotor assessments. Neurosurg Focus, 2016; 40(4):E7; doi: 10.3171/2016.1.FOCUS15610

Whitney

, Eagle

, Marchetti

, et al. Association of acute vestibular/ocular motor screening scores to prolonged recovery in collegiate athletes following sport-related concussion. Brain Inj, 2020; 34(6):842–847; doi: 10.1080/02699052.2020.1755055

DuPrey

, Webner

, Lyons

, et al. Convergence insufficiency identifies athletes at risk of prolonged recovery from sport-related concussion. Am J Sports Med, 2017; 45(10):2388–2393; doi: 10.1177/0363546517705640

10.

Moran

, Covassin

, Elbin

, et al. Reliability and normative reference values for the vestibular/ocular motor screening (VOMS) tool in youth athletes. Am J Sports Med, 2018; 46(6):1475–1480; doi: 10.1177/0363546518756979

11.

Patel

, LaBella

. Contributions of PCSS, CISS, and VOMS for identifying vestibular/ocular motor deficits in pediatric concussions. Sports Health, 2021; 13(6):565–572; doi: 10.1177/1941738121994116

12.

Yorke

, Smith

, Babcock

, et al. Validity and reliability of the vestibular/ocular motor screening and associations with common concussion screening tools. Sports Health, 2017; 9(2):174–180; doi: 10.1177/1941738116678411

13.

Kontos

, Sufrinko

, Elbin

, et al. Reliability and associated risk factors for performance on the Vestibular/Ocular Motor Screening (VOMS) tool in healthy collegiate athletes. Am J Sports Med, 2016; 44(6):1400–1406; doi: 10.1177/0363546516632754

14.

Knell

, Caze

, Burkhart

. Evaluation of the vestibular and ocular motor screening (VOMS) as a prognostic tool for protracted recovery following paediatric sports-related concussion. BMJ Open Sport Exerc Med, 2021; 7(1), e000970; doi:10.1136/bmjsem-2020-000970

15.

, Ortega

, Chrisman

, et al. King-Devick sensitivity and specificity to concussion in collegiate athletes. J Athl Train, 2021; 58(2):97–105; doi: 10.4085/1062-6050-0063.21

16.

Worts

, Schatz

, Burkhart

. Test performance and test-retest reliability of the vestibular/ocular motor screening and King–Devick test in adolescent athletes during a competitive sport season. Am J Sports Med, 2018; 46(8):2004–2010; doi:10.1177/0363546518768750

17.

Mucha

, Collins

, Elbin

, et al. A brief vestibular/ocular motor screening (VOMS) assessment to evaluate concussions: preliminary findings. Am J Sports Med, 2014; 42(10):2479–2486; doi: 10.1177/0363546514543775

18.

Martinez-Conde

, Otero-Millan

, Macknik

. The impact of microsaccades on vision: towards a unified theory of saccadic function. Nat Rev Neur, 2013; 14(2):83–96; doi: 10.1038/nrn3405

19.

Chatterjee

, Gavas

, Chakravarty

, et al. Eye movements–an early marker of cognitive dysfunctions. Conf Proc IEEE Eng Med Biol Soc, 2018:4012–4016; doi: 10.1109/EMBC.2018.8513307

20.

Hammer

, Ferguson

, Magill

, et al. Compact scanning laser ophthalmoscope with high-speed retinal tracker. Appl Opt, 2003; 42(22)4621–4632

21.

Leonard

, Kontos

, Marchetti

, et al. Fixational eye movements following concussion. J Vis, 2021; 21(13):11–11; doi: 10.1167/jov.21.13.11

22.

McCrory

, Meeuwisse

, Dvorák

, et al. Consensus statement on concussion in sport–the 5th international conference on concussion in sport held in Berlin, October 2016. Br J Sports Med, 2017; 51(11):838–847; doi: 10.1136/bjsports-2017-097699.

23.

Di Stasi

, McCamy

, Catena

, et al. Microsaccade and drift dynamics reflect mental fatigue. Eur J Neurosci, 2013; 38(3):2389–2398; doi: 10.1111/ejn.12248

24.

Alexander

, Macknik

, Martinez-Conde

. Microsaccade characteristics in neurological and ophthalmic disease. Front Neurol, 2018; 9:144; doi: 10.3389/fneur.2018.00144

25.

Iverson

, Williams

, Gardner

, et al. Systematic review of preinjury mental health problems as a vulnerability factor for worse outcome after sport-related concussion. Orthop J Sports Med, 2020; 8(10):2325967120950682; doi: 10.1177/2325967120950682

26.

Terry

, Huebschmann

, Maxwell

, et al. Preinjury migraine history as a risk factor for prolonged return to school and sports following concussion. J Neurotrauma, 2019; 36(1):142–151; doi: 10.1089/neu.2017.5443

27.

Turgut

, Turgut

. Neuro-ocular vestibular dysfunction: an overlooked illness. Adv Ophthalmol Vis Syst, 2017; 7(7):00255; doi: 10.1540/6/aovs.2017.07.00255

28.

Cohen

, Dai

, Yakushin

, et al. The neural basis of motion sickness. J Neurophysiol, 2019; 121(3):973–982; doi: 10.1152/jn.006474.2018

29.

Pearce

, Sufrinko

, Lau

, et al. Near point of convergence after a sport-related concussion: measurement reliability and relationship to neurocognitive impairment and symptoms. Am J Sports Med,, 2015; 43(12):3055–3061; doi: 10.1177/0363546515606430

30.

Arsiwala

, Guo

, Ramulu

, et al. Associations of visual function with cognitive performance in community-based older adults: the Eye Determinants of Cognition study. J Gerontol A Biol Sci Med Sci, 2022; 77(10):2133–2140; doi: 10.1093/gerona/glab349

31.

Peiffer

, MacDonald

, Duerson

, et al. The influence of binocular vision symptoms on computerized neurocognitive testing of adolescents with concussion. Clin Pediatr (Phila), 2020; 59(11):961–969; doi: doi.org/10.1177/0009922820927477

32.

Searle

, Rowe

. Vergence neural pathways: a systematic narrative literature review. Neuroophthalmology, 2016; 40(5):209–218; doi: 10.1080/01658107.2016.1217028

33.

Rucci

, Poletti

. Control and functions of fixational eye movements. Annu Rev Vis Sci, 2015; 1:499; doi: 10.1146/annurev-vision-082114-035742

34.

Krauzlis

, Goffart

, Hafed

. Neuronal control of fixation and fixational eye movements. Philos Trans R Soc Lond B Biol Sci, 2017; 372(1718):20160205; doi: 10.1098/rstb.2016.0205

35.

Elbin

, Sufrinko

, Anderson

, et al. Prospective changes in vestibular and ocular motor impairment after concussion. J Neurol Phys Ther, 2018; 42(3):142; doi: 10.1097/NPT.0000000000000230

36.

Alexander

, Macknik

, Martinez-Conde

. Microsaccades in applied environments: real-world applications of fixational eye movement measurements. J Eye Mov Res, 2019; 12(6); doi: 10.1691/0/jemr.12.6.15