Individual and Combined Effects of Sport-Related Concussion and Anterior Cruciate Ligament Injury on Neurocognitive and Neuromechanical Reaction Time

A total of 47 college individuals (age = 20.49 ± 1.3 years), consisting of 40 (85%) females and 7 (15%) males, participated in this study. Participants were categorized into 1 of 4 groups, based on their prior medical history: (1) a history of SRC, (2) a history of ACL tear and ACLR, (3) history of both SRC and ACLR (SRC + ACLR), and (4) a control group with neither SRC, nor ACL injury, nor any other lower extremity musculoskeletal injury history. Participants were additionally categorized into injury history consisting of any SRC (SRC and SRC + ACLR groups) and no SRC (ACLR and controls) history, and any ACLR (ACLR and SRC + ACLR) and no ACLR (SRC and controls) history groups.

Inclusionary criteria for this study consisted of individuals with any of the above individual or combined injury histories from participation in sport that were medically cleared by a physician for full return to sport, physical activity without restrictions, or both, along with ACL injury resulting from a noncontact mechanism. Time from SRC, ACLR, or both, was set within 7 years. This range was selected to include up to seniors in college who had ACLRs in high school. Exclusionary criteria consisted of those with a history of any foot, ankle, or hip surgeries. Individuals with a history of ACLR were excluded if their mechanism of injury was from direct contact, as opposed to noncontact. Participants were also excluded if they reported with any modifiers for concussion assessment, including individuals with attention or learning disorders; headache or migraine disorder; and depression, anxiety, or other psychiatric disorders; as well as any individuals that had diagnosed cognitive impairment or were still undergoing therapy or rehabilitation for any injury or pathology in the 6 months before testing.^21–25  Consistency between the number of males between groups was considered during the recruitment and enrollment stages of the study to eliminate any potential sex differences that may occur. Human subjects institutional review board approval was granted, and informed consent was obtained before testing. Participants were recruited through the university email servers as well as campus recreation athletic training services and were compensated monetarily for their participation in the study.

Participants reported to the athletic training research laboratory for a 1-time data collection session. After informed consent, participants completed a demographic questionnaire that inquired about age, biological sex, and injury history. Injury history consisted of self-reporting binary (yes or no) data to either prior SRC, ACLR, or both. Additional injury demographics included the previous number of each respective SRC, ACLR or both; time since the most recent SRC, ACLR, or both; and graft type (ACL only).

The Cleveland Clinic Concussion Application (C3 Logix) was used for the tablet-based neurocognitive battery, consisting of the Trail Making Test (TMT), simple RT, choice RT, and processing speed.²⁶  The TMT consists of 2 sections: Part A and B. Part A requires the test taker to connect 25 dots labeled 1–25 in ascending order (eg, 1 → 2 → 3). This engages spatial scanning and executive and neuromotor functions. Part B increases in difficulty, requiring the test taker to connect 25 dots labeled 1–13 and A–L, in which he or she alternates in switch-setting fashion (eg, 1 → A → 2 → B → 3 → C). This engages the same cognitive functions as Part A with the addition of switch setting. A final score, TMT B − A, is calculated. The second component of the C3 Logix battery was simple and choice RT tests. For simple RT, test takers were instructed to place their dominant finger on the designated box on the screen, and when the green circle appears in the middle of the screen, they lifted their dominant finger and tapped the circle as quickly as possible. Choice RT is completed with both fingers touching designated boxes on the left and right side at the bottom, center of the screen. A green circle and a blue circle (distractor) will both appear, to which the test taker lifts the finger on the side of the green circle and taps that green circle as quickly as possible, disregarding the blue circle. Both simple and choice RTs are completed for 25 trials to establish RT in milliseconds. The final component from the battery was a processing speed test, in which the test taker completed a digitized symbol and digit matching exercise over 2 minutes, producing the number of correct symbols and digit matches completed. The TMT A and B and symbol-digit match tests on the C3 Logix have demonstrated concurrent validity with the standard pen-and-paper versions and moderate to strong test-retest reliability.²⁶  The neurocognitive battery was completed first to negate possible improvements in neurocognitive RT after physical activity; however, all subtests within each battery were counterbalanced.²⁷  All assessments in the neurocognitive battery included a practice trial first, followed by the recorded trial.

The TRAZER technology was used for the software-based neuromechanical battery that combines immersive simulation, motion capture analytics, and RT-based challenges of simple, Flanker, and Stroop tasks.²⁸  The battery requires the test taker to stand approximately 10 feet away from the TRAZER system, in which a human-simulated character is reproduced on the screen, mirroring motion capture of his or her body. During the simple neuromechanical task, test takers watch the screen for a series of bumpers to appear in a 2 × 2 box, which instructs the individual to move either left, right, forward, backward, or diagonally to a combination of the prior 4 directions into the bumper and return to the middle of the box. This assessment is then repeated for choice RT measures of the Flanker and Stroop tasks to test RT to congruent and incongruent cues. The Flanker task uses left or right movements only. The screen flashes a series of 5 arrows in the middle of the screen, with arrows facing either similar (ie, → → → → →) or different (ie, ← ← → ← →) directions from each other, either left or right. The test taker is instructed to move to the bumper on the side of the box where the middle arrow faced and then return to the middle of the box. Participants then completed the Stroop task, which is like the Flanker task, but the stimuli consisted of a word that is a color (ie, red, blue, green, or purple). The color of the word may have matched or differed from the word that appeared on the screen. The test taker is instructed to move to the left or right bumper, whichever coincides with the word, not the color of the word. Three rings above the bumpers are coded with correct and incorrect stimuli. For example, if the word of the screen was “blue” and colored green, the test taker moved to the side with the bumper that had 3 blue rings over it and disregarded the side with 3 green rings. At times, the word and color may match, (eg, the word may be “red” and is also colored red). In this case, the test taker would move to the side with the red rings as the obvious correct answer and disregard the purple-ring bumpers. Single, Flanker, and Stroop tasks were completed for 1-minute each. The TRAZER RT test has demonstrated strong reliability for average velocity (interclass correlation coefficient = 0.84), acceleration (0.91), and deceleration (0.94).²⁸  All assessments in the neurocognitive and neuromechanical batteries included a practice trial first, followed by the recorded trial.

An a priori sample size power analysis with a 0.8 power and an 0.5 effect size⁹  was conducted to determine an adequate sample of 11 participants per group.⁹  General descriptive (ie, means, frequencies) and inferential statistics were used to summarize demographic variables and mean scoring. A series of Shapiro-Wilk tests were conducted to test for normality of all variables. All variables except neuromechanical simple RT were normally distributed (P = .027). A series on 1-way multivariate analyses of variance were used to examine differences between the SRC, ACLR, SRC + ACLR, and control groups on the TMT, neurocognitive RT, and on neuromechanical Flanker- and Stroop-task RTs. A Kruskal-Wallis H test was conducted to examine group differences on neuromechanical simple RT. A series of univariate analysis of variance tests were used to examine differences between any ACLR and no ACLR, along with any SRC and no SRC on all measures except neuromechanical simple RT, which was analyzed using a Mann-Whitney U test due to nonnormality. Due to nonnormality, median and interquartile range (IQR) were provided for only neuromechanical simple RT. To determine if time from injury was correlated with any dependent variables, a bivariate correlation was conducted between each dependent variable and the time from the last SRC (SRC group), time from the last ACLR (ACLR group), and time from the last injury, either SRC or ACLR (SRC + ACLR group). Strength of the correlation coefficients was interpreted as 0.00–0.30 = weak, 0.40–0.69 = moderate, 0.70–0.89 = strong, and 0.90 ≤ = very strong. A priori was set to P ≤ .05 for all analyses.

All 12 individuals in the SRC group reported sustaining multiple SRCs, with 5 (41%) reporting 2 previous SRCs, 3 (25%) reporting 3 SRCs, 2 (17%) reporting 4 SRCs, and 2 (17%) reporting 5 SRCs. In the ACLR group, 7 (58%) reported 1 prior ACL tear, while 4 (33%) reported 2 tears, and 1 (9%) reported 4 tears. Regarding limb side, 6 tore their left ACL, 3 tore their right, and 3 tore both. Two participants that tore their right and left ACLs had differing tendon grafts used (eg, 1 limb used hamstring graft, 1 used patellar). Aside from those 2 participants, the most common graft type was patellar (n = 6), followed by quadriceps (n = 2) and hamstrings (n = 1). In the SRC + ACLR group, 5 participants reported 1 prior SRC, while 2 each reported 2, 3, and 4 prior SRCs; 9 had 1 prior ACLR, and 2 had 2 prior ACLRs. Near even distribution was found between limb side, with 6 individuals tearing their right ACL and 5 tearing their left, while 5 had patellar tendon grafts, 4 had hamstring tendon grafts, 2 had quadricep tendon grafts (1 had quadricep and hamstring for multiple tears), and 1 had a gracilis/adductor tendon graft. A breakdown of age, sex, and full injury demographics by group are presented in Table 1.

No differences were noted between the SRC, ACLR, SRC + ACLR, and control groups on the TMT Part A (η² = 0.119, P = .13), Part B (η² = 0.143, P = .08), or B − A (η² = 0.151, P = .07) times (Table 2). Regarding neurological RT, no differences were observed between simple (P = .24) or choice (P = .93) RTs and processing speed (P = .14; Table 2). No differences were observed for neuromechanical RT, specifically during simple (H = 1.658, P = .64), Flanker (P = .99), and Stroop (P = .64) tasks (Table 2). Median (IQR) for simple RT consisted of SRC = 0.36 (0.09), ACLR = 0.40 (0.07), SRC + ACLR = 0.39 (0.04), and controls = 0.38 (0.08).

Significant differences were observed between groups for any SRC and any ACLR when compared with those without each respective injury history. Those with a previous SRC (SRC and SRC + ACLR) had a slower TMT B − A time (P = .03), which is the main outcome measure for that test. No differences were noted between Parts A (P = .96) and B (P = .08; Table 3). No differences were observed between any SRC and no SRC on neurocognitive simple (P = .15) and choice (P = .83) RTs, or processing speed (P = .055). No differences were observed for simple (U = 223.5, P = .26), Flanker (P = .95), or Stroop (P = .27) tasks on neuromechanical RT. Median (IQR) for simple RT consisted of any SRC = 0.37 (0.07) and no SRC = 0.40 (0.07).

Similar differences were apparent for the any ACLR group compared with the no ACLR group, with the any ACLR group producing worse Part A times (P = .02) on the TMT, despite no differences on Part B (P = .14) or B − A (P = .56; Table 3). No differences were observed for neurocognitive RT (P range = .21–.87), nor simple (U = 245.5, P = .51), Flanker (P = .87), or Stroop (P = .87) task RTs. Median (IQR) for simple RT consisted of any ACLR = 0.39 (0.07) and no ACLR = 0.37 (0.07).

A moderate, significant correlation was observed for time from the last SRC of the SRC group and Part B time (r = 0.440, P = .007) along with a moderate, significant correlation between the last SRC and B − A time (r = 0.581, P = .04), indicating that, as time from the last SRC increased, so did respective TMT time. Similar findings were noted with moderate, significant correlations between time from the last ACLR of the ACLR group and Part B time (r = 0.656, P = .02) and B − A time (r = 0.631, P = .03); however, a moderate, negative correlation was found with neurocognitive processing speed (r = −0.602, P = .04), indicating that, as time from the last ACLR increases, processing speed decreases. No significant correlations existed between time from the last injury, either SRC or ACLR, for the SRC + ACLR group and any dependent variables (r range = −0.526 to 0.560, P range = .07–.77).

Our findings suggest that, when comparing injury history groups of SRC, ACLR, a combination of SRC and ACLR, and previously uninjured controls, no differences exist on any battery items. Interestingly, when recategorizing participants based on history of any SRC, which contained the SRC and SRC + ACLR groups, slower times were noted on the TMT B − A score than those with no SRC history. Similar findings were noted on the TMT Part A only between any ACLR history and no history. In the previous literature, a history of concussion did not affect neurocognitive processing speed and RT, using the Immediate Post-Concussion and Assessment and Cognitive Testing (ImPACT) software.^29–31  This only partially supports our findings when comparing the SRC group to the ACLR and control groups and between concussion history groups (any SRC versus no SRC history) on RT only, when using C3 Logix neurocognitive assessment. Collins et al noted differences on the TMT between college football players with and without a history of concussion at baseline, using the pen-and-paper version, as opposed to the modern, tablet version.³²  Additionally, Collins et al showed that those players with multiple concussions scored the lowest on the TMT, which, given that all participants in our SRC group and roughly half in the SRC + ACLR group reported ≥2 previous concussions, this may indicate executive function and neuropsychological deficiencies for SRC and multiple SRC patients, 2+ years postinjury; however, future research is needed to confirm this hypothesis.³² 

It may be possible that all participants with a previously diagnosed SRC may have more negative effects of delayed processing. While no differences were observed between the ACLR group and the SRC, SRC + ACLR, and control groups on neurocognitive function, it was noted that those with a history of ACLR had clinically slower TMT Part A times than those without. This may be partially attributed to noncontact ACL injury having elements of visual search and processing speed noted as potentially contributing to the mechanism and noted deficits in ACLR patients.^33–35  The reduced ability to engage in a rapid cognitive task that has high spatial and visual search processing demands in addition to working memory may be secondary to the altered neural activity pattern seen after ACLR, including differences in sensory and visual integration brain regions.^36,37  While participants in our any ACLR group did not differ on neurocognitive processing speed, their clinically slower times on Part A may be attributed to the processing speed component of neuromotor function to connect the dots by processing scattered, numeric values (eg, 1 → 2 → 3) on the tablet as quickly as possible. Swanik et al used ImPACT, which at the time was only available as a computerized neurocognitive assessment using a mouse and keyboard interface, rather than a tablet-based tool like the C3 Logix.⁹  Differences in mouse versus touchscreen interfaces may yield differing results, as a maze-learning neurocognitive task was found to produce faster performance on an iPad.³⁸  Therefore, newer applications like the C3 Logix may be more sensitive to RT and processing speed due to screen latency.

The results from this current study found no group differences between the SRC, ACLR, SRC + ACLR, and uninjured control groups on simple RT and during Flanker and Stroop tasks. To date, no literature exists in which authors have examined combined effects of SRC and ACLR history on neuromechanical RT. While Swanik et al explored SRC screening tool performance after ACL injury, limited researchers have explored ACLR screening tools for SRC history.⁹  Authors of those few studies have examined drop biomechanics, as a measurement of dynamic motor control in those with and without a history of concussion, but these studies have yielded mixed findings.^39–41  Shumski et al noted no differences between college individuals with and without a history of concussion on ground reaction force, knee flexion and abduction angles, and landing rate, possibly explaining the lack of differences in this study between groups on neuromechanical RT.³⁹  Opposing findings were reported in adolescent athletes demonstrating no knee or dorsiflexion range of motion during a drop landing task nor single-legged squat in recreationally active college individuals.^40,41 

One metric that remains often unconsidered in injury-history studies is time since injury. In this study, the time since the last concussion or ACLR for both SRC, ACLR, and SRC + ACLR groups was approximately 3 years. While recovery time points vary across each injury and by both intrinsic and extrinsic factors, cognitive deficits and self-reported balance problems may persist up to 4 years postconcussion, with knee proprioception and postural stability deficits persisting up to 2 years post-ACLR.^42–46  In fact, some experts have suggested that medical clearance for full RTP post-ACLR should not be until 2 years postsurgery to allow full graft healing and joint health.⁴⁷  Therefore, it is possible that our participants may still be suffering knee- or brain-related physiological deficits, or both, despite medical clearance and unrestricted normal activity. We noted a significant correlation between time from the last SRC and Part B time, indicating that time from injury and recovery may contribute toward the findings. Moderate correlations were noted on variables of TMT performance for both time from the last SRC and time from the last ACLR. Moderate correlations were also observed between the last injury (SRC or ACLR) and neurocognitive and neuromechanical RT, all potentially influencing neurological function.

While neuromechanical RT differs from neurocognitive RT due to movement of the lower extremity and trunk, rather than the index finger, our findings may suggest the relationship of neurological RT in general, regardless of upper or lower extremity movement. The Flanker and Stroop tasks were incorporated into this study to serve as a neuromechanical choice RT task since participants had to respond to the congruent or incongruent stimuli before reacting, like the C3 Logix choice RT subtest, which noted a similar lack in differences between all groups in our study. Wilkerson et al recommended that clinical assessment of neurocognitive and neuromechanical function is needed to reduce the risk of lower extremity musculoskeletal injury, postconcussion.⁴⁸  These authors also provided a pathophysiological theoretical framework outlining the overlap of neurocognitive RT and functional brain activity via visual-kinesthetic disintegration quantified after concussion and lower extremity injury.⁴⁹  Evidence of an overlapping injury risk profile between concussion and ACL injury is growing, whereas the ability to engage in complex neuromechanical responses to environmental stimuli may influence both concussion and ACL injury risk.^50–54  This potential overlapping neural mechanism that drives neuromechanical responsiveness may predispose an athlete to lower extremity injury or further concussion.^48,55  Thus, the development of an integrated neurocognitive-mechanical assessment that incorporates key aspects of cognitive function, on visual-spatial navigation, RT, rapid online motor coordination, and decision making would progress the fields of both SRC and ACLR recovery and return-to-sport progressions. Classic metrics isolate each function independently, but in the real-world, these neurological functions occur together, thus, the development of an assessment that brings them together is highly warranted for both patient cohorts. By investigating the RT of both cognitive-motor response (ie, neurocognitive) and whole-body response (ie, neuromechanical), more detailed feedback and outcomes may provide important information on risk factors for performance as well as quantifying injury risk. Additionally, the integration of health-related quality-of-life surveys may also aid in clinical management and return-to-sport progressions in those individuals with individual and combined history of SRC and ACLR given prior findings that athletes with these injury histories report worse physical functioning than healthy controls 2+ years postinjury.⁵⁶ 

This study was not without limitations. First, this study was conducted in a sample of college-aged individuals, predominantly female, so future researchers should replicate the methodology in a sample including more males as well as at high school and collegiate athletes. Participants who sustained a previous ACL tear and ACLR were included in the study based on that specific ligamentous injury. It was not considered whether individuals had an isolated tear or multistructural damage (eg, meniscal tear, medial collateral ligament sprain). Future investigators should also employ strict time points for injury recovery, such as within 3 months of full RTP, 1 year post-RTP, and 2 years post-RTP. Investigating the immediate effects of an ACL tear will limit the ability of an individual to complete neuromechanical RT; however, future investigators should further validate the findings of neurocognitive RT between SRC and ACLR groups in the acute and subacute, especially in the acute phase of 1 injury (SRC or ACLR) after a history of the other. Lastly, most athletes in each group of this study had a history of multiple injuries in their categorized group. It is possible that multiple ACL injuries may present with osteopathic changes, and thus, future researchers should examine the effects of a dose response (eg, 1, 2, and 3+) of respective injuries.

In conclusion, no group differences were found between college-aged individuals with a history of SRC, ACLR, combination of SRC and ACLR, and healthy controls on RT or processing speed from a neurocognitive and a neuromechanical assessment battery. When further classified as having any SRC or any ACLR, group differences were observed on the TMT when compared with the noninjury history groups, indicating that apparent neurological deficits may exist related to spatial scanning and executive and neuromotor function 2 years post-SRC or post-ACLR. Future research is needed to examine neurocognitive and neuromechanical RT in athletes at various injury intervals and time points of recovery.

This project was funded by the Southeast Athletic Trainers’ Association (SEATA) Research Grant No. GR27596.