A Multifaceted Approach to Interpreting Reaction Time Deficits After Adolescent Concussion

METHODS

Study Design and Participants

We conducted a prospective investigation of adolescent athletes who sustained a concussion and uninjured control individuals. The participants with concussion were recruited from a pediatric sports medicine center associated with a regional tertiary care hospital, where they were receiving outpatient treatment after their concussion. Recruits were included if they were between 12 and 18 years of age. Recruits with concussion were included if they were diagnosed with a concussion by a Board-certified sports medicine physician,²⁰ seen within 14 days of injury, and reported a Post-Concussion Symptom Inventory (PCSI) score ≥9 during the initial assessment, ensuring that they were symptomatic at that time.⁵ We selected a PCSI score ≥9 as a minimal standard to be certain that participants had not experienced spontaneous recovery within the first 2 weeks after injury and before enrollment.²¹ Control participants were recruited from the surrounding community (ie, local high school athletes) to match the general characteristics of the concussion group (Table 1). We excluded recruits with coexisting lower extremity injuries affecting balance, a concussion within the past year (other than the current concussion for the concussion group), self-reported preexisting learning disability, documented neuroimaging of a structural brain injury, or a concussion sustained during a high-velocity impact (eg, motor vehicle collision).⁵ Volunteers were also excluded if they did not intend to return to sport after medical clearance.⁵ The local institutional review board approved the study protocol before data collection. All participants and their legal guardians (for participants <18 years of age) provided written informed consent or assent to engage in the study.

Table 1. Participant Demographics, Medical History, and Injury Characteristics

View Fullscreen

Participants with a concussion were assessed by a trained member of the study team within 14 days of their concussion and retested approximately 28 days later. This time frame was used for the control group as well (initial test with a follow-up test approximately 28 days later) and was selected a priori to reflect the typical RTP clearance window for adolescents after a concussion.^4,22 In this observational study, we did not make any recovery diagnosis at the 28-day follow-up time point but performed only the evaluations needed for the procedures outlined in the following section.

Outcome Variables

Clinical Drop Stick RT

As in previous studies, a 1.3-m measuring stick was coated with high-friction tape, marked every 1 cm, and attached to a hockey puck.^13,23,24 Participants were seated with their dominant forearm resting horizontally on a height-adjusted table so that the elbow was comfortable in 90° of flexion. The test administrator placed the drop stick device vertically between the participant’s thumb and index finger. Individuals were instructed to catch the stick as quickly as possible after it was dropped at random intervals of time. Each person was allotted 2 practice trials followed by 8 test trials. The distance the puck traveled was measured in centimeters. The average distance of the 8 trials was recorded and used for analysis. Distance was converted to RT (milliseconds) using the equation^5,7:

The clinical drop stick RT test has been established and validated in several studies of RT in athletes (Figure 1).^7,13,23,25

View Figure

• Download as Powerpoint
• Download Figure

Simple RT (Smartphone)

Simple RT data was gathered using an RT test on a smartphone (model Galaxy S8; Samsung Electronics Co, Ltd) with the Reaction Time Tester Pro application, described earlier in a feasibility trial examining clinical RT assessment tools.¹⁶ We placed the phone flat on a table at a 2-cm standardized distance in front of the participant’s self-selected dominant hand.¹⁶ A red dot appeared in the middle of the phone screen, and at random intervals, the dot turned green. The participant was instructed to tap the dot as quickly as possible once the color changed to green and return to the resting position (ie, finger on the table) between trials. Response time was recorded in the application as the time from the dot color changing to green to the participant tapping the screen. Average response time (milliseconds) was calculated from 30 successful trials.¹⁶

Stroop RT (Single Task and Dual Task)

During the single-task Stroop, participants were instructed to stand in place with feet together and eyes open for 30 seconds while responding to a series of auditory Stroop tasks.¹⁴ A computer-recorded voice spoke the word “high” or “low” in either a high or low pitch, with word and pitch combinations selected at random. Through headphones attached to the smartphone, individuals were instructed to audibly identify the pitch of the word, not the word itself, throughout each 30-second trial.¹⁵ A smartphone application (IMPROVE; Control One LLC) was used to elicit the Stroop task audio recordings and calculate the response time to detect the response accuracy as high, low, or unknown.^15,16 The smartphone recorded the time in milliseconds between each auditory cue and oral response. Reaction time was calculated by taking the difference between stimulus onset and oral response onset. Subsequently, the IMPROVE application classified each response as correct, incorrect, or unanswered. Four trials were completed per condition, and averages were obtained across the 4 trials.

Dual-task RT was assessed using a similar protocol as for the single-task Stroop, with the addition of a simultaneous walking task. Participants were instructed to walk at a self-selected pace along a 10-m pathway toward an object placed on the ground while simultaneously responding to the Stroop stimuli. On reaching the object, the person was asked to walk around it and return to the initial position.¹⁵ During the trial, the smartphone application recorded gait velocity, cognitive Stroop response accuracy, and response time. Four trials were completed, and averages were obtained across the 4 trials.

Symptom Severity (PCSI)

During initial and follow-up assessments, participants also completed the PCSI to report their current symptom severity. The PCSI is a patient-reported inventory with 22 questions measuring the perceived severity of postconcussion symptoms. Common concussion symptoms are rated from 0 (none) to 6 (severe), and the summed responses provide an overall symptom score.^14,26 Total scores range from 0 to 132. The reliability and validity of the PCSI are well documented.²¹

RESULTS

A total of 48 participants (27 adolescents with a concussion and 21 uninjured control adolescents) enrolled in the study and underwent initial RT assessment. Of those, 41 (85% retention; n = 23 in the concussion group and n = 18 in the control group) returned for the follow-up test and completed the multifaceted RT assessment at both time points. The 2 groups had similar demographic characteristics, although a greater proportion of the concussion group reported a preconcussion history of headache disorders and migraine than the control group (Table 1). In addition, the concussion group reported a higher concussion symptom burden on the PCSI than the control group during the initial assessment (Table 1). Thus, preinjury headache disorder, migraine, and PCSI score were included as covariates in multiple linear regression modeling.

During the initial evaluation, the concussion group demonstrated slower clinical drop stick (Figure 2A), single-task Stroop (Figure 2C), and dual-task Stroop (Figure 2D) RT. The AUC was outstanding for the dual-task Stroop RT, excellent for the clinical drop stick RT, and acceptable for the single-task Stroop RT (Figure 2). We observed a medium effect size but nonsignificant difference between groups for the simple RT measurement (Figure 2B). Classification accuracy was highest for the dual-task Stroop and clinical drop stick RT measurements (Table 2). After adjusting for preinjury headache disorder, migraine, and initial symptom burden, the concussion group had slower clinical drop stick (β = 58.8; 95% CI = 29.2, 88.3; P < .001) and dual-task Stroop (β = 464.2; 95% CI = 318.4, 610.0; P < .001) RTs at the initial visit than the control group. Simple (β = −3.54; 95% CI = −58.3, 51.2; P = .90) and single-task Stroop (β = 32.8; 95% CI = −124.6, 190.2; P = .68) RT results were not associated with group.

View Figure

• Download as Powerpoint
• Download Figure

Table 2. Cut Points With the Highest Classification Accuracy for Distinguishing Between the Concussion and Control Groups Within 14 Days of Concussion Across the 4 Reaction Time Measurements

View Fullscreen

At 1-month follow-up testing, differences were present between groups with large effect sizes for clinical, single-task Stroop, and dual-task Stroop RT and small or nonsignificant effects between groups for simple RT and symptom burden (Table 3). For participants with a concussion, this follow-up visit was associated with RTP clearance: patients were deemed clinically recovered from concussive symptoms.

Table 3. Reaction Time Performance Characteristics Between the Concussion and Control Groups at the 1-Month Follow-Up Test^a

View Fullscreen

DISCUSSION

The clinical drop stick and dual-task Stroop RT tests provided high clinical utility and may be useful during the initial and RTP concussion assessments. These approaches have high classification accuracy, sensitivity, and specificity to detect postconcussion RT deficits within the first 2 weeks postconcussion and may be useful for clinicians assessing postconcussion RT impairments as part of a multifaceted evaluation. Additionally, at 1-month follow-up testing, differences were found between groups for clinical, single-task Stroop, and dual-task Stroop RT. The RT assessments we selected are readily accessible to health care providers in a variety of settings and do not require laboratory or computerized testing to identify deficits, although some technological elements may still be required for the simple (eg, a smartphone) and Stroop RT assessments. They also are quick to perform and may offer insight into athletes’ motor function as they return to highly complex motor-cognitive tasks during sports.^6,16 The existing literature, including the 2017 Concussion in Sport Group consensus statement and the 2020 Concussion Clinical Practice Guideline, recommends including RT measurement in concussion evaluations^2,20 as an objective measure to monitor recovery beyond subjective symptom reporting.² Our work extends observations that RT assessments can supply objective assessment of persisting deficits after symptom resolution and highlights the need for continued rehabilitation before RTP.

Both the clinical drop stick and dual-task Stroop RT findings were slower among injured athletes than uninjured control participants within the first 14 days of concussion. These tests involve upper extremity function (clinical drop stick RT), whole-body movement (dual-task Stroop RT), and simultaneous cognitive and motor tasks (dual-task Stroop RT), each of which is relevant to the demands of sports and may be overlooked in other RT assessments. Simple and single-task Stroop RT did not identify differences between the concussion and control groups at initial assessment, which, along with lower sensitivity and specificity, indicates that these assessments may not be ideal for evaluating RT in the first 2 weeks after concussion. Furthermore, after adjusting for potential confounding variables, we noted that the univariable difference between groups for the single-task Stroop RT was no longer significant, suggesting that a history of a headache disorder, a history of migraines, and the initial symptom burden are possible modifying factors in the relationship between RT and concussion.

At follow-up assessment, which corresponded with the time of clearance for the participants with concussion, both the clinical drop stick and dual-task Stroop RT tests detected differences between adolescents with concussion and uninjured control adolescents, despite no difference in symptom burden. Prior authors determined that athletes were at an increased risk of MSK injury after concussion compared with a healthy cohort, which may be due in part to persisting impairments in dual-task cognitive processing, neuromuscular control, or both.^8,14,16–19 Although commonly used concussion assessments and mental health evaluations have not predicted subsequent MSK injuries postconcussion,⁸ the RT deficits noted in our study may represent an additional metric to consider given the statistically and clinically significant between-groups differences both initially postinjury and after symptom resolution. Moreover, testing RT under dual-task conditions may better reflect sport-specific demands.⁶ Clinically, our results have potential implications for the postconcussion injury risk. Reaction time may represent an interventional target, given that rehabilitation professionals can perform neuromuscular-training programs with athletes returning to sport after a concussion,¹⁶ with demonstrated effectiveness in reducing subsequent injuries.²⁸ Incorporating RT training into these programs with a rehabilitation specialist may be another avenue for addressing remaining deficits, thereby reducing the likelihood of subsequent injury.

Our study had several limitations that are important to consider when interpreting findings and generalizing to additional populations. Participants were seen at a specialty pediatric sports medicine clinic, 12 to 18 years old, and evaluated initially within 14 days of concussion. Those who had recovered by the time of the first assessment (PCSI score ≤ 9) were excluded; thus, our sample had a relatively similar profile upon enrollment, but this did not account for the variation in recovery within the first 2 to 4 weeks after concussion. Therefore, our findings may not be generalizable to patients seen in other care settings, such as emergency departments or on field, or to patients seen sooner or later postinjury (ie, within hours or >2 weeks). Also, the performance improvements in the control group may reflect a practice effect for the clinical drop stick RT assessment, indicating that healthy participants improve with repeat administration of the test. These practice effects may also exist for patients with concussion, independent of improvements in neuromuscular function as part of concussion recovery. Additionally, the time frame of follow-up was selected based on previous studies generalizing that concussion RTP typically occurs around 1 month postinjury.¹³ We acknowledge that the concussion recovery timeline varies and that athletes may recover in as few as 7 days or have lingering symptoms past our time frame for follow-up. Further, whether patients participated in formalized rehabilitation programs or received a home program focusing on addressing RT deficits was unknown. Future authors should examine whether dual-task RT deficits that persist after clearance to resume sport participation are related to a higher incidence of subsequent MSK injury. Moreover, future researchers should identify normative values and normative-based cutoff scores for each of the RT tests and provide minimal clinically important difference values or reliable change indices to determine a clinically meaningful threshold as it relates to postinjury assessment.

In conclusion, the clinical drop stick and dual-task Stroop RT tests are assessments to consider when quantifying concussion deficits after adolescent concussion. Dual-task Stroop RT testing may be the optimal test to incorporate because clinical RT performance improvements may be seen due to practice effects. Such approaches are available in many different clinical settings and may reduce the reliance on time-consuming or cost-prohibitive approaches.