Concussion Baseline Testing: Preexisting Factors, Symptoms, and Neurocognitive Performance

A cross-sectional design was used to compare baseline neurocognitive test results and total symptom score with preexisting factors. The independent variables were sex; treatment for headaches, migraines, or a psychiatric condition (eg, depression, anxiety); diagnosis of ADHD; or strenuous exercise within the previous 3 hours. The dependent variables were the 4 composite ImPACT baseline scores (verbal memory, visual memory, visual motor speed, and reaction time) and total symptom score.

The participants in this study were collegiate student-athletes attending a National Collegiate Athletic Association Division 1 university between the 2009−2010 and 2013−2014 academic years (n = 503). However, 17 participants lacked a valid baseline ImPACT, so the final sample was 486 participants. Invalidity was based on an index developed by ImPACT Applications, Inc.^19,20  Of the 486 participants, 176 were female (age = 18.9 ± 1.3 years, height = 170.4 ± 8.0 cm, weight = 65.6 ± 10.6 kg) and 310 were male (age = 19.0 ± 1.3 years, height = 184.1 ± 10.9 cm, weight = 92.6 ± 17.9 kg). The student-athletes were drawn from football (n = 181), women's soccer (n = 59), baseball (n = 55), men's soccer (n = 53), softball (n = 25), women's lacrosse (n = 25), men's basketball (n = 21), women's basketball (n = 21), women's track and field (n = 19), women's volleyball (n = 18), and women's tennis (n = 9). Before testing began, the study received institutional review board approval and all participants completed informed consent forms.

We used ImPACT to assess neurocognitive function and total symptom scores and collect self-reported demographic data (version 2.0 was used until 2011 and then version 2.1 was used; the important difference between the 2 programs was greater internal security to better comply with Health Insurance Portability and Accountability Act [HIPAA] regulations). The demographic data provided information concerning previous treatment for headaches, migraines, or psychiatric conditions; diagnosis of ADHD; and recent strenuous exercise. Both versions of ImPACT offer 6 test modules to produce 4 composite scores: verbal memory, visual memory, visual motor speed, and reaction time. Higher composite scores on visual and verbal memory and visual motor speed and a lower score on reaction time indicate better performance. The 6 modules measure attentional processes, verbal recognition memory, visual working memory, learning, and other measures of cognitive functioning. Total symptom score was based on responses to the 22-symptom list. Although some have suggested that ImPACT has poor reliability,²¹  most studies have demonstrated good test-retest reliability.¹³⁻¹⁵

As part of the university's concussion-management protocol, student-athletes were required to participate in baseline testing. Typically, baseline testing occurred before the first year of competition and was completed on a computer in a closed laboratory. No more than 2 student-athletes and 2 test administrators were present because testing in a group setting has been shown to negatively influence performance.²²  Testing was administered at various times during the day, depending on the availability of the participants. Generally, ImPACT was completed within 30 minutes.

A series of multivariate analyses of variance (MANOVAs) were conducted using preexisting factors of sex; self-reported previous treatment for headaches, migraines, or a psychiatric condition; diagnosis of ADHD; and recent strenuous exercise as independent variables. Dependent variables from ImPACT were verbal memory, visual memory, visual motor speed, and reaction time composite scores and total symptom score. If the MANOVA was significant, then univariate analyses were conducted to determine which factor(s) accounted for statistical significance. Because 6 separate MANOVAs were performed in these analyses, we used a Bonferroni correction to determine statistical significance (P < .0083).

Each participant self-reported sex (n = 176/486 females [36.2%], 310/486 males [63.8%]); history of treatment for headaches (n = 29/474, 6.1%), migraines (n = 29/473, 6.5%), or a psychiatric condition (n = 11/472, 2.3%); diagnosis of ADHD (n = 29/356, 8.1%); and strenuous exercise within the previous 3 hours (n = 104/359, 29.0%) as part of the demographic section completed at the beginning of ImPACT. The distribution of preexisting factors and the ImPACT scores are shown in the Table.

The ImPACT composite scores were different by sex (F_5,480 = 6.98, Wilks λ = .93, P < .001, partial η² = .07). Univariate analyses demonstrated effects of sex on verbal memory (F_1,484 = 11.2, P = .001, partial η² = .02), visual motor speed (F_1,484 = 12.37, P < .001, partial η² = .03), and reaction time (F_1,484 = 7.64, P = .006, partial η² = .02), with females performing better than males on these tasks. Sex did not affect the total symptom score (F_1,486 = 2.76, P = .098, partial η² = .01).

Treatment for headaches was associated with a difference in ImPACT scores (F_5,468 = 7.3, Wilks λ = .93, P < .001, partial η² = .07). Total symptom score showed a difference (F_1,472 = 30.23, P < .001, partial η² = .06): those who had undergone previous treatment for headaches reported more symptoms than those who reported no such treatment. Although the differences were not significant, those with previous treatment for headaches performed worse on visual memory (F_1,472 = 3.94, P = .048, partial η² = .01) and verbal memory (F_1,472 = 3.01, P = .083, partial η² = .01).

Treatment for migraines also had an influence on ImPACT composite scores (F_5,467 = 4.28, Wilks λ = .96, P = .001, partial η² = .04). Univariate analyses demonstrated that those with previous treatment for migraines reported more symptoms than those without such treatment (F_1,471 = 11.78, P = .001, partial η² = .02). Visual motor speed performance was worse in those who reported previous treatment for migraines, but the value did not achieve significance (F_1,471 = 5.94, P = .015, partial η² = .01).

Treatment for psychiatric conditions was associated with a difference in ImPACT scores (F_5,466 = 3.81, Wilks' λ = 0.96, P = .002, partial η² = .04). Those with previous treatment for a psychiatric condition reported more symptoms than those without such treatment (F_1,470 = 19.16, P < .001, partial η² = .04).

A diagnosis of ADHD affected ImPACT composite scores (F_5,350 = 7.32, Wilks' λ = .91, P < .001, partial η² = .10). Univariate analyses indicated effects for visual motor speed (F_1,354 = 7.21, P = .008, partial η² = .02) and total symptom score (F_1,354 = 29.75, P < .001, partial η² = .08). Those with a previous diagnosis of ADHD performed worse on visual motor speed and reported more symptoms than those who reported no such diagnosis. Visual memory (F_1,354 = 5.76, P = .017, partial η² = .02) and verbal memory (F_1,354 = 3.20, P = .075, partial η² = .01) were worse in those with a previous diagnosis of ADHD, but the values did not achieve statistical significance. No differences were noted for strenuous exercise in the previous 3 hours (F_5,353 = 1.47, Wilks' λ = 0.98, P = .200, partial η² = .02).

The aim of our study was to determine whether preexisting factors influenced baseline neurocognitive performance and total symptom scores on ImPACT. As we hypothesized, many of the preexisting factors studied influenced neurocognitive performance and total symptom score. The only 2 factors that affected ImPACT baseline neurocognitive performance were sex and a previous diagnosis of ADHD. Treatment for headaches, migraines, or a psychiatric condition or a previous diagnosis of ADHD affected total symptom score.

Our finding that females performed better than males on certain aspects of the neurocognitive test was consistent with previous data.^8,16,17,23  Covassin et al¹⁶  observed that at baseline, females performed better on verbal memory, but males performed better on visual memory. In our study, males did not perform better on neurocognitive testing. The differences in visual motor speed and reaction time we noted are consistent with those found by Covassin et al,¹⁷  who hypothesized that estrogen may be one explanation for better performance in females.¹⁶  Another possible explanation is that female student-athletes exert more effort or take baseline testing more seriously than their male counterparts.²⁴  Additionally, our finding that a previous diagnosis of ADHD influenced visual motor speed is supported by earlier researchers who found that visual motor speed^11,12  was influenced by ADHD. Visual memory,^6,11,12 reaction time,^11,12  and verbal memory^6,12  were also affected in other investigations, but we did not observe these differences. A possible explanation for these deficits in neurocognitive performance is that those with a diagnosis of ADHD have problems with impulsivity and inattention, which could influence their results. Further, we do not know if they were taking medications for their ADHD; a recent study by Littleton et al²⁵  demonstrated that those with ADHD who were taking medication performed as well on ImPACT as the control group.

No differences in neurocognitive performance were evident for those who self-reported being treated for headaches, migraines, or a psychiatric condition. This result contrasts with previous research demonstrating that neurocognitive performance was influenced by depression⁸⁻¹⁰ and headaches.^5,6  These variables may not have been factors in our study because they can be transient states and the participants may not have been experiencing them at the time of testing. Covassin et al⁸  showed that current depressive symptoms influenced neurocognitive performance. However, Register-Mihalik et al⁵  did not find headache symptoms at baseline to influence neurocognitive performance.

We failed to find a difference in total symptom score at baseline in females, which is inconsistent with the data of previous investigators,^8,16  who noted that females had higher scores. Our observation that total symptom score was influenced by headaches,^5,6  a psychiatric condition,⁸  and ADHD^6,11,12  is consistent with the literature. This makes sense: those who were diagnosed with and had received treatment for these conditions reported more of the same symptoms that overlap with postconcussion symptoms.

The only factor that did not affect neurocognitive performance or total symptom scores was strenuous exercise 3 hours before baseline testing. Few investigators have addressed the effects of exercise on neurocognitive test scores specific to baseline concussion testing despite the strong evidence²⁶  demonstrating the influence of exercise on cognitive function. Although Mrazik et al¹⁸  found that maximal aerobic exercise influenced symptom score, their tests were administered 10 minutes postexercise, whereas most of the athletes participating in our study took ImPACT at least 1 hour or more postexercise, as planned by the sports medicine staff scheduling the baseline tests for these student-athletes. Mrazik et al¹⁸  did not examine any outcomes of neurocognitive tests other than symptoms. Although this finding is not in accordance with our hypothesis, it does have clinical implications. Our results suggest that previous exercise need not be taken into consideration when scheduling and administering ImPACT. Strenuous exercise within 3 hours of testing might not influence neurocognitive performance outcomes, so the test could be taken after an intense practice or conditioning session.

Ultimately, all of the preexisting factors analyzed, except for sex and exercise, showed an influence on total symptom score. These results stress the importance of baseline neurocognitive testing as part of concussion management. Symptom reporting is also important in concussion management and when determining RTP. Often athletes are not allowed to begin their RTP protocol until they are asymptomatic; however, if an athlete reports symptoms on a daily basis when healthy, the RTP decision might be incorrectly delayed.

Several limitations were associated with this study. As mentioned earlier, the preexisting factors of previous treatment for headaches, migraines, or a psychiatric condition or a diagnosis of ADHD were reported by only a small number of participants. Additionally, the percentages reported in this sample for previous treatment of headaches, migraines, and psychiatric conditions were smaller than those in the general population.^27,28  The incidences of headaches and migraines (6% in our study compared with 16%)²⁷  and mental health disorders (2% in our study compared with 18%)²⁸  were less than the national estimates, but the incidence of ADHD was slightly higher than national estimates (8% in our study compared with 4%).²⁹ 

Along the same lines, all of the preexisting conditions were self-reported. Although self-reporting can be reliable, it is not the best method for obtaining a medical history, but we did not have access to medical records. Therefore, it was not possible to determine who was currently receiving treatment for many of these conditions. Another limitation was the distribution of participants across sports: we tested a large number of football student-athletes, yet many other sports were underrepresented. Finally, we examined each variable independently and did not analyze the combined effects of the preexisting factors. Some of the effects seen could be due to a combination of variables instead of 1 variable alone.

Our results suggest that preexisting factors, including sex and ADHD, can influence neurocognitive performance on ImPACT in collegiate student-athletes. Also, most preexisting factors increased total symptom scores. These findings stress the importance of baseline testing as part of concussion management. Even though normative data are available for most neurocognitive tests, including ImPACT, our observations indicate that preexisting factors can negatively influence baseline scores. Normative data may not be sufficiently reliable for use in determining RTP. Neurocognitive data for athletes with certain preexisting factors may differ from normative data and demonstrate a higher total symptom score, leading to incorrect RTP decisions.

This research was supported by a grant from the American Medical Society for Sports Medicine (K.P.B., E.E.H., coinvestigators).