Concussion History and Heart Rate Variability During Bouts of Acute Stress

As part of structured preseason medical screening and performance testing organized by Hockey Québec, Midget-AAA (ages = 15–18 years), participating adolescents from the surrounding area took part in our assessment battery. All participants and parents were instructed on the purpose and procedures before the study. To ensure that each participant met the inclusion criteria, we conducted a structured interview to gather relevant demographic data and medical history. All athletes reported that they were not taking any prescription or over-the-counter medications. Athletes with a diagnosed history of neurologic disease, psychiatric illness, non–sport-related brain injury, learning disability, attention deficit/hyperactivity disorder, or any conditions with a known influence on cardio-autonomic function were excluded. A waiver of assent and consent to analyze de-identified data was obtained from the local institutional review board.

After the screening process, individuals were assigned to either the control group (n = 18) or history-of-concussion group (n = 16). A past diagnosis of concussion was confirmed via medical record review and by a neuropsychologist (not an author). To prevent the possibility of an athlete with an undiagnosed concussion being included, we asked participants, “Following a blow to the head, neck, or body, have you ever experienced any of the following symptoms: headache, dizziness, confusion, blurred vision, balance problems, sensitivity to light and/or noise, fatigue, drowsiness, difficulty falling asleep, emotional, irritable, sad, or anxious?” Control participants who responded yes to any symptoms were excluded from the analyses. All participants in the history-of-concussion group were asymptomatic at the time of testing and currently engaged in regular team activities.

To minimize any effects of diurnal variation on cardio-autonomic function, all participants completed testing at approximately the same time of day (late morning). They were instructed to adhere to their normal eating habits and abstain from consuming any caffeinated beverages on the day of testing. All testing took place in a quiet room with regulated ambient temperature and stable humidity. The experimental procedure is illustrated in Figure 1A.

After arriving at the laboratory, participants were outfitted with a Zephyr (Medtronic) BioHarness strap, complete with BioModule sensor, which was used to collect continuous electrocardiographic (ECG) and respiratory data throughout testing. All ECG data were collected at a sampling frequency of 250 Hz. Participants underwent a 5-minute, eyes-open resting ECG assessment (HRV_REST). They then performed a series of baseline cognitive tasks, which are more thoroughly described herein and consisted of both homogeneous and heterogeneous conditions of the switch task (HRV_COG1). After the switch task, participants completed a bout of continuous, steady-state aerobic exercise on a stationary cycle ergometer. Finally, after a 10-minute rest period, they underwent a second round of cognitive testing, using an identical version of the switch task completed earlier (HRV_COG2). Total testing time was approximately 90 minutes.

The color-shape switch task is commonly used to investigate both working memory and cognitive behavior.¹⁵  Briefly, under 2 task conditions, participants were asked to respond according to specific rule sets mapped to either the color (COLOR-response) or shape (SHAPE-response; Figure 1B). They were then asked to complete the MIXED rule-response condition and apply either the color or shape rule, depending on the outline (solid and dashed) or the shape (Figure 1B).

Before each single-rule–set condition, participants completed 15 practice trials to familiarize themselves with the task. They then performed 1 block (30 trials) of each condition (COLOR-response and SHAPE-response). After the single-rule–set blocks, they completed 3 blocks of the MIXED rule-response condition, consisting of 60 trials each. Before the first block, participants performed a set of 30 practice trials. In the MIXED rule-response condition, the 2 rule sets were altered in an equiprobable and random manner.

The switch task was presented using the Psykinematix (version 1.5.1; KyberVision Japan LLC) software, and responses were recorded using a serial-port response pad (model RB-540; Cedrus Corp). Task stimuli were displayed centrally on a black background and consisted of circles or squares (7 cm × 7 cm, 4° visual angle), in either blue or green, with a white outline. All stimuli were presented until participants responded, with a maximum response duration of 2000 milliseconds and a 50-millisecond intertrial interval. Before each task, they received standardized oral instructions, and written instructions were provided on the monitor in front of them. Participants were asked to respond as quickly and accurately as possible.

The aerobic exercise protocol was carried out using a cycle ergometer (Lode). To begin, all participants completed a 5-minute incremental warmup. The warmup period was followed by a 20-minute steady-state exercise period at approximately 60% to 70% of the athlete's age-predicted (220 − age) maximum heart rate. Participants were instructed to maintain a cadence of approximately 60 rpm while the load (in watts) was adjusted to maintain the calculated target heart rate. After the 20-minute exercise period, participants were guided into a 2-minute active cool-down (unloaded cycling) and 10-minute seated, resting recovery.

Raw ECG waveforms were processed using Kubios HRV (version 2.0; Biosignal Analysis and Imaging Group). The RR intervals were quantified as the time interval between successive R peaks in the QRS complex gathered from the raw ECG recording. We then normalized the data by manually inspecting the RR intervals to identify ectopic beats and any artifacts caused by movements, creating normalized RR intervals (NN intervals). Next, a 10% Hanning window was applied to the remaining ECG data. Electrocardiograms collected during the HRV_REST were segmented into 2-minute epochs to match the duration of each switch-task block. Previous researchers¹⁶  have validated time-domain measurements collected during ultrashort (<5 minutes) HRV recordings.

We calculated linear, time-domain indices of HRV for each 2-minute epoch of the HRV_REST and for each 2-minute block of the MIXED rule-response condition of the switch task for HRV_COG1 and HRV_COG2. These time-domain parameters included the mean NN interval (NNmean), SD of NN intervals (SDNN), and root mean square of successive NN interval differences (RMSSD). Measures of HRV can be modulated by differences in respiration rate. To correct for potential differences in breathing during each assessment, we computed individual respiration rates.

Independent-samples t tests were conducted to identify any differences in characteristics between the groups. To compare resting HRV and the effect of physiological stressors on cardio-autonomic function group (control, history of concussion) × time (HRV_COG1, HRV_COG2), we performed repeated-measures analyses of covariance for each outcome measure with HRV_REST performance entered as the covariate. The Greenhouse-Geisser correction was applied to any measure for which the assumption of sphericity was violated. If we observed a main effect or interaction, we applied the Bonferroni correction post hoc. All statistical analyses were completed using SPSS statistical software (version 25; IBM Corp). The α level was set a priori at P ≤ .05.

The groups did not differ in age, body mass index, or number of years playing sport (P > .05), suggesting successful matching between groups (Table 1).

Cardio-autonomic profiles of each group and condition are presented in Table 2. Univariate analysis did not show any interaction or main effect for respiration rates (P > .05). Therefore, the respiration rate was not included in further HRV analyses. We did not observe group differences at HRV_REST (P > .05). Our repeated-measures analyses of covariance revealed a main effect of group for SDNN (F_1,31 = 4.31, P = .046, ηp² = 0.122) and RMSSD (F_1,31 = 4.88, P = .04, ηp² = 0.136). Pairwise comparisons showed that, during both HRV_COG1 and HRV_COG2, individuals with a history of concussion had greater SDNN and RMSSD than their matched control participants (Figure 2). No other interactions or main effects were found between groups.

We aimed to examine HRV indices of cardio-autonomic function in adolescent athletes with a history of concussion under periods of acute stress. The main findings of our study were as follows: at rest, athletes with a history of concussion did not demonstrate any differences in cardio-autonomic function relative to matched control individuals. During both HRV_COG1 and HRV_COG2, athletes with a history of concussion exhibited elevated HRV (SDNN and RMSSD) compared with matched control participants. We are the first to demonstrate similar trends after a bout of mental exertion in the chronic stage of concussion recovery.

Previous researchers^9,10  suggested that bouts of both physical and mental exertion in healthy populations were characterized by reductions in HRV (reduced vagal activity) as the body adjusted to meet the increased demands. During exercise, as the intensity of exercise increases, the increased metabolic demand of the body results in a subsequent increase in heart rate and SNS input to the heart.¹⁷  On the other hand, increased mental load is associated with increased activation of the prefrontal brain regions.¹⁸  Our matched control individuals without a history of concussion displayed this pattern of reduced HRV during both HRV_COG1 and HRV_COG2. However, athletes with a history of concussion showed uncharacteristic elevations in HRV.

Whereas athletes with a history of concussion demonstrated cardio-autonomic dysregulation during acute bouts of increased stress, we did not see any differences in HRV at rest. This finding adds to the findings of numerous researchers who suggested that stressors may exacerbate functional deficits not observed under resting conditions. Previous investigators showed that asymptomatic individuals with a history of concussion experienced abnormal HRV in response to exercise¹⁹  and prolonged recovery of HRV metrics after exercise.²⁰  These results in addition to ours suggest that abnormal cardio-autonomic responses to stress may indicate persistent disruptions in the ANS, limiting its ability to appropriately regulate physiological systems after concussion. However, some authors noted that HRV and other measures of ANS function recovered within the first few weeks after injury,²¹  even during periods of acute stress.¹⁴  It is possible that these differences are mediated by factors such as recovery status, postinjury management, concussion history, time since injury, and emotional status.^20,22  We attempted to account for these factors, but more work is needed to further identify how these confounding variables affect HRV after concussion.

Although the exact mechanisms underlying cardio-autonomic dysfunction after concussion are unknown, current theories suggest that HRV provides a means of quantifying the efficiency of neural communication among the central nervous system and other physiological systems (ie, cardiovascular). These theories propose that modulations of vagal tone (PNS activity) serve as a biological response to prepare the body to meet current or anticipated demands and to efficiently carry out goal-directed behaviors.²³  Frontolimbic structures in the central autonomic network (ie, prefrontal cortex, hypothalamus, amygdala, and hippocampus) are heavily involved in the organization and execution of goal-directed behaviors.²⁴  Additionally, these brain regions exert top-down control over the ANS and the intrinsic nervous system of the heart.²⁵  Furthermore, these brain regions and the interconnections among them are some of the areas and processes most frequently affected by concussions.²⁶  Accordingly, the different HRV responses observed in athletes with a history of concussion during acute bouts of either mental or physical stressors could indicate improper or inefficient communication among physiological systems in response to the current situational demands.

Regardless of the underlying mechanism, the fact remains that concussive injuries are characterized by widespread impairment that may affect several neurologic systems, including the cardio-autonomic pathways. Also, these deficits may persist for months or years after the initial injury. If unattended, these deficits may impede typical neurologic development and increase the individual's susceptibility to long-term cardiovascular and mental health concerns.²⁷ 

Our study had limitations. First, given the sample of only male adolescent athletes and the known sex differences in HRV, our findings may not be generalizable to a female population²⁸  or older, more physiologically mature athletes. Second, the Midget-AAA hockey system in Canada involves elite-level athletes for that age group, and HRV has been associated with age²⁹  and fitness.³⁰  Whereas no age differences were present in our sample, we did not assess individual fitness levels. Testing took place before the start of the season, and the athletes were of similar age, body mass index, and level of competition. Therefore, it is reasonable to assume that all participants had comparable fitness levels. Third, HRV is known to be sensitive to emotional status.¹¹  We excluded anyone presenting with a psychological or psychiatric condition, but we did not quantify acute emotional status and sleep patterns. Future authors should consider these potentially confounding factors, as they have been shown to affect HRV and other measures of cardio-autonomic function.

Our study is one of the first to investigate the long-term effect of concussive injuries on the cardio-autonomic responses to the bouts of physiological stress common in the everyday lives of student-athletes. Our findings indicated that concussion-related deficits in cardio-autonomic function may persist beyond the acute phase of injury. We also demonstrated that HRV may be used as a nonspecific indicator of neural health and recovery. Heart rate variability measurements are relatively fast and simple and dynamic enough to be collected under various testing conditions (exercise, cognitive tasks). Furthermore, our results highlighted the importance of assessing athletes under periods of physiological stress commonly experienced in everyday life, as deficits may not emerge under resting conditions. Measurements of HRV may help in the diagnosis of concussive injuries and the management of symptoms and provide objective biomarkers to inform return-to-play and return-to-learn policies.