Neurophysiological Effects of Repeated Soccer Heading in Youth

Abstract

Repeated head loading in sports is associated with negative long-term brain health, and there is growing evidence of short-term neurophysiological changes after repeated soccer heading. The objective of this study was to quantify the head kinematics and effects of repetitive soccer headers in adolescents using an instrumented mouthguard. Adolescent soccer players aged 13-18 years were randomly assigned to a kicking control, frontal heading, or oblique heading group. Participants completed neurophysiological assessments at three-time points: immediately prior to, immediately after, and approximately 24 h after completing 10 headers or kicks. The suite of assessments included the Post-Concussion Symptom Inventory, visio-vestibular exam, King-Devick test, modified Clinical Test of Sensory Interaction and Balance with force plate sway measurement, pupillary light reflex, and visual evoked potential. Data were collected for 19 participants (17 male). Frontal headers resulted in significantly higher peak resultant linear acceleration (17.4 ± 0.5 g) compared to oblique headers (12.1 ± 0.4 g, p < 0.001), and oblique headers resulted in significantly higher peak resultant angular acceleration (frontal: 1147 ± 45 rad/s2, oblique: 1410 ± 65 rad/s2, p < 0.001). There were no neurophysiological deficits for either heading group or significant differences from controls at either post-heading timepoint, and therefore, a bout of repeated headers did not result in changes in the neurophysiological measures evaluated in this study. The current study provided data regarding the direction of headers with the goal to reduce the risk of repetitive head loading for adolescent athletes.

Fig. 1

(Left) Peak resultant kinematics for each header (58 frontal, 44 oblique) are represented in beehive plots. *Difference between frontal and oblique heading groups (p < 0.004). (Right) Average resultant frontal and oblique kinematic time series plotted with 95th percentile confidence intervals (CI) of the mean.

Fig. 2

Average frontal and oblique kinematic time series by axis plotted with 95th percentile confidence intervals (CI) of the mean. *Peak kinematic difference between frontal and oblique heading groups (p < 0.017). AP = Anterior/Posterior; ML = Medial/Lateral; SI = Superior/Inferior.

Fig. 3

Post-concussion symptom inventory (PCSI) total symptom score for control (n = 8), frontal (n = 6), and oblique (n = 5) participants presented as mean ± standard error of the mean (SEM). Control and heading participants did not differ in pre-intervention PCSI total symptom scores (p > 0.080). There were no changes in PCSI relative to pre-intervention for control, frontal, or oblique participants at the 0 h post (p = 0.277, p = 0.185, p = 0.666, respectively) or 1 d post (p = 0.170, p = 0.233, p = 0.184, respectively) timepoints, and there were no differences between control, frontal, and oblique participants in total symptom score change from pre-intervention at the 0 h post (p > 0.824) or 1 d post (p > 0.090) timepoints.

Fig. 4

Near point of convergence (NPC) break distance for control (n = 8), frontal (n = 6), and oblique (n = 5) participants displayed as mean ± SEM. Control, frontal, and oblique participants did not differ in pre-intervention break distance (p > 0.310). There were no changes in NPC break relative to pre-intervention for control, frontal, or oblique participants at the 0 h post (p = 0.844, p = 0.151, p = 0.261, respectively) or 1 d post (p = 0.422, p = 0.121, p = 0.501 respectively) timepoints, and there were no differences between control, frontal, and oblique participants in change from pre-intervention of NPC break at the 0 h post (p > 0.244) or 1 d post (p > 0.157) timepoints. For right accommodation, oblique and control participants differed in change from pre-intervention at 0 h (*p = 0.018) and 1 d post (*p = 0.012). There were no other differences for left and right accommodation between groups at any timepoint (p > 0.033) or change from pre-intervention (p < 0.074).

Fig. 5

Tandem gait total errors + sway scores for control (n = 8), frontal (n = 6), and oblique (n = 5) participants displayed as mean ± SEM. Control and heading participants did not differ pre-intervention (p > 0.497). Compared to pre-intervention, control participants had lower scores at the 0 h post timepoint (*p = 0.020) but not at the 1 d post timepoint (p = 0.316). Frontal and oblique heading participants did not change from pre-intervention at the 0 h (p = 0.203, p = 0.815) or 1 d post (p = 0.363, p = 0.305) timepoints, and there were no differences between control, frontal, and oblique participants in change from pre-intervention at the 0 h post (p > 0.154) or 1 d post (p > 0.703) timepoints.

Fig. 6

King-Devick total duration to complete all three cards for control (n = 8), frontal (n = 6), and oblique (n = 5) participants displayed as mean ± SEM. Control, frontal, and oblique participants did not differ pre-intervention (p > 0.686). Compared to pre-intervention, oblique heading participants completed all 3 cards faster 1 d after heading (*p = 0.003). No other change from pre-intervention was significant (p > 0.033), and there were no differences between control, frontal, and oblique heading groups forchange from pre-intervention at 0 h (p > 0.466) or 1 d (p > 0.580).

Fig. 7

Composite sway index for the MCTSIB for control (n = 8), frontal (n = 6), and oblique (n = 5) participants displayed as mean ± SEM. There were no changes from pre-intervention (p > 0.103). There were no differences in sway index between control, frontal, and oblique participants at pre-intervention (p > 0.181) or change from pre-intervention at the 0 h post (p > 0.138) or 1 d post (p > 0.654) timepoints.

Fig. 8

Pattern reversal visual evoked potential P100–N75 amplitude for control (n = 8), frontal (n = 5), and oblique (n = 5) participants displayed as mean ± SEM. Oblique and frontal heading participants differed in change from pre-intervention at the 1 d post timepoint for P100-N75 amplitude (p = 0.002). There were no other differences between control, frontal, and oblique participants at any timepoint (p > 0.027).