Our data suggest a possible but non-significant association between individual basic and complex (executive) cognitive functions and muscoloskelettal injuries in professional male football players. However, due to the non-significant results across the assessed cognitive functions, our hypothesis could not be confirmed.
Although not statistically significant, we found that players with higher TMT-ratio-scores—indicating poorer cognitive flexibility or set-shifting ability—had 64% higher odds of injury. Athletes in open-skill sports like football must quickly adapt to unpredictable changes on the field (e.g., positions of teammates, opponents, and the ball) while making rapid decisions (e.g., pass, tackle, dribble) under time constraints16. This constant adjustment may require cognitive flexibility31. Athletes with lower cognitive flexibility may struggle to react and adapt motor responses to unexpected events, increasing their injury risk. Initial evidence suggests that lower cognitive performance, such as reduced cognitive flexibility, may lead to unfavourable biomechanics during jump landing tasks, which involve time-sensitive decisions and may increase the risk of knee injuries21. In this context, the observed potential association between higher TMT ratio scores and increased injury risk in the present study appears plausible.
Interestingly, performance on the TMT-4 test version, as well as the magnitude of the absolute TMT difference score—both measures of executive functions such as working memory and cognitive flexibility24—were far from sigmificantly related to injury risk. This may suggest that the proportional index of the TMT-ratio-score, which expresses executive performance relative to basic processing abilities, provides a more nuanced and sensitive measure of cognitive efficiency related to injury risk in professional football players. This makes sense insofar as playing performance and potential biomechanical injury risk factors in motor-cognitively challenging game situations seem to depend on a combination of basic and complex cognitive processes6,7,9,16, rather than a more isolated, specific domain like working memory or cognitive flexibility, which are predominantly reflected by the TMT-4 and the absolute difference score24. It is important to emphasize that this finding is hypothesis-generating and requires further investigation in larger, adequately powered studies. However, since the predictive value of the TMT-ratio-score was not statistically significant, showed only moderate sensitivity and specificity, and was only marginally better than age as a predictor—combined with its lower test-retest reliability compared to the individual TMT subtests28—this trend should be interpreted with caution. Future prospective studies should further explore and confirm the injury-predictive potential of these cognitive functions using the TMT or other established tests (e.g., task switching paradigms like PsyToolkit) before drawing more definitive conclusions about the predictive value of cognitive performance measures. Additionally, future research should investigate other executive functions not explicitly assessed in the present study such as working memory and inhibitory control, which are also critical for rapid decision-making during athletic tasks21. These domains can improve with training32 and may influence injury risk, making them important targets for assessment and prevention.
The association between higher age or more years of playing and an increased risk of injury in football players is well-documented33,34 and was also observed in our dataset. This relationship may reflect accumulated musculoskeletal load and degeneration, slower recovery times, or age-related declines in neuromuscular function35. Our findings suggest that any predictive influence of cognitive performance should be interpreted within the context of age-related risk. Future studies should further investigate whether cognitive factors interact with or mediate the effects of age on musculoskeletal injury risk.
Regarding more basic cognitive processes such as visual search and perception, we found a potential link to injury risk. Specifically, players with shorter TMT-1 completion times, indicating faster visual search, tended to show approximately 40% decreased odds of injury. However, this effect was not statistically significant, and thus the potential association should be interpreted with caution. Larger, adequately powered confirmatory studies are warranted to explore this relationship further. Previous studies have suggested that injured athletes might outperform uninjured ones on visual perception tests such as the TMT-1. For example, Stone et al.36 proposed that athletes recovering from ACL injuries might adopt a postoperative strategy that compensates for proprioceptive deficits by increasing visual control. This cortical compensatory mechanism has been supported by previous systematic evidence37. However, since injury history (e.g., prior ACL injuries) was not explicitly assessed in our sample, we cannot determine whether such a mechanism was present among the injured players in this study. As a result, any interpretation regarding visual compensation influencing TMT performance remains speculative and should be treated with caution. The same applies to the observed, non-significant trend between faster TMT-5 performance—indicative of faster hand motor execution speed—and an approximately 36% reduction in injury odds. Future research should explore whether pure motor speed represents an additional injury risk factor in football players, beyond cognitive processing speed.
Contrary to our hypothesis, we did not find a relationship between injury risk and either visuomotor vigilance (i.e., simple and choice reaction speed) or visuospatial short-term memory (CORSI). Regarding visuomotor vigilance, our findings diverge from previous evidence. Systematic reviews have reported that 7 out of 914 and 5 out of 6 studies6 identified a link between slower upper-extremity visuomotor speed and increased injury risk, primarily in team sport athletes. Several methodological and participant-related differences may account for this discrepancy. First, the methods used to assess visuomotor reaction time varied. For example, one study12 included in the reviews used the Dynavision Assessment and Training System, which involves standing participants completing reaction time tasks with upper extremity movements combined with visual search. This setup offers higher ecological validity, as it more closely mimics the visuomotor demands of real sports. In contrast, our test involved seated participants responding to visual stimuli on a computer screen via key presses, which requires lower motor demands and may therefore underestimate real-world cognitive–motor challenges. Although some of the other reviewed studies also used PC-based reaction tests in a seated position, most employed the ImPACT neurocognitive test battery, which may differ slightly from the tasks used in our PsyToolkit-based assessments. Second, participant characteristics differed substantially. Prior research primarily involved high school and collegiate athletes in team sports6,14, whereas our sample comprised older, professional male football players. It is possible that those athletes exhibit less variability in basic visuomotor reaction times due to their highly advanced performance levels. However, executive functions—such as cognitive flexibility—may still show meaningful individual differences even at high performance levels, potentially explaining their greater relevance in our findings.
Regarding visuospatial memory, previous reviews reported mixed findings. Four out of 9 studies14 and 4 out of 6 studies6 examined visual memory (mostly using ImPACT), but only one study (Swanik et al., 2007) in both reviews found a significant relationship with injury risk. Notably, none of these studies used the CORSI test to assess visuospatial memory specifically, limiting direct comparisons. Future research is warranted to determine whether visuospatial memory, as measured by tasks like CORSI, lacks predictive value for injury risk.
Practical implications and future research
Based on our study’s results, we cannot recommend integrating cognitive tests into pre-season performance and injury risk screenings for injury prevention in professional men’s football. However, beyond basic cognive processes such as visul perception, our findings provide initial indications of the potential relevance of executive function, specifically cognitive flexibility and set-shifting abilities. Further studies are needed to better understand the prognostic value of executive cognitive function measures before making specific recommendations for injury risk screening. Specifically, prospective studies, including female players and other key predictors such as age and years of playing as well as previous or preseason injuries, are needed to better understand the potential additional prognostic value of basic cognitive and particularly executive function measures. With regard to the outcome variable, we also suggest a more detailed analysis of injury occurrence, including mechanisms (contact vs. non-contact), locations (e.g., lower limbs), severity (e.g., downtimes), number and types of injuries (e.g., ligament sprains or tears), which was not feasible due to the limited sample size.
The ecological validity of many cognitive assessments—including those used in the present study, which often involve simple button presses performed in a seated position—has been questioned38,39,40,41,42. Such static tests may not adequately reflect the dynamic cognitive–motor demands athletes face in competitive settings, thereby limiting their potential to transfer to real-world performance contexts. Rather than relying on isolated cognitive screenings, practitioners should embed cognitive demands within sport-specific tasks that more accurately reflect real-game conditions. Araujo et al.39,40 emphasized the importance of ecological cognition, highlighting the dynamic interaction between perception, decision-making, and action. Although initial evidence suggests that combining cognitive and motor tasks may enhance transfer to on-field performance43, more ecologically valid approaches are needed. Agility tasks—such as reactive change-of-direction drills, or unplanned landing and cutting tasks that require time-constrained motor reactions to visual cues, as well as dual-task scenarios (e.g., jumping or cutting combined with simultaneous cognitive demands like counting backward or performing a Stroop interference test)44—inherently integrate perceptual, cognitive, and physical demands (e.g. change of direction and velocity45.
Agility-based training and assessment tasks offer a promising approach by simultaneously targeting concentric and eccentric strength as well as perceptual and cognitive skills46. Both agility and lower extremity strength directly—which are highly related and directly affect acceleration and deceleration cycles in athletes—are critical for team sport performance and may also play a role in injury prevention47. By requiring athletes to quickly perceive visual cues, make rapid decisions, and execute explosive movements, such tasks closely replicate the multifactorial demands of actual game situations. This integrated approach reflects sport-specific challenges more accurately than static, isolated tests. Initial evidence supports the effectiveness of agility-based training in improving sprint speed, change-of-direction ability, reaction time, lower-limb strength, and flexibility in team sport athletes48. Thus, beyond further exploring the injury predictive value of executive functions, research is needed to determine whether more ecologically valid approaches provide added value for injury risk assessment.
Limitations
One major advantage of our study is its prospective design, which allows us to quantify the predictive role of various cognitive performance measures on future injury occurrence in professional football players. Another strength is the high homogeneity of the sample, with both groups being similar in key characteristics such as sport, performance level, playing time, and anthropometrics. However, this study is not without limitations: Due to multicollinearity among the various TMT measures, separate regression analyses were necessary, precluding the identification of the single most predictive cognitive outcome within one comprehensive model. Nevertheless, the primary model yielded lower AIC and BIC values, indicating superior parsimony and thereby supporting the chosen analytical approach. Due to the limited sample size, the potential impact of playing positions on cognitive performance was not explored.
Injury diagnoses were made by each team’s medical staff using their usual procedures in real-world settings, which supports the ecological validity of the data, but these diagnoses were not centrally reviewed or validated by the authors. Although all teams employed licensed professionals and followed standardized definitions22, some variability in diagnostic accuracy and reporting consistency cannot be ruled out. Only musculoskeletal injuries from the observed seasons were reported, leaving the potential impact of head injuries (e.g., concussions) on cognitive performance and injury risk unclear49. Severe primary injuries (e.g., ACL tears) are a significant risk factor for re-injuries50, but since previous injuries were not reported, their influence on injury occurrence remains speculative. We only received the total number of injuries and downtime, so it was unclear how downtime was distributed to players with multiple injuries. This limited our ability to examine the effect of cognitive performance on injury severity. Finally, only four teams provided injury data, meaning the exclusion of potential selection bias and the representativeness of the results for the entire league cannot be guaranteed. Furthermore, our results regarding the relationship between cognitive performance and injury risk may only be partially generalizable to players in other national leagues, as varying performance levels could influence this association. However, when considering injury incidence rates alone, our data appear comparable to those reported in other professional leagues. Lopez-Valenciano et al.2 demonstrated that injury incidence in the top five European leagues does not significantly differ from that in other countries’ professional leagues. Additionally, we acknowledge limitations including the lack of data on injury type, mechanism, and history (e.g., ACL reconstruction), which could act as confounding variables influencing results and should be considered in future research.
A post-hoc sensitivity analysis based on effect sizes from comparable studies (e.g., Swanik et al., 200751 suggests that a sample of approximately 128 participants would be required to reliably detect medium effects (Cohen’s d = ~ 0.5, Alpha = 0.05, power = 0.80). Our sample of 78 players may therefore have limited statistical power to detect such effects.