Influence of stimulus material and effector specificity on performance in the measurement of interference control in adolescent handball and soccer players compared to non-player controls Authors: Florian Heilmann, Leif E Langsdorf, Torsten Schubert Published: 2026 ABSTRACT AbstractThis study investigated how stimulus material and effector specificity influence interference control (a form of inhibitory control) in adolescent players. Ninety-one males aged 15–18 years (28 handball players, 34 soccer players, 29 non-player controls) completed nine modified Flanker task versions varying in stimulus type (abstract, handball-specific, soccer-specific) and response modality (finger, hand, foot). Mixed-design ANOVAs revealed significant main effects of congruency, stimulus type, and response modality. Responses were slower, and interference effects were larger in incongruent than in congruent trials. In addition, sports-specific stimuli and more motorically demanding response modalities were associated with increased response times compared to abstract arrow stimuli and simple finger responses. A congruency × stimulus interaction revealed reduced flanker effects for sports-specific stimuli, indicating that domain-specificity shapes interference processing. However, no consistent group differences were found, suggesting adolescent players do not exhibit broad cognitive advantages, or such advantages are context-dependent. Overall, the findings suggest that performance was primarily shaped by perceptual and motor requirements of the task, i.e., bystimulus complexity and response modality rather than by differences between athlete and control groups. Longitudinal research is needed to clarify how training and development interact to shape cognitive–motor control.Keywords: Executive functions, Interference control, Effector specificity, Ecological validity, Adolescent sports players Introduction In many high-paced team sports, successful performance depends on the ability to rapidly interpret and respond to a stream of complex, often conflicting, sensory information while executing sports-specific motor actions under time pressure (Kibele, 2006, Xiong & Song, 2024). These demands require the application of open-motor skills, characterized by the execution of goal-directed actions in a dynamic, unpredictable environment (Gu et al., 2019, Heilmann et al., 2022). Beyond physical performance factors such as agility, endurance, and strength, cognitive domains in general, and executive functions (EF) in particular play a decisive role in enabling athletes to select appropriate responses, suppress inappropriate ones, and adapt to ever-changing in-game conditions (Cao et al., 2024, Haugan et al., 2025, Heilmann et al., 2022). Response inhibition, a core component of EF, refers to the capacity to withhold or override a prepotent or ongoing motor response (Diamond, 2013). The ability to selectively attend to relevant information and to deal with conflicting information, such as deceptive cues, is critical in a sports context. (e.g., Güldenpenning et al., 2017, Weigelt et al., 2020). The Flanker paradigm (Eriksen & Eriksen, 1974) is a well-established paradigm to investigate selective attention and interference control by presenting target stimuli flanked by congruent or incongruent distractors. Executive functions encompass several partially separable control processes, including response inhibition and interference control (Diamond, 2013). Response inhibition typically refers to the suppression or cancellation of a prepotent or ongoing motor action, as measured, for example, with Go/No-Go or Stop-Signal tasks. In contrast, interference control describes the ability to resolve competition between simultaneously activated stimulus–response representations, such as when distracting information activates an incongruent response tendency in the Flanker paradigm. Although interference control often involves inhibitory mechanisms, such as the suppression of task-irrelevant response activation both constructs are not identical. In the present study, we therefore use the term interference control when referring specifically to performance in the Flanker task, while acknowledging that inhibitory processes contribute to this form of control. Importantly, stimulus specificity (the degree to which stimuli match an athlete’s familiar perceptual domain) and effector specificity (the correspondence between task responses and sports-trained effectors) influence interference control performance (Abernethy et al., 2001, Mann et al., 2007). One limitation of conventional Flanker paradigms in the sports context is their limited ecological validity. Standard versions often use simple, non-sports-related stimuli (e.g., arrows or letters) and require minimal motor output, such as pressing a single key. In contrast, interference control in sports settings must be accomplished under conditions involving complex, dynamic visual stimuli and sports-specific motor responses (Hülsdünker et al., 2023). Importantly, the present study operationalizes sports relevance primarily through domain-specific visual content rather than through dynamic action information. Real sport situations are inherently dynamic and unfold over time, requiring continuous perception–action coupling and prediction of opponents’ movements. Static images allow strong experimental control and systematic manipulation of stimulus and response factors, but they represent only a simplified approximation of real-world visuomotor processing. Thus, the term sports-relevant stimulus material in the present study refers to increased domain specificity compared to abstract symbols, rather than full ecological realism. One may increase the complexity of the stimulus set and the response requirements by applying sports-relevant visual scenes rather than abstract symbols. This can also be achieved by requiring responses with both the dominant and the non-dominant effectors. Both manipulations have been shown to prolong reaction time (RT) in interference tasks, likely due to the increased perceptual and cognitive demands (Büchel et al., 2022, Buschman & Kastner, 2015, Scharfen et al., 2022). These modulations directly influence task performance measures such as flanker interference effects (Chanceaux et al., 2014). Building on this, the concept of effector specificity in interference control has recently been highlighted by Heppe & Zentgraf (2019), who demonstrated sports-related differences in inhibitory control between hand and foot responses. Their findings suggested that long-term sports-specific training leads to effector-dominant advantages, with handball players showing superior interference control for manual responses. While Heppe & Zentgraf (2019) demonstrated effector-related differences in response inhibition (e.g., interference control) using abstract arrow stimuli, their findings do not address stimulus specificity and were interpreted as reflecting partly shared control mechanisms across effectors; the present study, therefore, extends this approach by additionally manipulating sports-relevant stimulus material alongside response effectors. Recent reviews (Albaladejo-García et al., 2023, Simonet et al., 2023) have emphasized the need for modified inhibitory/interference control protocols that incorporate sports-specific perceptual and motor demands to advance our understanding of executive functions in athletic populations. An additional important consideration concerns the developmental stage of the participants. Executive functions, and inhibitory control in particular, continue to mature throughout adolescence and into young adulthood, reflecting ongoing structural and functional development of frontoparietal control networks (Casey et al., 2005, Diamond, 2013). Investigating adolescent athletes, therefore, provides a valuable opportunity to examine whether sport-specific training is already associated with differences in cognitive–motor control while these functions are still developing. Adolescence may represent a sensitive period in which maturational processes and training-related experiences interact, allowing insights into whether potential expertise-related adaptations emerge early or only become evident after full cognitive maturation. The current study directly addresses this research gap by implementing a modified Flanker paradigm that systematically varies both the stimulus domain (sports-relevant vs. abstract) and the response effector (hand vs. foot). By applying this protocol to handball players, soccer players, and a non-athlete control group, the study is designed to test whether aligning task demands with athletes’ sports-specific perceptual and motor expertise more accurately reflects cognitive–motor adaptations. Assessing both sports-specific and cross-domain transfer effects on interference control, therefore, provides a novel opportunity to extend the findings of recent reviews(Albaladejo-García et al., 2023, Simonet et al., 2020) and to generate deeper insights into the interplay between perceptual–motor expertise and executive functions in sports. Transfer question Based on these prior studies, we hypothesized that handball and soccer players would demonstrate superior interference control compared to non-athletes, with the most substantial advantages emerging when task demands align with their sports-specific perceptual and motor expertise. More specifically, athletes are expected to show reduced flanker interference when confronted with sports-relevant stimuli rather than abstract symbols, reflecting enhanced processing efficiency within familiar perceptual domains (Abernethy et al., 2001, Mann et al., 2007). Moreover, they are predicted to exhibit better inhibitory control when responding with their trained effector (hand for handball, foot for soccer) relative to their untrained effector, consistent with sports-related adaptations in motor control (Heppe & Zentgraf, 2019). Finally, while some cross-domain transfer of expertise is anticipated, the greatest performance benefits are expected when both the stimulus domain and the effector align with the athletes’ primary sports, in line with recent reviews emphasizing the need for sports-specific interference control protocols (Albaladejo-García et al., 2023, Simonet et al., 2023). Methodological question We were furthermore interested in whether more naturalistic sports-related stimulus material would show interference-related effects, i.e., longer reaction times and/or increased error rates in incongruent compared to congruent conditions. Such a result pattern is commonly reported when applying abstract stimulus material such as arrows. If that were the case, this would indicate similar interference-related processes for these different classes of stimuli (i.e., sports-related and abstract-symbolic stimuli). As a result, this would imply a validation of the common approach to use abstract stimuli for investigating interference processing, and would provide support for the assumption that, in the context of real-world sports situations, indeed, valid inferences could be drawn with both types of stimuli about the mechanisms of interference control. Methods Participants Ninety-one male participants took part in the study, including 28 handball players, 34 soccer players, and 29 non-athlete controls, aged 15 to 18 years (M = 16.57, SD = 1.61). Athletes were recruited from competitive youth teams participating in the highest or second-highest regional leagues within their respective age categories and engaged in structured training several times per week in addition to regular competition. Inclusion criteria for the athlete groups required continuous participation in organized handball or soccer training for multiple years and active competition at the time of testing. Control participants were recruited from local schools and were required not to participate in structured competitive team sports comparable to handball or soccer training. Recreational physical activity was permitted, but no participant in the control group had a formal training history in either target sport. Descriptive statistics for age, training, and experience across groups are presented in Table 1. ANOVA revealed significant group differences in age, F(2, 88) = 30.03, p <.001, η² = 0.41, with handball players being oldest (M = 17.86, SD = 0.85), followed by soccer players (M = 16.62, SD = 1.65), and controls (M = 15.28, SD = 1.03). T-tests with pooled SD (Bonferroni method) reveal significant differences between all groups (p <.01). Likewise, training experience differed significantly, F(2, 88) = 23.74, p <.001, η² = 0.38, with handball players reporting the longest practice history (M = 12.68 years, SD = 1.77), compared to soccer players (M = 7.25 years, SD = 3.99) and controls (M = 8.98 years, SD = 2.79, no history in handball or soccer) with significant differences between all groups (p <.01). In summary, significant group differences were found for age and training experience between all groups. Table 1Means, standard deviations, 95% confidence intervals, and ANOVA results for age, training age, height, and weight across groupsVariableGroupnMSD95% CI+95% CI-ANOVAAgesoccer3416.621.6517.2016.04F(2, 88) = 30.03, p <.001, η² = 0.41handball2817.860.8518.1917.53control2915.281.0315.6714.89Training agesoccer347.253.998.945.56F(2, 88) = 23.74, p <.001, η² = 0.38handball2812.681.7613.3611.99control298.982.7910.047.92M = mean; SD = standard deviation; CI = confidence interval; η² = partial eta squared. Confidence intervals represent 95% CI around the mean. Significance levels: * p <.05, ** p <.01, *** p <.001. Means, standard deviations, 95% confidence intervals, and ANOVA results for age, training age, height, and weight across groups M = mean; SD = standard deviation; CI = confidence interval; η² = partial eta squared. Confidence intervals represent 95% CI around the mean. Significance levels: * p <.05, ** p <.01, *** p <.001. An a priori power analysis was conducted using G*Power 3.1.9.4 based on the effect sizes reported by Heppe & Zentgraf (2019). To obtain a conservative minimum sample-size estimate, we modelled the simplest relevant contrast (athletes vs. controls) for a within–between interaction with two measurements (congruent vs. incongruent). Using the lower bound of the 60% confidence interval of the reported effect (converted to f = 0.33), assuming α = 0.05, power (1–β) = 0.80, and a mixed-design ANOVA, the required total sample size was N = 58. The present study included three groups (handball, soccer, controls) with a total sample of N = 91, thus exceeding this conservative minimum and providing sufficient sensitivity for the full 3 (group) × 2 (congruency) × 3 (stimulus) × 3 (response) design. The Declaration of Helsinki and the APA’s ethical guidelines were followed in the study protocol. The study was approved by the university’s ethical committee (approval number of the ethical committee of Martin-Luther-University Halle-Wittenberg: 2412HFS). Measurements The Flanker tasks were examined using Inquisit Lab 6 (Millisecond Software LLC, Seattle, WA, USA) on a 19-inch screen, a QWERTZ keyboard (PC [personal computer]; see Fig. 1). Experimental setup for the modified Flanker task. Note. Participants were seated in front of a 19-inch monitor and responded using a QWERTZ keyboard (finger responses), an external button panel (hand responses), or foot pedals (foot responses) In the present study, participants completed a series of modified Flanker tasks in which five stimuli were presented in a horizontal row, with the central element serving as the target (see Fig. 1). Each trial displayed a sequence of five figures (flanker–flanker–target–flanker–flanker) centered on the screen. Depending on the condition, the figures represented either (a) abstract arrow stimuli (“rightarrow.png”, “leftarrow.png”; available in the supplementary material), (b) handball-specific player figures with one arm extended laterally to the left or right, or (c) soccer-specific player figures with one leg extended laterally to the left or right. The central element served as the target, while the adjacent elements functioned as flankers that pointed either in the same direction (congruent trials) or in the opposite direction (incongruent trials). Stimuli were displayed at approximately 20% of the screen height (≈ 7–8 cm), subtending a horizontal visual angle of about 18°at a viewing distance of 60 cm. Flankers were positioned equidistantly from the target, centered at 26%, 38%, 50%, 62%, and 74% of the screen width. Each trial was preceded by a yellow fixation star (1 cm, ≈ 1° visual angle) presented for 1,000 ms. Response cues and instructions were displayed in Arial font (black text on a white background). Participants were instructed to respond to the direction of the target stimulus while ignoring the surrounding flankers. On congruent trials, all five stimuli pointed in the same direction, whereas on incongruent trials, the flankers pointed in the opposite direction to the target. Responses were made to the left when the target pointed left and to the right when the target pointed right (see Fig. 2). Examples of stimulus material and response modalities in the modified Flanker task. Note.Stimulus types included abstract (arrows), handball-specific (players with one arm extended), and soccer-specific (players with one leg extended) displays, each presented in congruent and incongruent conditions. Responses were given using (1) finger on a keyboard, (2) hand on a button, or (3) foot on a button The experiment consisted of nine distinct versions of the Flanker task, created by crossing two experimental factors: stimulus type and response modality. The factor stimulus type included three categories: (1) abstract arrow stimuli, (2) handball-specific player figures, and (3) soccer-specific player figures. The factor response modality comprised (1) finger responses, (2) hand responses, and (3) foot responses (see Fig. 2). Each of the nine task versions consisted of nine blocks, each containing four practice trials followed by 70 test trials. In all conditions, participants were instructed to respond to the direction of the central target stimulus while ignoring the flankers. Procedures All participants or their legal representatives gave their informed consent. Participants gave the informed consent form after they had been briefed on the procedure. The participants underwent tests at their training facilities or at school (control group). The Flanker tasks were conducted in randomized order. To minimize the effects of physical exertion, the players were assessed between 10:00 a.m. and 4:00 p.m., one hour before training. The participants had to fill out a questionnaire as well, which we will not report here, but which can be found in the supplementary material. Participants’ handedness was not formally assessed, and responses were performed according to standardized task instructions independent of individual hand dominance. Consequently, handedness was not included as a factor in the statistical analyses and should be considered when interpreting effector-related results. Data preparation For the Flanker task assessing interference control, several filtering steps were applied. First, all trials with incorrect responses were removed (⁓2.51% of all trials). Second, to account for extreme values, trials with response times below 200 ms or above 1,750 ms were excluded (e.g., Musculus et al., 2022; ⁓0.06% of all trials). Noparticipants were excluded due to incomplete datasets. Statistical analysis A 3-group × 2 congruency × 3 stimulus × 3 response mixed ANOVA was conducted separately for response times and accuracy. Group was entered as a between-subjects factor, whereas congruency, stimulus, and response were entered as within-subjects factors. Because the participant groups differed significantly in age and because executive functions continue to develop during adolescence, we additionally conducted an age-adjusted reanalysis that included age as a covariate. This analysis examined whether the observed effects, particularly those involving group, remained stable after controlling for age-related developmental differences. Greenhouse–Geisser correction was applied where appropriate. Significant main effects and interactions were followed up using Bonferroni-corrected post hoc pairwise comparisons to control for multiple testing. No outliers exceeding 1.5 times the interquartile range were identified; therefore, none were excluded. The factors had the following levels: group: handball, soccer, control; congruency: congruent, incongruent; stimulus: abstract, handball, soccer; response: finger, hand, foot. R and RStudio were used for all statistical analyses. The threshold for significance was fixed at p <.05. To resolve relevant main effects and/or interaction effects, Bonferroni-corrected post hoc simple contrasts were applied (e.g., Wilcox, 2011). The results of the age-adjusted analyses are reported alongside the original analyses to evaluate the robustness of the findings. Results Tables 2 and 3 present all relevant descriptive values. Table 2Descriptive results for response times and flanker effect across groups and conditionsGroupSoccerControlHandballMean [ms]SDMean [ms]SDMean [ms]SDResponse time [ms]abstractfinger395.85738.304398.76553.121385.51445.714Congruent trialshand370.23432.708392.94670.019372.64336.493foot418.03140.967416.02254.689413.56541.801handballfinger400.00343.783407.31952.678388.82439.005hand380.32833.712400.10672.076380.81141.360foot435.60553.249427.80559.297424.50446.812soccerfinger405.43739.330425.03666.568401.10539.386hand392.23245.768410.60568.116385.60539.787foot438.28248.117443.89760.801446.58249.278Response time [ms]abstractfinger429.06144.546427.62362.292408.25349.793Incongruent trialshand402.81635.887423.26181.266405.78844.130foot446.27941.299457.78262.915442.70652.878handballfinger418.92149.223418.66653.794401.30633.983hand392.68135.868413.13973.800393.12038.994foot440.56040.307440.56973.065436.34345.727soccerfinger414.28339.450437.53469.281409.66739.634hand405.49749.732426.01289.640393.26237.863foot453.01048.349446.83153.857454.06646.412Flanker effect [ms]abstractfinger33.20422.76728.85826.31922.73923.888hand32.58223.87830.31529.79233.14428.915foot28.24828.24741.76028.92429.14229.514handballfinger18.91821.78811.34721.78812.48222.937hand12.35320.11313.03320.88512.30816.265foot4.95527.67012.76521.87011.83915.978soccerfinger8.84715.99612.49724.7888.56318.621hand13.26421.98015.40628.4307.65815.332foot14.72727.2562.93329.9807.48431.088 Descriptive results for response times and flanker effect across groups and conditions Table 3 Descriptive results for accuracy across groups and conditionsGroupSoccerControlHandballMean [%]SDMean [%]SDMean [%]SDAccuracy [%]abstractfinger97.765.3596.864.5497.243.6Congruent trialshand99.182.0797.553.3697.014.82foot96.685.3596.574.5496.93.6handballfinger98.73.6397.454.597.133.72hand97.864.296.273.5997.133foot96.624.8495.953.5596.063.88soccerfinger97.743.3495.494.7197.133.12hand97.632.6396.573.4897.243.62foot95.165.2893.434.6195.634.56Accuracy [%]abstractfinger91.798.6794.857.3492.467.76Incongruent trialshand93.33592.666.1593.534.49foot93.995.1490.636.1190.38.3handballfinger95.385.0995.595.5895.915.26hand97.543.8697.064.3596.555.08foot95.345.7393.856.81955soccerfinger97.772.9996.324.8295.476.12hand98.023.3394.515.9994.44.45foot95.834.9494.854.4195.044.15 Descriptive results for accuracy across groups and conditions Response time Related to our research question of sports-reated modulations of interference control processes (transfer approach), neither the threeway interaction between the factors group, congruency, and stimulus, F(4, 176) = 0.45, p =.775, ηp² = 0.010 nor the three-way interaction between the factors group, congruency and response, F(4, 176) = 0.50, p =.735, ηp² = 0.011, reached significance. This was also not the case for the four-way interaction effect between the factors group, congruency, stimulus, and response, F(8, 352) = 1.62, p =.117, ηp² = 0.036. Hence, we found no evidence for the assumption of group-specific modulation of interference control processes (see Fig. 3). Mean reaction times (RTs) as a function of the factor group (control subjects, handball players, and soccer players), congruency (congruent, incongruent trials), stimulus type (handball, soccer, abstract stimuli), and response type (foot, finger, or hand responses). Error bars represent the standard error of the mean Related to our methodological approach, we found several relevant effects, such as a large main effect of congruency, F(1, 88) = 253.51, p <.001, ηp² = 0.742, indicating slower responses on incongruent (m = 424 ms) than on congruent trials (m = 406 ms) (see Fig. 4A). This indicates that our experimental setup was effectively inducing interference effects. Mean reaction times (RTs) as a function of significant main effects and interactions. Error bars represent the standard error of the mean. An asterisk indicates a significant effect, and the ns indicates a non-significant effect. Black horizontal lines indicate the comparison between conditions. A) Shows the effect of congruency on RTs, with longer RTs in the incongruent condition compared to the congruent condition. B) Depicts increased RTs to soccer stimuli compared to abstract and handball stimuli. C) Depicts the longest RTs for foot compared to finger and hand responses. D) Indicates a significant congruency effect (difference between incongruent and congruent RTs) for all stimulus types. The congruency effect for abstract stimuli was increased compared to handball stimuli and soccer stimuli, while it was not different for the sport-related stimuli. E) Depicts that soccer players show a larger difference between finger-hand responses compared to the other groups. While all comparisons between groups were non-significant The effect of the factor congruency was further specified by the interaction effect between the factors of congruency and stimulus, F(2, 176) = 64.10, p <.001, ηp² = 0.421. As can be seen in Fig. 4D, for all stimulus types, we obtained a congruency effect, all ps < 0.001, reflecting longer RTs in the incongruent compared to the congruent conditions. This effect was further increased in abstract stimuli, for which the congruency effect (m = 31 ms) was increased compared to the congruency effect with soccer stimuli (m = 10 ms), t(177.9) = 8.84, p =.001, d = 0.66, and with handball stimuli (m = 12 ms), t(172.7) = 8.31, p =.001, d = 0.63; whereas the congruency effects were comparable for soccer and handball stimuli, t(178.2) = − 0.97, p =.331, d = 0.07. Overall, in line with our prediction, this result pattern indicates that for sports-related stimulus material, similar interference effects emerge compared to when using abstract stimulus material. This provides further validation for the experimental approach. We also obtained a significant main effect of the factor stimulus, F(2, 176) = 14.31, p <.001, ηp² = 0.140, reflecting longer RTs for the soccer (m = 422 ms) compared to the handball stimuli (m = 412 ms), t(90) = −5.58, p =.001, d = 0.59 (see Fig. 4B). Similarly, RTs to soccer stimuli were prolonged compared to abstract stimuli (m = 412ms), t(90) = 4.65, p =.001, d = 0.49. In contrast, RTs to handball and abstract stimuli (m = 412 ms) remained comparable, t(90) = − 0.18, p =.99, d = 0.02. This effect pattern probably reflects the observation that visual complexity was increased for the soccer stimuli compared to the other stimulus types, thereby prolonging information processing. Furthermore, we obtained a significant main effect of the factor response, F(2, 176) = 90.09, p <.001, ηp² = 0.506 (see Fig. 4C). RTs were highest when participants responded with their foot (m = 438 ms), followed by finger responses (m = 410 ms), and fastest with their hands (m = 398 ms). In particular, foot responses were slower than finger responses, t(90) = −9.16, p =.001, d = 0.96, and finger responses were slower than hand responses, t(90) = 4.92, p =.001, d = 0.52; hand responses and foot responses also differed significantly, t(90) = −10.47, p =.001, d = 1.10. Finally, we obtained a significant interaction effect between the factors group and response, F(4, 176) = 2.94, p =.022, ηp² = 0.063, reflecting larger differences between hand and finger responses in the soccer group compared to the handball and control groups (see Fig. 4E). In detail, this was mainly driven by a heterogeneous pattern of the RT differences within the different groups, as all differences across groups were non-significant, all p´s < 0.2. Relatedly, within the soccer group, the difference between hand (m = 393 ms) and finger (m = 413 ms) responses was especially pronounced as reflected by a significant difference, t(88) = 4.39, p =.001, d = 0.47. In contrast, this was not the case in the handball group (396 ms vs. 406 ms), t(88) = 1.97, p =.15, d = 0.21, d = 0.21, and neither in the control group (402 ms vs. 411 ms), t(88) = 2.37, p =.06, d = 0.25. These results indicate that soccer players show improved effector-related processing differences compared to the other groups. All other effects were not significant, such as the main effect of the factor group, F(2, 88) = 0.92, p =.402, ηp² = 0.020. Consequently, athletes and their control counterparts were comparable in their processing speed; group × congruency, F(2, 88) = 0.56, p =.574, ηp² = 0.013; group × stimulus, F(4, 176) = 0.43, p =.787, ηp²= 0.010; congruency × response, F(2, 176) = 0.46, p =.635, ηp²; stimulus × response, F(4, 352) = 0.58, p =.678, ηp²; congruency × stimulus × response, F(4, 352) = 0.93, p =.449, ηp² = 0.011. Accuracy Related to our transfer approach research question of sports-related modulations of interference control processes, none of the relevant interactions reached significance. In particular, neither the three-way interaction between the factors group, congruency, and stimulus, F(4, 176) = 1.11, p =.353, ηp² = 0.025, nor the three-way interaction between the factors group, congruency, and response, F(4, 176) = 1.46, p =.215, ηp² = 0.032, reached significance (see Fig. 5). The four-way interaction between group, congruency, stimulus, and response was also not significant, F(8, 352) = 1.92, p =.56, ηp² = 0.042, indicating no evidence for sports-related transfer effects on interference-control performance. Overall, these results indicate no sports-related transfer effects on interference control processes. Error rates as a function of the factors group (control subjects, handball players and soccer players), congruency (congruent, incongruent trials), stimulus type (handball, soccer, abstract stimuli), and response type (foot, finger, or hand responses). Error bars represent the standard error of the mean In contrast, for our research question related to the methodological approach, we obtained several significant main effects and interaction effects. In detail, we obtained a strong main effect of the factor congruency, F(1, 88) = 67.76, p <.001, ηp² = 0.435, indicating that participants made more errors on incongruent (m = 5.3%) than congruent (m = 3.2%) trials (see Fig. 6A). This demonstrates a similar effect pattern as on the RTs and is in line with previous results (Heppe & Zentgraf, 2019). Mean error rates (%) as a function of significant main effects and interactions. Error bars represent the standard error of the mean. An asterisk indicates a significant effect, and ns indicates a non-significant effect. Black horizontal lines indicate the comparison between conditions (A) Shows the effect of congruency on error rates, with higher error rates in the incongruent condition compared to the congruent condition. (B) Depicts differences in error rates across stimulus types, with the highest error rates for abstract stimuli compared to soccer and handball stimuli. The error rates were not different between handball and soccer stimuli. (C) Shows error rates as a function of response type, with an increased error rate for foot responses compared to finger and hand responses. The error rates were not different between finger and hand responses. (D) Illustrates a significant congruency effect (difference between incongruent and congruent conditions) for abstract and handball stimuli, showing the largest congruency effect for abstract stimuli. While the congruency effect was not different between sport-related stimuli Similar to the results observed for reaction times, the effect of congruency on error rates was further specified by a significant interaction between congruency and stimulus type, F(2, 176) = 34.55, p <.001, ηp² = 0.282, indicating that the magnitude of the congruency effect depended on the stimulus category (see Fig. 6D). More specifically, a congruency effect in error rates was observed for handball and abstract stimuli (both ps < 0.001), but not for soccer stimuli (p =.269). Post hoc comparisons revealed the largest congruency effect for abstract stimuli (m = 4.7%) compared with handball stimuli (m = 1.2%), t(170.2) = 5.78, p <.001, d = 0.44, and soccer stimuli (m = 0.4%), t(169) = 7.15, p <.001, d = 0.55, whereas the difference between handball and soccer stimuli was not significant, t(179.97) = − 1.54, p =.125, d = − 0.11. We furthermore obtained a significant main effect of the factor stimulus, F(2, 176) = 17.21, p <.001, ηp² = 0.163, reflecting that participants made the fewest errors responding to the handball stimuli (m = 3.6%), followed by the error rate to soccer stimuli (m = 4%) and abstract stimuli (m = 5.1%) (see Fig. 6B). In detail, the error rate to handball and soccer stimuli was comparable, t(90) = −1.77, p =.238, d = −0.19. While the error rate was decreased for handball stimuli compared to abstract stimuli, t(90) = −5.268, p <.001, d = −0.55, which was also the case for soccer compared to abstract stimuli, t(90) = −4.007, p <.001, d = −0.42. These results indicate that sports-related naturalistic stimuli were effectively differentiated by the participants (see Fig. 6B). We also obtained a significant main effect of the factors response, F(2, 176) = 13.76, p <.001, ηp² = 0.135, reflecting the highest error rate with foot responses (m = 5.2%), followed by the error rate of finger responses (m = 3.8%), and hand responses (m = 3.7%) (see Fig. 6C). Comparisons demonstrated comparable error rates between finger and hand responses, t(90) = 0.543, p =.99, d = 0.06. In contrast, the error rates differed for finger and foot responses, t(90) = 3.860, p <.001, d = 0.40, as well as between hand and foot responses, t(90) = −4.730, p <.001, d = −0.50. These results suggest that it was more challenging for participants to respond correctly with their feet, mirroring results on the RTs, in line with previous evidence (Heppe & Zentgraf, 2019). Finally, the interaction effect between the factors congruency, stimulus, and response reached significance, F(4, 352) = 3.16, p =.014, ηp² = 0.035, indicating that congruency effects varied modestly depending on both the type of stimulus and response. All other effects failed to reach significance: the factor group did not reach significance, F(2, 88) = 1.65, p =.197, ηp² = 0.036, reflecting similar accuracies across groups. The interaction between group and congruency failed to reach significance, F(2, 88) = 0.80, p =.452, ηp² = 0.018. Furthermore, the interaction between group and stimulus failed to reach significance, F(2, 88) = 1.25, p =.292, ηp² = 0.028. The interaction effect between group and response did not reach significance, F(2, 88) = 0.77, p =.292, ηp² = 0.017. The interaction effect between congruency and response did not reach significance, F(2, 88) = 0.6, p =.941, ηp² = 0.013. The interaction effect between stimulus and response failed to reach significance, F(2, 88) = 0.53, p =.711, ηp² = 0.012. Finally, the interaction effect between group, stimulus, and response failed to reach significance, F(2, 88) = 1.03, p =.406, ηp² = 0.023. Age-adjusted reanalysis Because the groups differed significantly in age, we repeated the main analyses with age entered as a covariate. For response times, age was a significant covariate, F(1, 87) = 5.74, p =.019, ηp² = 0.062, whereas the main effect of group remained non-significant after controlling for age, F(2, 87) = 2.68, p =.075, ηp² = 0.058. Importantly, the critical group-related interactions relevant to the transfer hypothesis remained non-significant: group × congruency × stimulus, F(4, 174) = 1.04, p =.388, ηp² = 0.023; group × congruency × response, F(4, 174) = 0.80, p =.529, ηp² = 0.018; and group × congruency × stimulus × response, F(8, 348) = 1.55, p =.138, ηp² = 0.034. The main task-related effects remained significant after controlling for age, including congruency, F(1, 87) = 253.72, p <.001, ηp² = 0.745; stimulus, F(2, 174) = 14.74, p <.001, ηp² = 0.145; response, F(2, 174) = 90.88, p <.001, ηp² = 0.511; and the congruency × stimulus interaction, F(2, 174) = 65.46, p <.001, ηp² = 0.429. A comparable pattern emerged for accuracy. Age was not a significant covariate, F(1, 87) = 0.13, p =.720, ηp² = 0.001, and the main effect of group remained non-significant, F(2, 87) = 1.08, p =.345, ηp² = 0.024. The critical group-related interactions again remained non-significant after controlling for age: group × congruency × stimulus, F(4, 174) = 1.03, p =.395, ηp² = 0.023; group × congruency × response, F(4, 174) = 1.40, p =.235, ηp² = 0.031; and group × congruency × stimulus × response, F(8, 348) = 1.22, p =.286, ηp² = 0.027. The main task-related effects also remained significant, including congruency, F(1, 87) = 66.98, p <.001, ηp² = 0.435; stimulus, F(2, 174) = 17.11, p <.001, ηp² = 0.164; response, F(2, 174) = 13.64, p <.001, ηp² = 0.136; congruency × stimulus, F(2, 174) = 34.58, p <.001, ηp² = 0.284; and congruency × stimulus × response, F(4, 348) = 3.19, p =.014, ηp² = 0.035. Taken together, the age-adjusted analyses indicate that the central conclusions were not attributable to age differences between groups: performance was primarily shaped by task-related factors, whereas no consistent sport-specific transfer effects were observed. Discussion The present study examined whether adolescent handball and soccer players show sport-related advantages in interference control compared to controls, and whether sports-related stimuli elicit interference effects comparable to abstract stimuli. Based on previous research, we expected athletes to show superior interference control, particularly when stimulus material and response modality matched their sport-specific perceptual and motor experience (Abernethy et al., 2001, Mann et al., 2007, Heppe & Zentgraf, 2019). Specifically, we predicted reduced flanker interference for sports-related stimuli and enhanced performance when responding with the trained effector, namely the hand in handball players and the foot in soccer players. Because the Flanker paradigm assesses interference control arising from response competition, whereas the Stop-Signal Task primarily measures action cancellation, the present study extends previous work by targeting a related but distinct inhibitory mechanism. In addition, we examined whether sports-related stimuli produce interference effects similar to abstract stimuli, which would support the use of controlled Flanker paradigms to investigate cognitive control processes relevant to sport contexts (Albaladejo-García et al., 2023). The present results provided only partial support for our predictions. We obtained robust interference effects, which confirmed the sensitivity of the paradigm and demonstrated influences of stimulus type and response modality; however, we did not observe the expected sport-specific advantages in athletes compared to controls. Thus, the findings suggest that task-related perceptual–motor demands influenced performance more strongly than, sport-related expertise, highlighting the importance of contextual task characteristics rather than athletic status alone. Overall, the analyses revealed no main effect of group, indicating comparable performance across soccer players, handball players, and controls. In contrast, strong main effects of congruency, stimulus, and response confirmed that task demands substantially influenced performance, with slower responses for incongruent trials, sports-specific stimuli, and motorically demanding effectors. Significant congruency × stimulus and group × response interactions showed that interference varied across stimulus types and that response-related differences depended modestly on group, although post-hoc analyses revealed no consistent athlete-specific advantages (Hüttermann et al., 2014, Voss et al., 2010). These findings reinforce the notion that task demands and ecological manipulations exert a much stronger influence on inhibitory performance than group membership alone. The absence of a group main effect aligns with recent reviews that question the robustness of broad, domain-general executive function advantages in athletes (Kalén et al., 2021). While earlier meta-analyses suggested athletes outperform non-athletes on inhibition tasks (Heilmann et al., 2022, Voss et al., 2010), more recent evidence, such as the present study, suggests that such effects are inconsistent and may depend heavily on the specific task context and the match between motor and perceptual demands (Botvinick et al., 2001, Ridderinkhof et al., 2004). The robust main effects of congruency, stimulus, and response highlight the sensitivity of interference control to perceptual and motor context. Responses were slower for incongruent than for congruent trials and were influenced by stimulus type and response modality, indicating that both perceptual complexity and effector requirements shaped performance. These findings replicate and extend previous work showing that more realistic task demands increase cognitive load (Heppe & Zentgraf, 2019). From an embodied cognition perspective, response selection and conflict resolution arise from interactions between sensory input, motor representations, and task demands rather than from abstract cognitive processes alone (Barsalou, 2008). Thus, the present findings suggest that interference control is shaped by perception–action couplings. Related to the methodological validation of sports-related stimuli, interference effects were observed not only for abstract stimuli but also for sports-related stimulus material. Although abstract stimuli produced the largest interference effects in RTs and error rates, the presence of comparable effects across stimulus types suggests that similar cognitive control and conflict-monitoring mechanisms were engaged (Botvinick et al., 2001; Ridderinkhof et al., 2004). Thus, controlled Flanker paradigms using domain-specific, sports-related stimuli may provide a useful approximation for investigating interference control in sport contexts. Developmental factors provide a likely explanation for the limited group effects. Adolescence is characterized by ongoing maturation of the prefrontal cortex and frontoparietal control networks that underpin inhibition (Arain et al., 2013, Casey et al., 2005, Tervo-Clemmens et al., 2023). Sports-related neural adaptations may therefore be latent or emerging, becoming more evident only in adulthood when neural specialization stabilizes (Gordon-Murer et al., 2021, Holfelder et al., 2020). Furthermore, individual variability in adolescent development which is likely linked to hormonal, educational, and environmental factors, can overshadow subtle sports-specific cognitive differences (Gerván et al., 2024, Sammy Ahmed et al., 2024). Longitudinal research is thus essential to disentangle maturational from training-related effects. The main effect of the stimulus shows a difference between soccer and the abstract and handball stimuli. Classical flanker-task models assume participants direct attention first to the central target stimulus (e.g., letter or arrow) and ignore flankers (Eriksen & Eriksen, 1974). In our sports-specific stimulus set, however, the salient feature may not simply lie at the geometric centre (e.g., head) but may shift toward motor-relevant zones (e.g., feet for soccer players). Accordingly, it is possible that attention in the football-foot condition was directed initially toward foot position rather than head position, thereby altering the standard locus of processing and interference dynamics. Although we lack direct eye-tracking data to confirm this attentional shift, the strong main effect of stimulus and interaction trends suggests that such effector- and domain-driven attention allocation may play a role. The strong main effect of response on reaction times indicates that performance varied substantially across effectors, with slower reaction times for hand and foot responses compared to finger responses. Such differences are consistent with well-established findings that motor execution speed depends on both the size and complexity of the effector system involved (Burle et al., 2004, Schmidt & Lee, 2020). In flanker and choice-reaction paradigms, distal effectors such as fingers typically yield faster responses due to lower inertial load and shorter corticospinal conduction distances (Anson, 1989, Pfister et al., 2014). In contrast, foot responses involve larger muscles and longer transmission times, leading to systematically increased latencies (Simonen et al., 1995). Beyond biomechanics, effector differences may also reflect variations in cortical representation and sensorimotor precision: finger movements engage highly specialized motor areas with dense cortico-cortical connectivity, whereas foot actions rely on less differentiated control regions (Schieber, 2001). In the context of the present study, these response-dependent differences suggest that interference control is tightly coupled with motor-execution constraints. The finding that soccer players showed somewhat larger distinctions between finger and hand responses than the other groups could indicate that habitual foot-dominant training modifies cross-effector coordination or attentional allocation during upper-limb responses (Heppe & Zentgraf, 2019, Swinnen & Wenderoth, 2004). More broadly, the robust response effect underscores the need to consider motor-specific factors in executive function research, as effector selection itself alters the temporal dynamics of inhibition and response selection (Gálvez-García et al., 2018, Tecilla et al., 2022). A direct comparison with Heppe & Zentgraf (2019) helps contextualize the absence of clear group effects in our data. Their study used the Stop-Signal Task, which measures action cancellation, whereas our Flanker task taps interference control, meaning the two paradigms target partly different inhibitory mechanisms. They also tested fully mature adult athletes, while our 15–18-year-old participants are still in a developmental phase in which inhibitory functions continue to mature, potentially reducing observable sport-related advantages. Developmental work on stop-signal performance suggests that response inhibition (e.g., interference control) refines into early adulthood, which may contribute to the divergent findings. Moreover, the adult athletes in Heppe & Zentgraf’s (2019) study competed at a higher and more homogeneous expertise level than our youth athletes. These methodological, developmental, and expertise-related differences likely explain why their study revealed clear group effects while ours did not. The age-adjusted reanalysis further supports this interpretation. Because the groups differed significantly in age, and because adolescence is a developmental period in which executive functions are still maturing, age was included as a covariate in additional analyses. Importantly, controlling for age did not change the central pattern of findings. For response times, age significantly contributed to performance, indicating that developmental differences were related to processing speed. However, the critical group-related interactions relevant to the transfer hypothesis remained non-significant after age adjustment, including the group × congruency × stimulus, group × congruency × response, and group × congruency × stimulus × response interactions. Similarly, for accuracy, age was not a significant covariate, and the critical group-related interactions remained non-significant. At the same time, the robust task-related effects of congruency, stimulus type, and response modality remained significant for both response times and accuracy. Thus, the absence of consistent sport-specific transfer effects cannot be explained simply by age differences between groups. Rather, the age-adjusted findings strengthen the conclusion that performance in the present paradigm was primarily shaped by task-specific perceptual and motor demands, whereas group membership played only a minor role. The significant congruency × stimulus interaction reflects the observation that the congruence effect varies in magnitude depending on the stimulus. Although the classic flanker effect (incongruent > congruent) was present for all stimulus types, the interference was attenuated for the sport-specific conditions, indicating that ecological manipulations altered the nature of interference processing. The slower overall responses to handball and soccer figures suggest that ecologically valid stimuli impose greater perceptual and cognitive demands, likely due to increased visual complexity and the activation of action-related representations (Davelaar, 2012, Kalén et al., 2021, Musculus et al., 2022). In line with the embodied cognition framework (Barsalou, 2008) and models of perception–action coupling in sport (Mann et al., 2007, Hüttermann et al., 2014), sport-specific figures may engage broader neural networks encompassing visuospatial, attentional, and motor areas, thereby increasing processing time while simultaneously narrowing interference effects. By contrast, abstract arrow stimuli are perceptually simple and highly overlearned, eliciting faster but less contextually integrated responses (Ridderinkhof et al., 2021). Together, these findings suggest that increasing ecological validity enhances processing demands yet refines selective attention, leading to a reduced flanker effect despite slower overall reaction times. The findings underscore two key implications. First, ecological validity is important: interference control performance is highly context-dependent and varies substantially with perceptual and motor task demands. Traditional paradigms using simple, abstract stimuli likely underestimate real-world cognitive–motor integration. Second, the lack of robust group differences suggests caution in using executive function tasks for talent identification or selection in youth sports (Fransen, 2024, Reinhard et al., 2025). Rather than relying on domain-general cognitive measures, training programs should emphasize context-specific, perceptually grounded exercises that directly engage the sensory–motor couplings central to sports performance (Heilmann & Schubert, 2025). Next, we would like to acknowledge the limitations of our study. The cross-sectional design does not permit causal conclusions regarding training-related adaptations. Differences, or the lack thereof, could reflect pre-existing individual differences rather than consequences of sports participation. A longitudinal design following athletes across training years would be necessary to establish causality. The adolescent sample represents another limitation. Adult athletes show more pronounced adaptations that were not yet evident in our younger participants. Furthermore, accuracy rates were generally high, raising the possibility of ceiling effects that may have masked potential group differences. A further limitation concerns the static nature of the stimulus material. Although the use of photographs increased domain specificity compared to traditional abstract stimuli, real sport environments are characterized by dynamic and continuously evolving action sequences. Dynamic information provides kinematic cues that are critical for anticipation and action selection and may engage partially different perceptual and visuomotor processing mechanisms than static displays. Consequently, the present paradigm captures only a controlled approximation of sport-specific perception–action demands. Future research should therefore extend interference-control paradigms toward video-based or interactive designs to better approximate the temporal dynamics of real competitive situations and to examine whether expertise effects emerge more clearly under dynamic conditions. Furthermore, handedness was not formally assessed or controlled for, which may have influenced performance differences between manual response conditions and should be addressed in future research. Despite these limitations, the study provides important contributions to the literature. It demonstrates convincingly that interference control performance is sensitive to ecological manipulations of task condition, reinforcing the argument that executive functions are not fixed capacities but context-dependent processes. At the same time, it tempers claims of broad sports-related cognitive benefits, at least in adolescent populations. This dual message is significant for both theoretical debates in cognitive neuroscience and practical considerations in sports science. In conclusion, the present study demonstrates that interference-control performance in adolescents is strongly influenced by stimulus characteristics and response effectors, highlighting the importance of perceptual–motor task demands in executive-function research. Despite these robust task-related effects, soccer and handball players did not show consistent performance advantages over controls, suggesting that sport-related cognitive adaptations may be subtle, context-dependent, or not yet fully developed during adolescence. By specifying the conditions under which sport-specific effects may emerge, the study contributes to a more differentiated understanding of how athletic training interacts with cognitive development. Future research should employ longitudinal approaches and more ecologically representative paradigms, and may benefit from combining behavioral and neurophysiological measures to better disentangle the roles of training, maturation, and neural plasticity in shaping executive functions across development. Acknowledgements We want to thank Gina-Maria Winkler, Pascal Schneider and Tobias Donath for data curation. Author contributions F.H. contributed to the conceptualization and design of the study, conducted data sampling and data analysis, and drafted the first version of the manuscript; further contributed to critical editing and review.L.E.L. contributed to data analysis, writing (manuscript preparation), and critical reviewing and editing of the manuscript.T.S. contributed to conceptualization and design, provided supervision throughout the project, and critically reviewed and edited the manuscript.All authors read and approved the final version of the manuscript and agree to be accountable for all aspects of the work. Funding Open Access funding enabled and organized by Projekt DEAL. The authors received no financial support for the research, authorship, or publication of this article. Data availability The data that support the findings of this study are openly available at https://osf.io/c3h67/overview? view_only=635a285a86e1486eae8435d9baea92b8. Declarations Ethical approval This study was reviewed and approved by the appropriate institutional ethics committee. All procedures were conducted in accordance with relevant guidelines and regulations. Patient consent statement No identifiable patient data were used in this study. Patient consent was obtained where required, and no additional consent was necessary for the use of anonymized information. Competing interests The authors declare no competing interests. Footnotes Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. References Associated Data Data Availability Statement The data that support the findings of this study are openly available at https://osf.io/c3h67/overview? view_only=635a285a86e1486eae8435d9baea92b8.