Predictors of Latent Reaction Speed in Athletes: The Role of Performance Level and Stress Tolerance at Different Competitive Levels

Abstract

Background

Reaction speed represents a fundamental cognitive-motor capacity in sport, yet empirical findings regarding its psychological predictors remain inconsistent, particularly across stages of athletic expertise. This study aimed to investigate the associations of performance level and stress tolerance with latent reaction speed in athletes and to determine whether these associations differ across competitive levels.

Results

A cross-sectional sample of 304 competitive athletes (amateur, pre-elite, and elite) completed standardized computerized psychophysiological assessments. Latent reaction speed was modeled as a unified cognitive-motor construct indicated by simple and choice reaction speed and motor execution measures. Performance level and stress tolerance were assessed as task-based indices of performance efficiency and cognitive load tolerance under standardized conditions. Structural equation modeling was used to examine predictors of latent reaction speed, followed by multi-group analyses across athlete competitive levels. Both performance level (β = 0.19, p = 0.004) and stress tolerance (β = 0.17, p = 0.010) showed small but statistically significant positive associations with latent reaction speed, together explaining 14.2% of its variance. Age was negatively associated with latent reaction speed, while gender effects were small. Multi-group analyses revealed that performance level (β = 0.28, p < 0.001) and stress tolerance (β = 0.17, p = 0.028) significantly predicted latent reaction speed only among pre-elite athletes, whereas no significant associations were observed in amateur or elite groups.

Conclusions

These findings suggest that the contribution of performance-related psychological factors to reaction speed is stage-specific, with relevance during the pre-elite phase of athletic development. The findings indicate statistically significant but modest associations between performance-related psychological factors and latent reaction speed across stages of athletic development.

Key Points

1. Performance level and stress tolerance are positively associated with latent reaction speed in athletes, indicating that higher task efficiency and better tolerance to cognitive load are linked to faster cognitive-motor responses.

2. These associations are stage-specific, emerging only in pre-elite athletes, while no significant relationships were observed in amateur or elite groups.

3. Developmentally targeted training approaches may be warranted, as cognitive-motor performance appears most responsive to performance-related psychological factors during the pre-elite phase of athletic development.

Background

Reaction Speed in Athletic Performance

Reaction speed is widely used in sports science to describe athletes’ ability to detect, process, and respond to task-relevant stimuli under time constraints [1, 2]. Faster reaction speed is often considered a distinguishing characteristic of high-level athletes, with previous research indicating that trained athletes may exhibit substantially shorter reaction times than non-athletes [3]. Importantly, reaction speed reflects a composite cognitive-motor process, integrating perceptual information processing with motor execution, rather than a purely reflexive response. Evidence from dual-task research indicates that concurrent cognitive demands can alter balance control and motor coordination in athletes, further supporting the view that reaction performance reflects an integrated cognitive-motor process rather than an isolated reflexive response [4].

Empirical findings, however, remain mixed regarding the direct relationship between reaction speed and sport performance. While some studies report faster simple and choice reaction times in physically active individuals and athletes [2], others have failed to demonstrate clear associations between reaction time measures and competitive performance indicators, particularly at higher levels of expertise [5]. Experimental training interventions have likewise yielded inconsistent effects, with improvements often limited to specific reaction components or task types [1].

These divergent findings suggest that reaction speed should be conceptualized not as a single observable variable, but as a latent cognitive-motor construct reflecting coordinated perceptual processing and motor execution efficiency. Computerized psychophysiological assessments, such as those provided by the Vienna Test System (VTS), allow for objective and standardized measurement of these underlying components, thereby offering a more nuanced understanding of reaction-related performance in athletes [6]. The importance of psychometrically sound assessment tools in sport psychology has also been emphasized in research examining the development and validation of psychological instruments for athlete populations [7].

Performance Level as an Indicator of Skill and its Link to Reaction Speed

In sport psychology, performance level is determined not only by competition results but also by the athlete’s ability to execute tasks effectively and consistently while maintaining a high level of attention, optimal pacing, and accuracy over time. Performance quality in sport reflects not only physical preparation but also cognitive-motor efficiency and the ability to maintain stable task performance under demanding conditions [8].

A consistently high-performance level is therefore considered an indicator of skill and mastery, as top-level athletes typically maintain precision and intensity throughout competition. For example, research on handball players has shown that athletes with higher levels of attentional focus and concentration are better able to make effective decisions during the most demanding moments of a match [6]. It can be hypothesized that performance level is positively associated with latent reaction speed, as athletes capable of sustaining high cognitive functioning are also more likely to demonstrate faster and more stable motor output by processing stimulus efficiently while minimizing errors.

Although some studies report mixed findings regarding the relationship between reaction-related measures and performance outcomes [5], the broader body of evidence suggests that elite athletes outperform non-athletes in a range of neurocognitive domains, including sustained attention and information processing speed. A meta-analytic review further demonstrated that high-level athletes achieve superior performance in both reaction time and attention-related tasks compared to non-athletes, reflecting cognitive advantages developed through training and skill acquisition [9]. At the same time, these cognitive-motor advantages may manifest differently depending on sport-specific demands. The presence of divergent findings underscores the need for a multidimensional evaluation of performance level that integrates physical, cognitive, and psychological factors.

In the present study, performance level is operationalized as an objectively assessed index of task-based performance efficiency under standardized conditions, rather than competitive sport performance.

Stress Tolerance and Cognitive-Motor Performance Under Pressure

The development of cognitive-motor skills in sport is closely related to the broader process of athlete expertise development [8]. Research on athlete development suggests that perceptual–motor and cognitive performance capacities evolve progressively as athletes transition from early participation stages toward elite performance [10, 11]. During intermediate phases of athlete development, often described as the pre-elite stage, substantial variability may exist in perceptual processing, attentional control, and cognitive-motor efficiency [12]. Competitive sport environments are characterized by high perceptual load, time pressure, and psychological stress, all of which can negatively affect reaction speed and decision accuracy. Stress tolerance, defined as the ability to maintain stable cognitive-motor performance under such conditions, has therefore been identified as a critical psychological factor in athletic performance.

Research using psychophysiological assessments indicates that athletes with higher stress tolerance demonstrate more consistent reaction-related performance under increasing task demands, whereas athletes who experience cognitive overload show slowed responses and reduced decision accuracy [13]. Experimental studies examining cognitive fatigue further demonstrate that increased mental load may impair reaction time, perceptual processing speed, and decision accuracy in athletes performing cognitively demanding tasks [14]. Similarly, fatigue-related changes in motor technique have been documented in high-performance adolescent athletes, indicating that movement efficiency and coordination may be particularly sensitive to fatigue during developmental stages of athletic expertise [15]. While some studies report clear advantages for elite athletes in stress-based reaction tasks [16], others suggest that cognitive-motor differences may diminish once a certain level of expertise is reached [17].

These mixed findings highlight the need to examine stress tolerance not only as a general predictor of reaction speed, but also in relation to athletes’ developmental stage and competitive level. Furthermore, neuromuscular training interventions have been shown to improve movement control and reduce injury risk in athlete populations, suggesting that cognitive-motor performance is responsive to structured training during key developmental periods [18].

Study Aim and Hypotheses

The present study had a twofold aim: (1) to examine whether performance level and stress tolerance predict latent reaction speed in athletes; and (2) to investigate whether the strength of these relationships differs across athlete competitive levels (amateur, pre-elite, and elite), in order to determine whether the contribution of performance level and stress tolerance to latent reaction speed is stage-specific across athletic development.

Latent reaction speed was conceptualized as a unified cognitive-motor construct reflecting stimulus perception, information processing and response selection, as well as motor initiation and execution speed, as measured by simple and choice reaction time tasks.

Based on the theoretical framework outlined above, the following hypotheses were tested:

Hypothesis 1

Performance level will be positively associated with latent reaction speed. Athletes with higher performance-level scores are expected to exhibit faster overall reaction times (reflecting quicker perceptual processing and motor execution).

Hypothesis 2

Stress tolerance will be positively associated with latent reaction speed. Athletes who can maintain performance under high cognitive or emotional load are hypothesized to demonstrate superior reaction speed performance.

Hypothesis 3

The strength of the relationships between both predictors (performance level and stress tolerance) and latent reaction speed will vary across competitive athlete levels. Specifically, these associations are expected to differ in magnitude between amateur, pre-elite, and elite groups (indicative of stage-specific effects in athletic development).

Methods and Materials

Participants

The sample for this cross-sectional study consisted of N = 304 competitive athletes, including n = 183 males (60.2%) and n = 121 females (39.8%). Participants were on average in late adolescence (M = 19.63 years, SD = 3.64) and reported a mean sport-specific training history of 9.22 years (SD = 4.18).

The sample comprised athletes from both team sports (n = 177, 58.2%) and individual sports (n = 127, 40.5%). Team sports represented in the study included basketball, association football (soccer), volleyball, beach volleyball, ice hockey, floorball, handball, rugby union, and field hockey. Individual sports included athletics (track and field), cycling, luge, skeleton, orienteering, tennis, judo, boxing, kickboxing, table tennis, alpine skiing, cross-country skiing, figure skating, gymnastics, climbing sports (including bouldering), sport dancing, equestrian sports, canoeing, triathlons, fencing, golf, bowling, motocross, inline skating, and running disciplines.

Participants were recruited through collaboration with national sport federations, sport schools, and accredited sport clubs across Latvia. Eligibility criteria required athletes to have a minimum of three years of competitive sport experience and to be actively engaged in organized training and competition at the time of data collection. Athletes were excluded if they reported current injury, acute illness, or medical conditions that could interfere with psychophysiological assessment. All assessments were administered on an individual basis within a controlled laboratory environment.

For the purposes of group-level analyses and Hypothesis 3, athletes were classified into three competitive levels (amateur, pre-elite and elite) using objective performance-related criteria. Athletes’ classification was based on a combination of training volume, level of competitive participation, and accumulated sport-specific experience. This approach ensured that group assignment was grounded in objective performance indicators rather than subjective self-classification. In cases where athletes met criteria corresponding to more than one tier, classification was determined using a hierarchical decision rule prioritizing the highest verified level of competitive participation (e.g., national championship results or international competition participation).

Elite athletes were classified as those training within advanced performance settings, with weekly training volumes of at least 8 training sessions or a minimum of 12 h per week, and a sport-specific training history of no less than five years. In the present study, elite athletes were defined as those who had achieved podium placements at national championships or had participated in major international competitions (e.g., European, World, or Olympic events). These criteria were used to represent athletes competing at the highest performance level within their respective sport systems.

Pre-elite athletes were actively competing at organized national-level events or within structured regional and university-based leagues, with training volumes of at least five sessions or a minimum of 7.5 h per week. National-level competition was defined as participation in officially sanctioned championships organized by national sport federations and included in formal competition calendars, ensuring that athlete classification was based on standardized and verifiable performance criteria across sports. Pre-elite athletes were typically engaged in structured development systems (sport academies or national youthprograms) but hadnot yet achieved consistent involvement at the eliteinternational level. At the time of data collection, these athletes typically participated in national-level competitions or structured development leagues but had not yet achieved podium placements at national championships or qualified for major international competitions (e.g., European or World Championships). Consistent involvement was operationalized as regular participation in organized competitions and structured training programs for at least three consecutive competitive seasons.

Amateur athletes engaged in organized training at least twice per week or for a minimum of four hours weekly and primarily participated in non-professional competitive settings, such as local, regional, or lower-tier national events. These athletes did not hold professional contracts and were not involved in national level team programs.

The resulting sample consisted of 54 elite athletes (17.8%), 163 pre-elite athletes (53.6%), and 87 amateur athletes (28.6%).

Design and Data collection

This study used a cross-sectional, laboratory-based design. Data was collected over an 18-month period between May 2024 and December 2025. Data collection spanned multiple competitive seasons, testing was conducted under standardized laboratory conditions independent of training phase, thereby reducing the likelihood that seasonal variation substantially influenced the measurement outcomes. Participants completed the testing protocol individually in a controlled research setting designed to minimize external distractions and ensure consistent assessment conditions. Each testing session lasted approximately 60 min per athlete. All testing sessions were administered by trained laboratory personnel following standardized procedures to ensure consistency across participants.

Prior to assessment, athletes were informed about the study objectives, testing procedures, data processing, and confidentiality measures. Written informed consent was obtained before participation. Participants subsequently completed a short questionnaire capturing demographic information and sport-related background, including age, gender, sport discipline, training history, and competitive level.

Subsequently, athletes completed the VTS test batteries in a fixed order to ensure procedural consistency across participants: (1) Determination Test (DT), (2) Attitude Toward Work Test (AHA), (3) Reaction Test (RT), including the simple (S1) and choice (S5) reaction formats. Standardized instructions and practice trials were provided before each task to familiarize participants with the response procedures and minimize potential learning effects during testing. After completing the assessment, participants were offered individualized feedback and a standardized results report upon request.

The study was approved by the Ethics Committee of the Latvian Academy of Sport Education (Protocol No. 8, Statement No. 1, April 19, 2024) and was conducted in line with the principles of the Declaration of Helsinki. Participation was entirely voluntary, and athletes retained the right to discontinue participation at any point without justification or adverse consequences. All collected data were anonymized, stored securely with restricted access, and used exclusively for research purposes. Data handling and storage procedures adhered to the registered ARGOS Data Management Plan (OpenAIRE).

Measures

Testing procedures were administered using the VTS software (version 8.26.10; Schuhfried GmbH, Mödling, Austria). All time-critical tests were conducted using the standardized VTS USB response panel with color-coded buttons and foot-operated keys to ensure precise timing of responses. The system operates with a temporal resolution of approximately 1 ms (1 kHz), corresponding to a sampling frequency of approximately 1000 Hz [19]. Visual stimuli were presented on a standard laboratory LCD monitor configured in accordance with the Vienna Test System technical recommendations, including a screen diagonal within the range of 14–27 inches and a resolution of at least 1280 × 1024 pixels. These specifications are consistent with manufacturer guidelines for accurate stimulus presentation in time-critical testing environments. The monitor operated at a standard refresh rate of 60 Hz, which represents the typical configuration recommended for VTS-based reaction-time assessment. The same laboratory setup, including the monitor and response panel configuration, was maintained throughout the data collection period to ensure consistency of testing conditions.

Latent reaction speed, performance level, and stress tolerance were evaluated using the Vienna Test System (VTS), a computerized psychophysiological assessment battery widely applied in sport science, neuroscience and psychology for objective measurement of cognitive-motor performance. VTS demonstrates high internal consistency for the main variables across its test batteries (r = 0.90–0.99). Previous validation studies and technical documentation also report good test–retest reliability for the reaction time and determination test measures, with coefficients typically ranging between r = 0.80 and r = 0.90 [20]. Independent studies using standardized electronic reaction time measurement systems in sport populations have demonstrated acceptable reliability and validity of response time assessments [21]. Evidence for construct and criterion validity has been documented across several application domains, including sport psychology and cognitive performance assessment.

All assessments were administered individually in a controlled laboratory environment to ensure standardized testing conditions and to minimize the influence of external distractions. Athletes received identical instructions and practice trials prior to testing, thereby supporting administrative and procedural validity. The following VTS test batteries were used in the present study: (1) Determination Test (DT; assessment of stress tolerance); (2) Attitude Toward Work Test (AHA; assessment of performance level); (3) Reaction Test (RT; assessment of latent reaction speed).

The full testing procedure required approximately 45–60 min per athlete. Test outcomes were available in multiple standardized formats, including raw scores, T-scores (M = 50, SD = 10), and percentile ranks. For statistical analysis, T-scores were selected because they provide a normalized metric that allows comparison across different test domains and between individuals, independent of age and sport discipline. All variables described were retained for multivariate analyses and subsequent modeling procedures. Higher scores indicated superior performance, reflected by faster reaction and motor execution speed, greater performance efficiency, and higher stress tolerance.

Stress Tolerance

Stress tolerance was measured using the Determination Test (DT). The DT examines the capacity to sustain fast and accurate responding when exposed to concurrent visual and auditory stimuli presented under temporal pressure. The task requires sustained attention, rapid stimulus discrimination, and continuous motor responses while coping with increasing cognitive load [20]. During the task, coloured visual stimuli and auditory signals appear on the computer screen in rapid succession. Participants are required to respond as quickly and accurately as possible by pressing the corresponding-coloured buttons or foot pedals on the standardized VTS response panel. The task progressively increases in stimulus frequency, requiring athletes to maintain rapid and accurate responses under conditions of high perceptual load.

The DT is theoretically grounded in the cognitive ability model (the Cattell–Horn–Carroll), particularly reflecting components of processing speed, reaction speed, and decision speed [22]. Within this framework, stress tolerance is operationalized as the capacity to maintain stable cognitive-motor performance under conditions of high stimulus density and time pressure. This is particularly relevant in competitive sport settings, where athletes are required to interpret complex information and make rapid decisions under conditions of psychological and time-related pressure. Stress tolerance reflects the ability to maintain rapid and accurate responses under high perceptual load and time pressure conditions that closely parallel decision-making demands in competitive sport.

Performance Level

Performance level was assessed using the Attitude Toward Work Test (AHA). The AHA evaluates performance-related behavioral tendencies during continuous task execution, including sustained concentration, work pace, accuracy, and productivity over time. In the task, participants are presented with a continuous sequence of simple symbols on the screen and are required to identify and mark predefined target stimuli using the response keys. The task emphasizes sustained concentration, work pace, and accuracy over time while minimizing errors. The test provides an objective index of performance efficiency under standardized conditions [23]. Performance level was represented by the AHA overall efficiency score (standardized T-score), reflecting sustained work pace and accuracy during continuous task execution, with higher scores indicating higher task-based performance efficiency. In the present study, performance level refers to an objectively assessed performance-related efficiency under standardized task conditions, rather than competitive sport outcomes. The AHA was administered using Version 31, Revision 1 (2019), which represents the current standardized version available within the Vienna Test System at the time of data collection.

In the context of sport performance, the AHA reflects an athlete’s ability to maintain consistent task engagement, effort regulation, and goal-directed behavior, particularly in situations requiring prolonged focus and resistance to performance decrement. These characteristics are closely aligned with performance demands in competitive training and competitive environments.

Latent Reaction Speed

Latent reaction speed was measured using the Reaction Test (RT). The RT was administered in two formats: S1 assessing simple reaction and S5 assessing choice reaction. These forms yielded four reaction speed indicators: simple reaction speed (perceptual latency), simple motor speed (movement execution following stimulus detection), choice reaction speed, and choice motor speed [24].

These indicators capture both stimulus detection and processing speed as well as motor execution efficiency, under conditions with and without response selection demands. During the reaction task, visual stimuli appear on the computer screen and participants must respond by pressing the designated response button as quickly as possible. In the simple reaction format (S1), a single stimulus requires a single response, whereas in the choice reaction format (S5) different stimuli require selecting the corresponding response button from several alternatives. RT therefore provides a comprehensive assessment of sensorimotor responsiveness and decision-making efficiency, which represent fundamental cognitive-motor capacities in many sport disciplines. Reaction speed variables were automatically registered by the Vienna Test System as the time interval (in milliseconds) between stimulus presentation and the participant’s button press response. Perceptual latency reflects stimulus detection and processing time, whereas motor speed represents the execution time of the response. Incorrect responses and omissions are also recorded by the system, meaning that faster responses that lead to errors do not improve task performance, thereby reducing potential speed–accuracy trade-offs [24].

The term “latent” is used in a statistical sense to denote a structural equation modeling construct representing shared variance across multiple reaction and motor execution indicators and does not refer to a specific temporal component (perceptual or decision latency) within the reaction process.

Statistical Data Analysis

Statistical analysis was carried out using IBM SPSS Statistics and the R statistical computing environment [25]. SPSS was used for data screening and descriptive analyses, whereas structural equation modeling was performed in R. Means, standard deviations, and observed ranges were calculated for all study variables. Pearson correlation coefficients were computed to examine bivariate associations among reaction time measures, performance-related variables, and demographic characteristics, and were visualized using a correlation heatmap.

Prior to structural equation modeling (SEM), univariate normality was assessed using skewness and kurtosis statistics. Values within commonly accepted thresholds (|skewness| < 3 and |kurtosis| < 10) were considered indicative of acceptable normality [26]. No severe departures from normality or extreme univariate outliers were observed. Given minor deviations from normality, models were estimated using robust maximum likelihood estimation.

SEM was conducted using the lavaan package. The application of SEM to examine relationships between physical activity, cognitive performance, and psychological factors has been previously demonstrated in sport and exercise science research [2, 27]. A measurement model was first specified to define a single latent construct representing reaction-related performance (latent reaction speed), indicated by four observed variables derived from the computerized reaction time test: simple reaction speed, simple motor speed, choice reaction speed, and choice motor speed. Reaction speed was fixed to 1.00 to establish the scale of the latent factor. Model fit was evaluated using multiple indices, including the chi-square statistic (χ²), Comparative Fit Index (CFI), Tucker-Lewis Index (TLI), Root Mean Square Error of Approximation (RMSEA) with 90% confidence intervals, and the Standardized Root Mean Square Residual (SRMR). Confidence intervals and p-values were derived using the same robust maximum likelihood estimation procedure. However, minor discrepancies between statistical significance and confidence interval boundaries may occur when robust standard errors are applied, particularly in subgroup analyses with smaller sample sizes. Model fit was interpreted according to established guidelines [28]. Sample size adequacy was evaluated using both methodological recommendations for structural equation modeling and model-sensitivity assessment. According to Kline [26], SEM models of moderate complexity typically require a minimum sample size of approximately 200 observations to ensure stable parameter estimation. A post-hoc power analysis based on RMSEA criteria was conducted to assess model sensitivity within the SEM framework. Assuming RMSEA₀ = 0.05, RMSEA₁ = 0.08, α = 0.05, and df = 13, the RMSEA-based sensitivity estimate for the present sample (N = 304) was 0.938, indicating adequate sensitivity of the model to detect potential misfit rather than statistical power to detect specific structural effects and supporting the adequacy of the sample size for the primary SEM analyses [29]. However, this RMSEA-based analysis reflects model-level sensitivity to detect misfit rather than the statistical power to detect specific structural path coefficients. To address this limitation, an additional approximate Monte Carlo simulation was conducted based on the standardized parameter estimates obtained in the present study, corresponding to the small-to-moderate structural effects observed in the final model, and supplemented by sensitivity scenarios using slightly smaller and small effect sizes [30]. Across scenarios, the simulation indicated adequate-to-moderate power in the full sample, but substantially lower power in the subgroup analyses. For the observed effects scenario (β = 0.19 for performance level and β = 0.17 for stress tolerance), estimated power in the full sample (N = 304) was 0.870 and 0.726, respectively, whereas in the pre-elite group (n = 163) it was 0.618 and 0.458, in the amateur group (n = 87) 0.256 and 0.250, and in the elite group (n = 54) 0.140 and 0.096. These results suggest that the subgroup analyses, particularly in the amateur and elite groups, had limited statistical power to detect small-to-moderate effects of the observed magnitude. Therefore, non-significant subgroup findings should be interpreted with caution rather than as evidence of absence. Because the simulation was informed by the observed parameter estimates rather than externally specified effect sizes, these power estimates should be interpreted as approximate and conditional on the assumed population model. Because this analysis was conducted post-hoc using the observed sample data, it should be interpreted as an estimate of model sensitivity rather than as an a priori assessment of statistical power. No formal correction for multiple comparisons was applied, as the structural equation modeling framework evaluates parameters simultaneously within a single model, and the primary focus was on theory-driven associations rather than isolated hypothesis tests.

A structural model was then specified in which latent reaction speed was regressed on performance level (assessed using the AHA test), stress tolerance (assessed using the DT test), age (centered), and gender. Standardized path coefficients were reported to facilitate interpretation. The proportion of variance explained in latent reaction speed (R²) was computed to quantify overall model explanatory power.

To examine whether structural relationships differ across stages of athletic expertise, multi-group SEM was used to estimate group-specific structural models for amateur, pre-elite, and elite athletes. Prior to multi-group structural comparisons, measurement invariance of the latent reaction speed construct was evaluated across groups. Configural invariance was tested to establish equivalence of factor structure, followed by metric invariance, in which factor loadings were constrained to equality across groups. Scalar invariance, involving additional equality constraints on indicator intercepts, was also examined exploratorily. Invariance was evaluated primarily using changes in CFI (ΔCFI), with values close to or below 0.01 considered indicative of acceptable invariance, in line with established recommendations [31]. Because metric invariance was supported, comparisons of standardized structural path coefficients across groups were considered statistically valid. All statistical tests were two-tailed, with a significance level of p < 0.05. Because SEM estimates parameters simultaneously within a single structural model rather than through independent hypothesis tests, corrections for multiple comparisons were not applied, consistent with recommendations for model-based multivariate analyses [26].

Results

Descriptive Statistics and Correlation Analysis

Table 1 reports the means, standard deviations, and observed ranges for all reaction speed, motor speed, and performance-related variables included in the analyses. All variables are reported as standardized T-scores, with higher scores indicating better performance (faster reaction speed, higher performance level or greater stress tolerance). The mean simple reaction speed was 61.40 (SD = 8.98), with observed values ranging from 40.00 to 80.00. Simple motor speed showed a comparable distribution (M = 62.40, SD = 8.69; range = 44.00–80.00). Measures involving decision-making demands yielded higher mean scores: choice reaction speed had a mean of 67.00 (SD = 9.78; range = 44.00–80.00), while choice motor speed averaged 65.60 (SD = 8.41; range = 44.00–80.00). The mean performance level score was 61.60 (SD = 6.13), with values ranging from 47.00 to 80.00. Stress tolerance had a mean of 54.30 (SD = 7.61), with observed values ranging from 32.00 to 73.00.

Table 1.

Descriptive statistics of reaction speed, motor speed, performance level, and stress tolerance measures

No.	Variable	M	SD	Min.	Max.
1.	Simple reaction speed	61.40	8.98	40.00	80.00
2.	Simple motor speed	62.40	8.69	44.00	80.00
3.	Choice reaction speed	67.00	9.78	44.00	80.00
4.	Choice motor speed	65.60	8.41	44.00	80.00
5.	Performance level	61.60	6.13	47.00	80.00
6.	Stress tolerance	54.30	7.61	32.00	73.00

Figure 1 presents the Pearson correlation matrix among simple reaction speed, simple motor speed, choice reaction speed, choice motor speed, performance level, stress tolerance, and selected demographic variables. Strong positive correlations were observed among reaction speed indicators, most notably between simple motor and choice motor speed (r = 0.75, p < 0.001), supporting the conceptualization of a common latent reaction speed construct. Performance level and stress tolerance demonstrated small to moderate positive correlations with both simple and choice-based reaction time measures (r = 0.16–0.43, p < 0.05). Age and athlete competitive level showed weak to moderate negative correlations with reaction speed variables (up to r = − 0.22, p < 0.001), indicating slower response patterns with increasing age and athletes’ competitive level. Athlete competitive level was coded in ascending order (1 = elite, 2 = pre-elite, 3 = amateur), such that higher numeric values represent lower levels of competitive expertise. Although the reaction-time measures were expressed as age-adjusted T-scores, the residual negative association with age may reflect sport-specific performance dynamics or cohort-related differences in training experience that are not fully captured by normative adjustments. All correlations are reported for descriptive purposes and do not imply causal relationships.

Fig. 1.

Pearson correlation heatmap of reaction speed measures, performance level, stress tolerance, and demographic variables. *p < 0.05. **p < 0.01. ***p < 0.001.

Measurement model and overall model fit

Prior to testing the hypothesized structural equation model, preliminary analyses were conducted to assess data quality and statistical assumptions. All observed variables exhibited acceptable univariate distributions, with skewness and kurtosis values within recommended thresholds [26]. Inspection of bivariate relationships and residuals revealed no evidence of extreme outliers or violations that would compromise model estimation. Given minor deviations from normality, model parameters were estimated using robust maximum likelihood procedures.

The final analytical sample comprised 304 athletes representing different competitive levels (amateur, pre-elite, and elite). This sample size exceeded commonly recommended minimum ratios of cases to estimated parameters for structural equation modeling and was considered adequate for stable parameter estimation and model testing [32].

A latent variable structural equation model (SEM) was specified to examine predictors of reaction speed performance (see Fig. 2). The measurement model defined a single latent construct, latent reaction speed, reflecting general cognitive-motor processing capacity underlying reaction performance. This construct was indicated by four observed reaction measures derived from the RT test: simple reaction speed (latency of signal detection), simple motor speed (movement execution following signal detection), choice reaction speed, and choice motor speed. These indicators represent increasing levels of cognitive and motor processing demands, with the latter two involving response selection under decision-making conditions. The measurement model was subsequently embedded within a structural model specifying performance level, stress tolerance, age, and gender as predictors of the latent construct.

Fig. 2.

The structural equation model, and the corresponding parameter estimates are reported in Table 4. Note: Standardized path coefficients are shown. Latent reaction speed was modeled as a latent cognitive-motor construct indicated by reaction speed, motor speed, choice reaction speed, and choice motor speed. Performance level and stress tolerance were specified as primary predictors, age and gender were included as control variables. All depicted paths are statistically significant (p < 0.05). Gender coded as 0 = male, 1 = female. *p < 0.05. **p < 0.01. ***p < 0.001

The structural model specified performance level (AHA test), stress tolerance (DT test), age (centered), and gender as predictors of the latent reaction speed construct. Performance level reflects the efficiency of information processing and goal-directed task execution under time constraints [23], whereas stress tolerance represents the individual’s ability to maintain performance stability under conditions of increased cognitive load and stress [20]. Age and gender were included as control variables due to their well-documented associations with reaction time and motor speed performance.

The hypothesized structural equation model demonstrated a good fit to the data, χ² (13) = 28.12, p = 0.008; χ²/df = 2.16; CFI = 0.969; TLI = 0.948; RMSEA = 0.062, 90% CI [0.030, 0.093]; SRMR = 0.043; N = 304 (see Table 2). Both the Comparative Fit Index and Tucker-Lewis Index exceeded recommended cutoff values, indicating excellent comparative fit [28]. The RMSEA value suggested a close approximate fit, with the confidence interval remaining within acceptable limits [32]. Furthermore, the χ²/df ratio fell well within commonly accepted thresholds for structural equation models of comparable complexity [26].

Table 2.

Model fit indices for the structural equation model

χ²	df	p	χ²/df	CFI	TLI	RMSEA	RMSEA 90% CI	N
28.12	13	0.008	2.16	0.969	0.948	0.062	0.030–0.093	304

CFI Comparative Fit Index, TLI Tucker-Lewis Index, RMSEA Root Mean Square Error of Approximation, CI Confidence Interval

All factor loadings were statistically significant or fixed for identification purposes, supporting the validity of the latent reaction speed construct (see Table 3). Simple reaction speed was specified as the marker indicator (λ = 1.00, fixed). Simple motor speed (λ = 2.08, SE = 0.28, p < 0.001), choice reaction speed (λ = 1.19, SE = 0.17, p < 0.001), and choice motor speed (λ = 1.91, SE = 0.30, p < 0.001) all loaded strongly and significantly on the latent factor. Standardized loadings indicated that motor speed and choice motor speed contributed most strongly to latent reaction speed, reflecting the combined contribution of perceptual-motor processing and response execution components. Notably, simple reaction speed showed a comparatively weaker standardized loading (Std. λ = 0.41), suggesting that the latent construct was driven more strongly by motor execution and choice-related components than by perceptual latency alone.

Table 3.

Factor loadings for latent reaction speed construct

Latent Variable	Indicator	λ	SE	p	Std. λ
Latent reaction speed	Simple reaction speed	1.000	–	–	0.41
Latent reaction speed	Simple motor speed	2.077	0.279	< 0.001	0.89
Latent reaction speed	Choice reaction speed	1.192	0.168	< 0.001	0.45
Latent reaction speed	Choice motor speed	1.909	0.304	< 0.001	0.84

Factor loadings (λ) are unstandardized estimates. Simple reaction speed was fixed to 1.00 for model identification. Std. λ = standardized factor loading

Structural Model Predicting Latent Reaction Speed

The structural equation model examined the relationships between performance level, stress tolerance, and latent reaction speed (see Table 4; Fig. 2). Latent reaction speed was specified as a unified cognitive-motor construct indicated by four reaction time measures derived from computerized testing: simple reaction speed, simple motor speed, choice reaction speed, and choice motor speed.

Table 4.

Path coefficients for the structural model

Predictor		Outcome	β	SE	p
Performance level	→	Latent reaction speed	0.19	0.04	0.004
Stress tolerance	→	Latent reaction speed	0.17	0.03	0.010
Age	→	Latent reaction speed	– 0.20	0.06	0.001
Gender	→	Latent reaction speed	0.12	0.05	0.045

Standardized path coefficients (β)

Performance level, assessed using the AHA test, showed a positive and statistically significant association with latent reaction speed (β = 0.19, SE = 0.04, p = 0.004), indicating that athletes with higher performance level scores demonstrated faster perceptual processing and motor execution across both simple and choice reaction tasks.

Stress tolerance, measured using the DT test, was also positively associated with latent reaction speed (β = 0.17, SE = 0.03, p = 0.010), indicating that athletes who are better able to maintain performance under cognitive and emotional pressure exhibited faster reaction speed performance.

Age was negatively associated with latent reaction speed (β = −0.20, SE = 0.06, p = 0.001), indicating slower reaction speed performance with increasing age within the athletic sample. Gender showed a small positive association with latent reaction speed (β = 0.12, SE = 0.05, p = 0.045), indicating slightly higher latent reaction speed scores among female athletes compared to male athletes, however, the magnitude of this effect was modest. Because reaction speed indicators were analysed using standardized T-scores derived from normative data that may partially account for age and gender related differences, the observed association should be interpreted cautiously and does not necessarily reflect a true gender-based performance advantage.

At the measurement level, all reaction speed indicators loaded significantly on latent reaction speed. Motor speed and choice motor speed exhibited the highest standardized loadings, indicating that motor execution speed contributed more strongly to the latent construct than perceptual latency alone. This pattern supports the interpretation of latent reaction speed as a unified cognitive-motor construct encompassing both signal processing and movement execution components across simple and choice reaction paradigms.

The predictors accounted for 14.2% of the variance in latent reaction speed (R² = 0.142), indicating that reaction speed performance is influenced by performance level and stress tolerance but also substantially shaped by other factors not included in the model. Although statistically significant, the observed standardized effects were small to moderate in magnitude, indicating that performance level and stress tolerance represent contributory rather than dominant determinants of latent reaction speed. This magnitude of explained variance is consistent with prior research on cognitive-motor reaction speed, which is known to be influenced by multiple neurocognitive, neuromuscular, and task-specific factors [33, 34].

Measurement Invariance and Multi-Group Structural Models Across Athlete Competitive Levels

To examine whether the associations between performance level, stress tolerance, and reaction speed performance differed across stages of athletic expertise, separate structural equation models were estimated for amateur, pre-elite, and elite athletes. Prior to multi-group structural comparisons, measurement invariance of the latent reaction speed construct was evaluated across competitive athlete levels.

Configural invariance was supported, indicating that the same latent factor structure was applicable across amateur, pre-elite, and elite athletes. However, absolute fit indices for the configural model indicated suboptimal model fit (χ²(6) = 82.443, p < 0.001, CFI = 0.818, RMSEA = 0.355, and SRMR = 0.101). An RMSEA value of this magnitude indicates substantial model misspecification in at least one subgroup and therefore warrants cautious interpretation of multi-group comparisons. Evaluation of measurement invariance therefore focused on changes in comparative fit indices rather than absolute fit values, consistent with established recommendations for multi-group SEM [31]. Imposing equality constraints on factor loadings (metric invariance) resulted in a small change in model fit (CFI = 0.829; ΔCFI = + 0.01), suggesting that the measurement structure and factor loadings were largely equivalent across athlete competitive levels.

Imposing additional equality constraints on item intercepts (scalar invariance) led to a reduction in model fit (CFI = 0.812; ΔCFI = − 0.017), indicating that full scalar invariance was not achieved. However, because the primary focus of the present study was on comparing structural relationships rather than latent mean differences, the establishment of metric invariance was considered sufficient for multi-group structural equation modeling [31]. Accordingly, comparisons of standardized path coefficients across athlete competitive levels were considered methodologically appropriate.

In all models, latent reaction speed was specified using the same four reaction speed indicators, ensuring measurement consistency across groups. Standardized path coefficients are reported in Table 5 and illustrated in Fig. 3.

Table 5.

Path coefficients predicting latent reaction speed across athlete competitive levels

Predictor	Elite (n = 54)			Pre-elite (n = 163)			Amateur (n = 87)
	β	p	95% CI	β	p	95% CI	β	p	95% CI
Performance level	– 0.02	0.84	[– 0.31, 0.27]	0.28*	< 0.001	[0.15, 0.41]	0.20	0.15	[– 0.07, 0.46]
Stress tolerance	0.19	0.106	[– 0.04, 0.42]	0.17*	0.028	[– 0.04, 0.42]	0.15	0.41	[– 0.19, 0.49]

*p < 0.05. For the pre-elite subgroup, the 95% confidence interval for the stress tolerance path includes zero despite a statistically significant p value (p = 0.028). This finding should be interpreted cautiously and considered exploratory

Fig. 3.

Comparison of standardized paths to latent reaction speed across athlete competitive level. Note: Error bars represent 95% confidence intervals of standardized path coefficients. Corresponding confidence intervals are reported in Table 5. The horizontal reference value of β = 0 indicates the null effect threshold. Estimates for the elite group should be interpreted cautiously, as wide confidence intervals reflect substantial sampling variability associated with the relatively small subgroup size.

The pattern of predictors differed across athlete competitive levels (see Fig. 3). Among pre-elite athletes, performance level demonstrated a statistically significant positive association with latent reaction speed (β = 0.28, p < 0.001), and stress tolerance also showed a significant positive association (β = 0.17, p = 0.028). Although the path coefficient for stress tolerance in the pre-elite group reached statistical significance, the corresponding confidence interval included zero, suggesting that this effect should be interpreted with caution.

Among elite athletes, neither performance level nor stress tolerance significantly predicted latent reaction speed. Similarly, among amateur athletes, neither predictor showed statistically significant associations with latent reaction speed.

Taken together, these findings indicate that the relationship between latent reaction speed and performance-related factors differs across stages of athletic development, with performance level and stress tolerance emerged as the strongest significant predictors during the pre-elite phase.

These findings indicate that the associations between performance level, stress tolerance, and latent reaction speed differ across competitive athlete levels. The absence of statistically significant associations in the elite and amateur groups should be interpreted cautiously, given smaller subgroup sample sizes and potentially restricted variance in reaction speed at the extremes of expertise. Because statistical significance can vary across groups due to power and range restriction, these results should be interpreted as stage-specific associations rather than definitive evidence that coefficients differ in magnitude between groups.

Discussion

Key Findings

The present study examined whether performance level and stress tolerance predict latent reaction speed in athletes and whether these associations differ across stages of athletic expertise. In the athletes’ sample, both performance level and stress tolerance showed statistically significant positive associations with latent reaction speed, supporting Hypotheses 1 and 2. Athletes who demonstrated higher task-based performance efficiency and greater tolerance to cognitive load exhibited faster reaction-related performance across both simple and choice reaction paradigms.

A small gender effect favoring female athletes was observed; given the modest magnitude, this finding should be interpreted cautiously and may reflect sample composition (sport type, age distribution) rather than a generalizable sex difference in reaction performance.

Multi-group analyses further revealed that these associations were not uniform across athletes’ competitive levels. Performance level and stress tolerance emerged as significant predictors of latent reaction speed only among pre-elite athletes, whereas no statistically significant relationships were observed in amateur or elite groups. Accordingly, Hypothesis 3 was partially supported, indicating that the contribution of performance-related factors to reaction speed performance is stage-specific and appears to be particularly pronounced during the pre-elite phase of athletic development.

Performance Efficiency and Stress Tolerance as Cognitive-Motor Correlates of Reaction Speed

The positive associations observed between performance level (AHA-derived task efficiency), stress tolerance, and latent reaction speed in the athletes’ sample suggest that cognitive-behavioral efficiency under standardized task conditions is meaningfully related to cognitive-motor responsiveness in athletes. Performance level assessed by the AHA test battery of VTS reflects sustained concentration, work pace, and accuracy during continuous task execution which are capacities that have been previously linked to faster information processing and more efficient perceptual decision-making in sport and cognitive-motor tasks [35, 36].

In this context, sustained task efficiency can be interpreted as a behavioural indicator of performance monitoring capacity, reflecting the ability to maintain stable output under cognitive and time constraints. These characteristics are closely aligned with efficient perceptual processing and motor execution, which likely facilitate faster responses across both simple and choice reaction paradigms. In simple paradigms, participants respond to a single stimulus with a predefined response, whereas in choice paradigms the correct response must be selected from multiple alternatives [3].

Stress tolerance, measured through the DT test battery of VTS, captures the ability to maintain rapid and accurate responses under conditions of high perceptual load and time pressure. Prior research has shown that the ability to regulate cognitive resources under stress is associated with superior reaction and decision-making performance, particularly in dynamic and time-constrained environments typical of competitive sport [37, 38]. The present findings are consistent with this literature, indicating that athletes who better tolerate cognitive load demonstrate enhanced reaction-related performance. Experimental research also shows that mental fatigue and sustained cognitive effort can impair reaction time and decision accuracy in athletes, highlighting the importance of maintaining cognitive resources during demanding tasks [14].

Although the observed effects were small to moderate and the predictors explained a limited proportion of variance in latent reaction speed, this pattern is consistent with prior research suggesting that reaction performance is a multifactorial construct influenced by multiple cognitive and motor processes represented within the latent variable. Taken together, these results support the conceptualization of latent reaction speed as a unified cognitive-motor construct influenced by higher-order performance-related capacities, while also confirming that reaction speed is a multifactorial trait shaped by neurocognitive, neuromuscular, and task-specific factors [3, 34]. This pattern suggests that the latent reaction speed construct in the present study primarily reflects motor execution efficiency rather than perceptual detection speed, highlighting the functional importance of rapid response implementation in sport-related tasks.

Differences in Cognitive-Motor Predictors Across Athletic Expertise

An important contribution of the present study lies in demonstrating that the relationship between performance related factors and reaction speed differs across stages of athletic competitive level. Among pre-elite athletes, both performance level and stress tolerance emerged as significant predictors of latent reaction speed, indicating that individual differences in cognitive-behavioral efficiency are particularly relevant during this developmental phase. Given the relatively small subgroup size and the confidence interval spanning zero, this association should be considered exploratory and interpreted with caution. These findings should be interpreted as reflecting stage-specific associations rather than definitive evidence of differences in effect magnitude between groups, given potential power limitations and restricted variance at the extremes of expertise.

One possible interpretation is that the pre-elite stage represents a period of heightened differentiation in cognitive-motor capacities [37, 38]. This pattern is consistent with developmental frameworks suggesting that transitional stages of athlete development are characterized by asynchronous maturation of physical, cognitive, and psychological systems, resulting in increased interindividual variability in performance-related capacities. Athletes at this level are typically exposed to increasing training demands, competitive pressure, and performance expectations, while their cognitive-motor systems may still be undergoing refinement and consolidation. Previous research on expertise development suggests that this transitional phase is characterized by substantial interindividual variability, during which cognitive control, attentional regulation, and stress-related coping processes play a more pronounced role in performance outcomes [39, 40]. Under such conditions, individual differences in task efficiency and stress tolerance may exert a stronger influence on reaction speed performance. Evidence from sport training research also suggests that structured sport participation during developmental stages can facilitate both motor and cognitive development. Studies on youth athletes have shown that systematic training in sports is associated with improvements in motor coordination, attention, and cognitive processing speed, indicating that sport-specific practice may support cognitive-motor development during the transition toward higher levels of athletic expertise [11, 41, 42].

The absence of significant associations in amateur athletes may reflect insufficient training exposure or limited task specialization for higher-order performance-related capacities to meaningfully influence reaction speed. At earlier stages of skill acquisition, reaction-related performance is often driven primarily by basic sensorimotor speed rather than by advanced cognitive regulation or performance efficiency mechanisms [3, 43].

For elite athletes, reaction speed is likely to be highly automatized and optimized through extensive deliberate practice, resulting in restricted variability, potential ceiling effects, or range restriction in performance-related indicators that reduce observable associations with general performance-related capacities measured under laboratory conditions. Research on high-level athletic performance consistently indicates that elite athletes rely on highly specialized and domain-specific perceptual-motor adaptations that are robust to variations in general cognitive efficiency [44, 45]. Consequently, laboratory-based measures of performance efficiency and stress tolerance may be less sensitive to performance-relevant differences in this group.

Implications for Training and Athlete Development

The current findings suggest that the relationship between latent reaction speed and its predictors (performance level and stress tolerance) appears to depend on the developmental stage of the athlete. This observation highlights the importance of adopting developmentally informed training approaches rather than applying uniform intervention strategies across all levels of athletic competence [39, 40].

Consistent with the observed stage-specific effects, performance level and stress tolerance emerged as significant predictors of latent reaction speed only among pre-elite athletes, suggesting that targeted training interventions may be particularly relevant during this developmental phase to support athlete progression and successful transition to elite performance. Training programs for pre-elite athletes may therefore benefit from the systematic integration of cognitively demanding reaction tasks, including exercises that combine rapid stimulus discrimination, decision-making, and motor execution under time pressure. Because the present study employed a cross-sectional design, the current findings cannot determine whether such training interventions would directly improve reaction speed performance in this population. Experimental intervention studies further indicate that structured coordination and rhythm-based training programs can improve reaction speed and motor performance in developing athletes, highlighting the benefits of integrating cognitively demanding motor tasks into pre-elite training environments [46]. Furthermore, the significant contribution of stress tolerance highlights the value of incorporating psychological skills training aimed at improving stress regulation, attentional control, and emotional stability, particularly within performance-relevant and sport-specific contexts.

Consistent with prior evidence, elite athletes do not necessarily demonstrate superior performance in general cognitive functions, such as simple reaction time or basic attention tasks, compared to less experienced athletes or even non-athletes [45]. In the context of the present study, this evidence is consistent with the finding that performance level and stress tolerance no longer significantly predict latent reaction speed in elite athletes. At this level of expertise, the mechanisms underlying reaction performance are likely to be highly stabilized and automatized, thereby reducing the relevance of general cognitive and psychological factors for explaining individual differences. From a practical perspective, this suggests that training interventions for elite athletes should place less emphasis on enhancing general reaction speed or domain-general cognitive abilities and instead prioritize the development of sport-specific perceptual-motor processes, anticipatory skills, and contextual decision-making under competitive conditions. Recent meta-analytic evidence also indicates that EEG-based neurofeedback training can enhance cognitive, attentional, and motor performance in elite athletes, suggesting that targeted neurocognitive interventions may represent a promising complementary approach for high-performance sport environments [47].

For amateur athletes, the absence of significant associations between performance level, stress tolerance, and latent reaction speed suggests that reaction-related performance at this stage may be primarily influenced by more fundamental factors, such as basic motor coordination and limited exposure to time pressure or competition-like situations. Evidence from systematic reviews indicates that reaction time represents a trainable cognitive-motor capacity across different sports and levels of expertise [44]. Most studies report significant improvements in reaction time following structured visual and perceptual-cognitive training, with particularly strong effects observed for choice reaction time. Gradual exposure to reaction-based tasks under controlled and progressively increasing time constraints may therefore help amateur athletes develop more stable and efficient cognitive-motor responses. Recent evidence suggests that structured physical exercise can support cognitive performance during developmental stages, highlighting the adaptive relationship between motor activity and cognitive functioning in youth populations [48]. In addition, emerging technology-assisted performance monitoring systems may support the integration of cognitive-motor assessment into routine athlete development and performance evaluation.

Limitations and Future Directions

Several limitations of this research should be acknowledged. First, the cross-sectional design excludes causal inferences regarding the relationships between performance level, stress tolerance, and latent reaction speed. Although the use of SEM allows for the examination of theoretically grounded associations, longitudinal designs are required to clarify the developmental dynamics of cognitive-motor predictors across stages of athletic expertise.

Second, while the overall sample size was adequate for SEM analyses, the elite athlete subgroup was relatively small. This may have reduced statistical power to detect subtle associations within this group and warrants cautious interpretation of null findings among elite athletes. Given that elite athletes represent a highly selective population competing at the highest national or international levels, the attainable sample size in such studies is naturally limited. Future studies with larger elite samples or targeted recruitment of high-performance athletes are needed to confirm the observed pattern of results.

Third, all cognitive-motor measures were obtained under standardized laboratory conditions using computerized test batteries. Although this approach ensures high internal validity and measurement reliability, it may not fully capture the complexity and contextual demands of sport-specific reaction performance. In addition, the use of a 60 Hz refresh-rate monitor may introduce minor temporal variability in stimulus presentation (approximately 16.67 ms), representing a technical limitation inherent to standard laboratory display systems. Reliability coefficients reported in the Vienna Test System technical documentation indicate good test–retest stability. However, these estimates are based on general normative samples and may not fully generalize to highly trained athletic populations characterized by specific performance profiles. Test–retest reliability estimates for the VTS tasks were derived from manufacturer normative samples, and athlete-specific reliability data were not available for the present study. Furthermore, the use of standardized T-scores derived from normative reference samples assumes comparability between the study population and the normative distribution; therefore, potential differences between competitive athletes and general population samples should be considered when interpreting standardized performance indices. Future research could complement laboratory-based assessments with ecologically valid tasks embedded in sport-relevant scenarios.

Fourth, the sample included athletes from a wide range of individual and team sports. While this enhances the generalizability of the findings, it may also obscure sport-specific differences in reaction speed and its cognitive-motor determinants. Future investigations may benefit from examining open- and closed-skill sports separately or focusing on single sport disciplines. The present study did not include sport-specific or sex-specific subgroup analyses, which represents an important direction for future research given potential variability in cognitive-motor performance across sport disciplines and athlete characteristics.

Finally, the present model focused on performance level and stress tolerance as key predictors of latent reaction speed. Other potentially relevant factors, such as fatigue, sleep quality, injury history, motivational states, or sport-specific tactical demands, were not included. Incorporating these variables in future models may provide a more comprehensive understanding of the multifactorial nature of reaction speed performance in athletes.

Further methodological consideration concerns the model fit of the configural invariance model across athlete groups in the multi-group SEM analysis. The CFI value for the configural model (CFI = 0.818) was below commonly recommended thresholds for acceptable model fit. However, measurement invariance was evaluated primarily using changes in CFI (ΔCFI), which remained within recommended limits [31]. Because the factor structure was conceptually identical across groups and metric invariance was supported, comparisons of standardized structural path coefficients were considered interpretable. Nevertheless, the relatively low absolute fit suggests that group-specific findings should be interpreted with caution. An RMSEA value of this magnitude indicates substantial model misspecification in at least one subgroup and therefore warrants cautious interpretation of multi-group comparisons. This pattern may reflect developmental heterogeneity in performance characteristics across stages of athletic expertise rather than a fundamental difference in the conceptual meaning of the latent reaction speed construct. Accordingly, future studies with larger subgroup samples are warranted to further examine measurement equivalence across competitive levels. Accordingly, these procedures should not be interpreted as formal statistical power analyses in the strict sense, but rather as post-hoc indicators of model sensitivity based on the observed sample. The post-hoc RMSEA-based sensitivity analysis and Monte Carlo simulation should therefore be interpreted as internal consistency checks rather than as formal a priori statistical power analyses for detecting structural effects.

Conclusion

The present study provides evidence that reaction-related performance in athletes can be conceptualized as a latent cognitive-motor construct influenced by higher-order performance-related capacities. Across the sample, performance level and stress tolerance were positively associated with latent reaction speed, highlighting the role of cognitive-behavioral efficiency in cognitive-motor responsiveness.

The findings demonstrate that these relationships are not uniform across stages of athletic expertise. Performance level and stress tolerance emerged as meaningful predictors of reaction speed primarily among pre-elite athletes, whereas no significant associations were observed in amateur or elite groups. This stage-specific pattern suggests that the pre-elite phase represents a critical window during which individual differences in reaction-related performance.

From a practical standpoint, the results underscore the importance of adopting developmentally informed assessment and training strategies. In particular, evaluating and targeting cognitive-motor efficiency and stress regulation may be especially beneficial during the pre-elite stage, whereas such general measures may be less informative at earlier or more advanced levels of expertise. Future research employing longitudinal and sport-specific approaches may further clarify the developmental mechanisms underlying reaction speed in athletes.

Author Contributions

K.V. conceived the study, developed the methodology, acquired funding, supervised the project, conducted the investigation, curated the data, performed the formal analyses, validated the findings, prepared the visualizations, and wrote the original draft of the manuscript. G.U. participated in the investigation. Z.V. and G.U. contributed to data curation, formal analysis, and validation, and assisted in writing and reviewing and editing the manuscript. All authors have read and approved the final version of the manuscript.

Funding

The author(s) declared financial support was received for the research, authorship, and/or publication of this article. This research is funded under the Grant No. RSU/LSPA-PA-2024/1–0010 of the project No. 5.2.1.1.i.0/2/24/I/CFLA/005 “RSU Internal and RSU with LASE External Consolidation” (funded by the European Union Recovery and Resilience Facility and the budget of the Republic of Latvia).

Data Availability

The used datasets can be accessed at Riga Stradiņš University Dataverse repository: Volgemute, Katrina. 2025. “Athletes Personality Traits, Psychological Skills, and Psychophysiological Performance.” Rīga Stradiņš University Institutional Repository Dataverse. https://doi.org/doi:10.48510/FK2/0B871H. For long-term access or institutional inquiries, data requests may also be directed to the RSU Research Department at research@rsu.lv.

Declarations

Ethics Approval and Consent to Participate

This study was approved by the Ethics Committee of the Latvian Academy of Sport Education (Protocol No. 8, Statement No. 1, April 19, 2024) and adhered to the ethical guidelines outlined in the Declaration of Helsinki.

Informed consent was obtained from all participants included in the study and participants were fully informed that their data would be used solely within the framework of this research. Confidentiality was strictly maintained, with all data securely stored to protect participant privacy. Participants were also informed of their right to withdraw from study at any time without penalty.

Competing Interests

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.