Abstract

Acute hormonal and neuromodulatory responses to maximal exertion are key indicators of physiological stress in elite athletes, yet it is unclear whether these responses vary with pre-exercise psychological state or differ by sex. This study examined sex-specific links between pre-exercise mood and acute physiological reactions. Twelve elite female (18.75 ± 0.75years) and sixteen elite male (19.43 ± 1.09 years) rowers performed a maximal-effort 2000 m test. Pre-exercise mood was assessed using the Profile of Mood States (POMS). Venous blood collected at PRE, POST, and 3-hour recovery (3H REC) was analyzed for cortisol, testosterone, serotonin, dopamine, gamma-aminobutyric acid (GABA), and tryptophan. Two-way ANOVA and exploratory correlations evaluated sex-specific physiological and psychophysiological responses. Cortisol showed a significant Time × Group interaction (p = 0.0002), with males displaying increases at POST and 3H REC, while females showed only small, non-significant changes. Dopamine also exhibited a significant interaction (p = 0.0107): males showed PRE-POST elevations and declines during recovery, whereas females showed minimal fluctuations. Tryptophan demonstrated main effects of Time (p = 0.0437) and Group (p < 0.0001), with larger decreases in females. Psychophysiological associations were sex-specific. In females, negative mood dimensions correlated with lower cortisol at 3H REC (r = -0.63 to -0.66). In males, lower vigor and higher depression correlated with higher serotonin and lower dopamine POST (r = -0.52 to -0.56). Pre-exercise psychological state was associated with distinct acute hormonal and neuromodulatory responses to maximal exertion, with clear sex-specific patterns. Incorporating mood assessment alongside biochemical markers may support more individualized interpretation of fatigue and recovery markers in elite sport.

Trial registration:

This observational study was retrospectively registered atClinicalTrials.gov (NCT07243613).

Introduction

Elite athletes frequently undertake periods of intensified training designed to provoke physiological adaptation through repeated high-intensity sessions and limited recovery opportunities. These microcycles create a sustained physiological load that accumulates over time, making athletes more responsive to the stress of additional maximal-effort tests. As noted by Meeusen et al. (2013), the combined influence of training load and acute exercise provides an ecologically valid context for examining psychophysiological mechanisms underlying fatigue and recovery, including psychological and cognitive responses associated with accumulated training stress (Valdesalici et al., 2026).

Insufficient recovery during intensified training can lead to transient disturbances across multiple physiological systems. These include elevated cortisol secretion, an imbalance between anabolism and catabolism, increased muscle-damage biomarkers, and activation of inflammatory pathways (Halson, 2014). Such responses often precede measurable decrements in performance and are therefore considered early indicators elevated physiological strain and insufficient recovery. Studying acute responses in this context allows for a more comprehensive understanding of how athletes adapt, or show physiological and psychological responses to accumulated physical and emotional stressors. These multidimensional responses are increasingly recognized as important components of athlete monitoring in high-performance sport (Tian et al., 2026). Mood state is one of the most sensitive indicators of training-related stress. Prolonged or intensified training consistently produces increases in fatigue, tension, and confusion, accompanied by reductions in vigor and motivation (Morgan et al., 1987; Bonin et al., 2025). These shifts frequently emerge before objective performance declines and are commonly detected through routine monitoring. The Profile of Mood States (POMS), in particular, provides a validated framework for detecting early changes in psychological state associated with elevated training stress and recovery demands. Beyond its role in tracking athlete readiness, the POMS has also been associated with performance outcomes. A meta-analysis of 25 studies demonstrated that pre-competition Total Mood Disturbance (TMD) reliably forecasts competitive performance across sports (Lochbaum et al., 2021). Therefore, psychological state may be considered both a marker of accumulated training stress and a potential correlate of subsequent physiological stress reactivity.Cortisol and testosterone provide key endocrine indicators of the body’s response to intense exercise. Cortisol rises in response to metabolic or psychological stress via the activation of the hypothalamic-pituitary-adrenal (HPA) axis, while testosterone supports anabolic recovery and tissue repair through the hypothalamic-pituitary-gonadal (HPG) axis. Their testosterone to cortisol ratio (T/C) is commonly interpreted as an index of anabolic-catabolic balance. Intensified training often elevates cortisol while leaving testosterone unchanged or slightly reduced, resulting in transient reductions in T/C that signal elevated physiological strain (Rowell et al., 2018; Ficarra et al., 2024; Aydemir et al., 2025). Similar endocrine responses have been observed in athletes exposed to intensified training camps or competitive tournament loads, reflecting the cumulative physiological demands associated with repeated exercise exposure (Ostapiuk-Karolczuk et al., 2024; Pancar et al., 2026). These hormonal changes are integral to understanding the recovery process, as they influence energy mobilization, immune function, and tissue remodeling.

Acute intense exercise also affects circulating neurotransmitter-related markers that respond dynamically to exercise stress and may provide additional information about psychophysiological responses associated with fatigue and recovery. A widely accepted model suggests that the balance between serotonergic and dopaminergic activity shapes motivation, arousal, and effort regulation (Meeusen, 2006). Increased serotonergic activity has been associated with lethargy and heightened perceived exertion, while dopaminergic signaling supports alertness and motor drive. In parallel, circulating serotonin and dopamine, derived largely from peripheral sources including platelets, the gastrointestinal tract, adrenal tissue, sympathetic nerve terminals, and immune cells (Eisenhofer et al., 2004; Berger et al., 2009), are increasingly examined in exercise physiology as serum neurochemical markers of systemic responses to acute exercise stress. Although still relatively rarely investigated in elite sport settings, these peripheral markers may provide valuable complementary information for psychophysiological monitoring of fatigue and recovery in athletes (Ostapiuk-Karolczuk et al., 2025b).

Studies consistently report elevated circulating serotonin after strenuous exercise, typically accompanied by increased fatigue perception (Heijnen et al., 2016; Zimmer et al., 2016; Ostapiuk-Karolczuk et al., 2025b). Dopamine generally rises due to sympathetic activation, though the magnitude and timing vary across exercise modalities (Meeusen, 2006). When serotonin rises more than dopamine, it may indicate a shift toward monoamine-related fatigue processes. Emerging evidence suggests that γ-aminobutyric acid (GABA) may also respond to exercise and recovery interventions, potentially reflecting muscle-to-brain signaling pathways (Lyssikatos et al., 2023a; Yang et al., 2024). Although current knowledge about GABA responses to exercise is limited, available evidence suggests that GABA may clarify neurochemical responses associated with fatigue and post-exercise recovery.

In addition to these neurochemical mechanisms, biological sex may further shape physiological and psychological responses to exertion. Some studies indicate that female athletes may show more pronounced mood disturbances or reduced cortisol variability under heavy training loads (Casanova et al., 2016). Other work suggests that males exhibit stronger acute cortisol responses, whereas females display higher anticipatory tension before exertion (Beckner et al., 2024). However, not all studies report such patterns. Raglin et al. (1991) observed similar psychophysiological responses among male and female swimmers exposed to identical training regimens, indicating that sex-related differences may be context-dependent. Notably, much of the available literature focuses on male athletes (Costello et al., 2014), leaving limited evidence on sex-specific interactions between acute fatigue, mood, and endocrine responses, particularly in rowing, where female participation at the elite level has increased substantially in recent years. Moreover, rowing represents a suitable model for examining sex-specific psychophysiological responses because female and male athletes perform the same standardized competitive task and are exposed to highly comparable training structures and performance demands (Ingham et al., 2002; Cerasola et al., 2020). Maximal rowing exercise also elicits pronounced endocrine, metabolic, and neuromodulatory responses, making it an informative model for investigating acute psychophysiological stress responses in elite athletes [16.

Maximal-effort exercise provides a controlled means of eliciting acute physiological stress, enabling simultaneous assessment of psychological state and biochemical reactivity. Despite extensive research on endocrine responses, mood disturbance, and neurotransmitter dynamics, the relationships between pre-exercise mood and acute hormonal or neuromodulatory responses remain poorly characterized. Addressing this gap requires an integrative psychophysiological approach capable of capturing sex-specific patterns of acute reactivity.

The present study aimed to examine whether pre-exercise mood state was associated with acute hormonal and neuromodulatory responses to a maximal-effort 2000 m rowing test in elite athletes, and whether these responses differed between females and males. We hypothesized that maximal exertion performed after intensified training would induce increases in cortisol and serotonin concentrations together with alterations in dopamine-related responses, and that these responses would differ between female and male athletes. We further hypothesized that pre-exercise mood state would be associated with endocrine and peripheral neurochemical responses during recovery, with sex-specific patterns of association. By integrating mood assessments with acute biochemical responses to maximal exertion, this study was designed to identify sex-specific patterns relevant to fatigue and short-term recovery in high-performance sport.

Materials and methodsEthics approval

The study was approved by the Bioethics Committee of the Nicolaus Copernicus University in Toruń, Collegium Medicum in Bydgoszcz (decision no. KB 312/2024), and conducted in accordance with the Declaration of Helsinki, and retrospectively registered at ClinicalTrials.gov (NCT07243613; 19 November 2025), as an observational study to ensure transparency and compliance with reporting standards. All participants were informed about the study procedures and potential risks and provided written informed consent before participation. Participation was voluntary, and athletes could withdraw from the study at any time without providing a reason.

Participants

Twelve trained female and sixteen trained male rowers from the Polish Youth National Rowing Team took part in this study. All participants were medically fit and free from any acute or chronic health problems.

All participants completed the same 7-day intensive training camp in the same facility, beginning on the same day (see Table 1). Female and male athletes followed identical session schedules and were exposed to the same environmental, nutritional, and recovery conditions. The maximal exertion test was performed on day 8 under identical circumstances for both groups. In national-team practice, such uniform programming is standard, as rowing workloads are prescribed using relative intensities (e.g., %VO2max, %HRmax, %power at threshold), while managing it is individualized. This approach ensured standardized external training exposure across all participants, although internal physiological load was not directly monitored during the training camp.

Male and Female1234567Total training time, min/day11090200105200105100Time rowed, min/day65–110951209540Distance rowed, km/day12–2020221810Training for force development, min/day––80–––60Extensive endurance rowing training time, min/day50–110351009540High intensity endurance rowing training time, min/day15––6020––Unspecific training (running, etc.), min/day459010106010–

Training program for the 7 days before the rowing test.

Inclusion criteria were: at least five years of rowing training experience, active membership in the national youth team, completion of a standardized 2000-meter ergometer test, and completion of the Profile of Mood States (POMS) questionnaire. Exclusion criteria included the presence of acute or chronic inflammation, pain or injury, use of anti-inflammatory medications, or non-compliance with the study protocol.

All female participants were non-pregnant, non-lactating, and reported regular menstrual cycles (28 ± 4 days). The menstrual cycle phase was not controlled because all measurements were constrained by the fixed schedule of the national-team training camp, which determined the timing of testing and did not allow for individualized adjustments of assessment days. However, this design enabled testing under highly homogeneous training, nutritional, environmental, and recovery conditions during the same preparation period.

Anthropometric measures included body mass, body composition (% fat, fat-free mass) assessed using a bioelectrical impedance analyzer (Tanita BC-418; Tanita Corporation, Tokyo, Japan), and body height measured with a stadiometer (SECA 213; seca GmbH & Co. KG, Hamburg, Germany). Measurements were taken in the morning, after an overnight fast. Performance variables from the 2000-m test (time, mean power, relative power) and blood lactate concentrations before and immediately after the exercise were also recorded. As expected, anthropometric and performance characteristics differed significantly between females and males due to established physiological and morphological sex differences. Descriptive characteristics of study participants are presented in Table 2.

VariableFemales (N = 12)Male (N = 16)Age (years)18.75 ± 0.7519.43 ± 1.09Body mass (kg)73.63 ± 3.5288.74 ± 5.12Height (m)176.5 ± 4.23188.5 ± 5.79%Fat18.24 ± 2.3311.53 ± 2.43Fat-free Mass (kg)60.05 ± 4.0578.46 ± 4.41Time of exercise (s)439.83 ± 7.52380.19 ± 9.29Power (W)263.83 ± 13.53408.87 ± 29.45Relative power (W·kg-¹)3.59 ± 0.304.61 ± 0.33LA pre-exercise (mmol·L-¹)1.14 ± 0.542.70 ± 0.93LA post-exercise (mmol·L-¹)12.43 ± 2.6010.95 ± 1.95

Participant characteristics and 2000-m rowing test results.

Values are mean ± SD. LA: blood lactate.

Since circulating tryptophan levels were analyzed, dietary composition and amino acid intake during the 24 h preceding the exercise test were also recorded. These data are presented in Table 3. Dietary intake during the 24 h preceding the exercise test was recorded and analyzed by a certified sports nutritionist using a validated dietary assessment software DietetykPro (DietetykPro Wrocław, Poland), which enabled precise quantification of total energy intake, macronutrients, and amino acid composition, including tryptophan and branched-chain amino acids.

VariableFemales (24 h)Males (24 h)p-valueEnergy kcal·kg-¹40.56 ± 10.2340.30 ± 8.170.8899Protein g·kg-¹1.76 ± 0.512.09 ± 0.570.1843Carbohydrate g·kg-¹6.46 ± 1.545.73 ± 1.390.2396Fat g·kg-¹1.06 ± 0.351.06 ± 0.370.7091Tryptophan mg·kg-¹20.76 ± 5.5924.72 ± 6.350.1090Leucine mg·kg-¹112.90 ± 24.36139.12 ± 39.030.0718Isoleucine mg·kg-¹70.51 ± 16.8890.07 ± 21. 100.0259Valine mg·kg-¹90.44 ± 23.1593.75 ± 29.860.1154Tyrosine mg·kg-¹59.33 ± 12.3864.08 ± 15.730.1763Phenylalanine mg·kg-¹65.39 ± 12.8278.28 ± 20.560.0929

Dietary composition and amino acid intake 24 h before exercise test.

Values are mean ± SD. Bolded values indicate statistically significant differences between groups (p < 0.05).

Study design

The study was conducted during the preparatory phase of the annual training cycle. Athletes first completed anthropometric assessments and the POMS questionnaire, followed immediately by the first blood collection.

Venous blood samples were obtained from the cubital vein at three time points: PRE (collected in the morning between 07:00 and 09:00 to limit circadian influence), POST (within 2 minutes after the 2000-m test), and 3H REC (3 hours post-exercise). Samples were collected into serum tubes with clot activator (Sarstedt AG & Co., Germany), allowed to clot for ~30 min, centrifuged at 3000 rpm for 10 min at 4 °C, aliquoted, and stored at -80 °C until analysis. Following the collection of the PRE blood sample, participants commenced the maximal-effort 2000 m rowing test, which was performed after their usual morning meal, consumed at the training-center dining facility.

The exercise test consisted of a 2000 m maximal-effort trial performed on a Concept2 rowing ergometer (Concept2 Inc., USA). This protocol is widely regarded as the standard laboratory test for assessing rowing performance in elite athletes (Ingham et al., 2002; Cerasola et al., 2020). Before the test, each participant completed a 5-minute individualized warm-up. During the test, participants were verbally encouraged to achieve maximal performance. Moreover, the test outcome contributed to the selection of athletes for championship-level crews; all participants were highly motivated to perform at their maximal capacity. The total time to complete the 2000 m distance and the mean power output were recorded as performance indicators (Table 2). All testing procedures were conducted by qualified personnel in compliance with biosafety standards and under identical laboratory conditions to minimize pre-analytical variability.

Serum concentrations of testosterone and cortisol were measured using enzyme-linked immunosorbent assay (ELISA) kits supplied by DiaMetra (Spello, Italy), according to the manufacturer’s instructions. The analytical sensitivity of the assays was 0.10 ng·mL-¹ for testosterone and 2.42 ng·mL-¹ for cortisol. According to the manufacturer’s specifications, the cortisol assay showed intra-assay and inter-assay coefficients of variation of <8.1% and <6.7%, respectively, while the testosterone assay showed corresponding values of <4.2% and <6.7%.

Serum levels of serotonin, dopamine, GABA, and tryptophan were determined using ELISA kits from SunRed Biotechnology Company (Shanghai, China), following the manufacturer’s protocols. The analytical sensitivity of the assays was as follows: serotonin - 0.388 ng·mL-¹, dopamine - 7.043 nmol·L-¹, GABA - 1.826 µg·dL-¹, and tryptophan - 0.233 ng·mL-¹. According to the manufacturer, the ELISA kits for serotonin, dopamine, GABA, and tryptophan showed intra-assay and inter-assay coefficients of variation below 10% and 12%, respectively. All analyses were performed using a SPECTROstar Nano microplate reader (BMG Labtech, Germany).

The testosterone-to-cortisol (T/C) and serotonin-to-dopamine (S/D) ratios were calculated to assess anabolic-catabolic and serotonin-dopamine balance.

Lactate concentration (LA) was assessed in capillary blood samples collected immediately after exercise using the Lactat Photometer and a commercially available diagnostic kit (Diaglobal GmbH, Berlin, Germany), according to the manufacturer’s instructions.

Psychological assessment

Mood states were assessed using the Profile of Mood States questionnaire in Polish adaptation by Dudek and Koniarek (1987). The POMS is a standardized and widely used instrument designed to evaluate transient mood fluctuations associated with fatigue and recovery processes. It consists of 65 adjectives describing current emotional states, which are rated by participants on a 5-point Likert scale ranging from 0 (“not at all”) to 4 (“extremely”).

The questionnaire yields scores in six subscales: Tension-Anxiety, Depression-Dejection, Anger-Hostility, Vigor-Activity, Fatigue-Inertia, and Confusion-Bewilderment. A Total Mood Disturbance index was also calculated by summing the negative mood dimensions (Tension, Depression, Anger, Fatigue, Confusion) and subtracting the score for Vigor.

Participants completed the POMS questionnaire in the morning, before pre-exercise blood sample collection, following standardized instructions and under the supervision of the research team.

Statistical analysis

Statistical analyses were performed using Statistica 14.0 (TIBCO Software Inc., Palo Alto, CA, USA) and GraphPad Prism 10.0 (GraphPad Software Inc., San Diego, CA, USA). In accordance with methodological recommendations by Lakens (2022), a sensitivity analysis was performed instead of an a priori power analysis because no consistent effect-size estimates are available for acute hormonal and neurotransmitter responses in elite athletes. Sensitivity analysis (G*Power 3.1) indicated that the achieved sample size (females = 12; males = 16) allowed detection of moderate-to-large effects (dz ≈ 0.7-0.8).

The normality of data distribution was assessed using the Shapiro-Wilk test, and the homogeneity of variances was verified with Levene’s test. For all normally distributed variables, a two-way ANOVA (Time × Group) was applied to examine the effects of Time (within-subject factor: PRE, POST, and 3H REC) and Group (between-subject factor: females, males), as well as their interaction (Time × Group). When the assumption of sphericity was violated, the Greenhouse-Geisser correction was applied.

When significant main effects or interactions were observed, Bonferroni-corrected post hoc tests were conducted to identify pairwise differences between specific time points and between sexes. In addition, independent samples t-tests (Welch correction) were used for variables measured at a single time point to compare females and males, including dietary variables, and POMS scores. For non-normally distributed data, the Friedman test (within-group) and the Mann-Whitney U test (between-group) were employed.

Exploratory Pearson’s correlations were computed separately for each time point (PRE, POST, and 3H REC) to examine concurrent relationships between mood states and biochemical markers. Given the sample size and the number of explored associations, the correlation analyses were interpreted as exploratory and used to identify biologically plausible psychophysiological relationships that may guide future research. The strength of correlations was interpreted as small (0.10-0.29), moderate (0.30-0.49), or large (≥0.50).

Effect sizes were reported alongside p-values to evaluate the magnitude of observed effects. For ANOVA models, partial eta squared (η²p) was calculated and interpreted as small (≥ 0.01), medium (≥ 0.06), or large (≥ 0.14), following conventional criteria (Cohen, 1988; Richardson, 2011). For paired and independent comparisons, effect sizes were calculated as Cohen’s d and interpreted as trivial (<0.20), small (0.20-0.49), moderate (0.50-0.79), or large (≥0.80). The strength of Pearson’s correlations was interpreted as small (0.10-0.29), moderate (0.30-0.49), or large (≥0.50), consistent with established conventions for behavioral and sport science research (Cohen, 1988; Hemphill, 2003). Data are expressed as mean ± standard deviation (SD). The level of statistical significance was set at p < 0.05.

All participants completed the full testing protocol, and no missing data occurred at any measurement point.

ResultsMood states (POMS)

Mood profiles assessed using the Profile of Mood States are presented in Figure 1. Both male and female athletes showed a similar distribution of scores across subscales, with the highest values observed for Vigor and the lowest for Depression and Confusion. Female athletes demonstrated slightly higher scores across most negative mood dimensions, resulting in a higher Total Mood Disturbance than male athletes (6.84 vs. 6.19, respectively) however, this difference did not reach statistical significance (p = 0.068).

Profile of Mood States in female and male rowers. Values are mean ± SD; TMD - Total Mood Disturbance; * significant difference at p < 0.05.

A significant sex difference was observed only in the Anger subscale, with female athletes reporting higher scores p = 0.049). No significant sex differences were found for Tension, Depression, Vigor, Fatigue, or Confusion.

Hormonal responses to exercise

For cortisol, significant main effects of time (p < 0.0001, η²p = 0.5201) and group (p = 0.0002, η²p = 0.5022), as well as a significant time × group interaction (p = 0.0002, η²p = 0.3394), were observed. In males, cortisol increased significantly from PRE to POST and from POST to 3H REC (d = 1.14 - 3.04), resulting in a significant overall elevation from PRE to 3H REC. In females, cortisol exhibited non-significant time-dependent changes, with effect sizes indicating small-to-moderate increases (d = 0.59 - 0.83). Consequently, cortisol concentrations were significantly higher in males at POST and 3H REC (p < 0.001) (Figure 2).

Changes in cortisol, testosterone, and testosterone-to-cortisol ratio (T/C) in male and female participants at PRE, POST, and 3H REC. Data are presented as mean ± SD. PRE - before exercise; POST - immediately after exercise; 3H REC - 3 hours of recovery. ap < 0.05 vs. PRE; bp < 0.001 vs. PRE; cp < 0.001 vs. POST; ***p < 0.001 between groups.

For testosterone, only the main effect of group was significant (p < 0.0001, η²p = 0.9352), whereas neither time nor the interaction reached significance. Testosterone remained stable across all time points in both sexes (d = 0.04 - 0.37). Male athletes consistently showed significantly higher concentrations than female athletes (p < 0.001), reflecting expected physiological sex differences (Figure 2). This very large group effect reflects the expected physiological difference in circulating testosterone concentrations between female and male athletes.

For the T/C ratio, significant main effects of time (p = 0.0346, η²p = 0.1213) and group (p < 0.0001, η²p = 0.8269) were found, with no significant interaction. Both groups demonstrated non-significant decreases across time (d = 0.05 - 0.72). At all time points, males exhibited significantly higher T/C values than females (p < 0.001), consistent with physiological differences in anabolic-catabolic balance (Figure 2).

Neurotransmitter responses

For serotonin, no significant main effects of time, group, or time × group interaction were detected (p > 0.05). Although no significant overall ANOVA effects were observed for serotonin, exploratory post hoc comparisons indicated a significant PRE-to-POST increase in males (d = 0.79), followed by a return toward baseline at 3H REC (d = 0.47–0.58), whereas females showed only small, non-significant fluctuations across time (d = 0.11–0.24). Serotonin concentrations were significantly higher in female athletes at POST (p = 0.045), with no significant between-group differences at PRE or 3H REC (Figure 3).

Changes in serotonin, dopamine, and serotonin-to-dopamine ratio (S/D) in male and female athletes at PRE, POST, and 3H REC. Values are mean ± SD. PRE - before exercise; POST - immediately after exercise; 3H REC - 3 hours of recovery. ap < 0.05 vs. PRE; bp < 0.05 vs. POST; *p < 0.05, **p < 0.01 between groups.

For dopamine, significant main effects of time (p = 0.0038, η²p = 0.1924) and a significant time × group interaction (p = 0.0107, η²p = 0.1599) were observed, while the main effect of group was not significant. In males, dopamine increased significantly from PRE to POST and decreased significantly from POST to 3H REC (d = 1.07 - 1.38), resulting in a minimal PRE to 3H REC difference (d = 0.21). In females, dopamine remained stable across time, with small, non-significant effect sizes (d = 0.01-0.36). Between-group comparisons showed higher dopamine in females at PRE (p < 0.01) and higher concentrations in males at 3H REC (p < 0.05), with no significant difference at POST (Figure 3).

For the S/D ratio, significant main effects of time (p = 0.0074, η²p = 0.1187) and group (p = 0.0068, η²p = 0.2491) were identified, whereas the interaction was not significant. In males, the S/D ratio decreased significantly from PRE to POST and showed a non-significant increase at 3H REC (d = 0.49 - 0.84). In females, the S/D ratio displayed non-significant time-dependent fluctuations (d = 0.16 - 1.16). At 3H REC, S/D values were significantly higher in males (p < 0.05), with no significant between-group differences at PRE or POST (Figure 3).

GABA and tryptophan

For GABA, no significant main effects of time, group, or time × group interaction were observed (all p > 0.05). In males, GABA showed small, non-significant fluctuations across measurements, with effect sizes indicating minimal PRE-to-POST, PRE-to-3H REC, and POST-to-3H REC differences (d = 0.12 - 0.30). In females, GABA also remained stable over time, with very small effect sizes (d = 0.04 - 0.07) and no statistically significant within-group changes. No significant between-group differences were detected at PRE, POST, or 3H REC (Figure 4).

Changes in gamma-aminobutyric acid (GABA) and tryptophan concentrations in male and female athletes at PRE, POST, and 3H REC. Values are mean ± SD. PRE - before exercise; POST - immediately after exercise; 3H REC - 3 hours of recovery. *p < 0.05, **p < 0.01 between groups; bp < 0.05 vs. POST.

For tryptophan, significant main effects of time (p = 0.0437, η²p = 0.1133) and group (p < 0.0001, η²p = 0.9359) were identified, as well as a significant time × group interaction (p < 0.0097, η²p = 0.1632) were identified. In females, tryptophan decreased progressively across the three time points, with a significant reduction from PRE to 3H REC (p < 0.05) and effect sizes indicating moderate to large changes (d = 0.47 - 1.35). In males, tryptophan values decreased from PRE to POST and showed a slight increase at 3H REC, but none of these changes reached statistical significance, with small-to-moderate effect sizes (d = 0.15 - 0.57). Between-group comparisons demonstrated higher tryptophan concentrations in females at PRE (p < 0.01) and POST (p < 0.05), while no significant difference was noted at 3H REC (Figure 4). The very large group effect should be interpreted in the context of marked baseline differences in circulating tryptophan concentrations between female and male athletes.

Correlations between biochemical markers and mood

Exploratory correlation analyses were conducted to examine associations between pre-exercise mood dimensions and biochemical markers at each time point (Table 4).

TimepointBiomarkerPOMSFemales
(r [95% CI])P valueMales
(r [95% CI])P value3HRecoveryCortisolAnger-0.66 [-0.85, -0.21]0.019––3HRecoveryCortisolDepression-0.64 [-0.84, -0.18]0.026––3HRecoveryCortisolTension-0.65 [-0.83, -0.22]0.021