Methods

Experimental Approach

A randomized crossover design was used to investigate the acute effects of SSGt and SSGf on neurochemical responses, cognitive performance, and technical actions in male soccer players. In SSGt, a maximum of two ball touches per player was imposed, whereas in SSGf, players were allowed to play without any touch restriction. The outcome measures included serum BDNF levels, blood lactate concentration, technical actions (interceptions, lost balls, successful and unsuccessful passes, and successful and unsuccessful shots), and executive function. The study consisted of a familiarization session followed by two experimental SSG sessions, as shown in Figure 1. During the familiarization session, participants were introduced to the Stroop test, testing procedures, and experimental protocol to minimize potential learning effects. Given their competitive level and training background, all players were already accustomed to SSG formats involving touch restrictions. Each participant completed both formats using a randomized counterbalanced crossover design. A 7‐day washout period was implemented between sessions to ensure sufficient physiological recovery and minimize potential carryover effects.

FIGURE 1: Study design. BDNF, Brain‐derived neurotrophic factor; GK, Goalkeeper; RPE, Rating of perceived exertion; RSME, Rating scale of mental effort; SSGf, Small‐sided games free‐play; SSGt, Small‐sided games with touch restriction.

Participants

Thirty‐two male soccer players (age: 21.7 ± 1.7 years, body height: 178 ± 4 cm, body mass: 76.3 ± 3.6 kg; body mass index: 24 ± 1 kg·m−2; body fat percentage: 12 ± 5.95; VO2max: 50.7 ± 3.7 mL · kg−1 · min−1) voluntarily participated in this study during the 2024–2025 off‐season period (Table 1). As part of their off‐season program, they completed prescribed individual training sessions 5 days per week. All participants were contracted professional players recruited from the same soccer club. The sample comprised all available outfield players who were regularly involved in the club's off‐season training program and met the inclusion criteria. Recruiting players from a single training squad was intended to reduce between‐team variability in training background, tactical familiarity, coaching exposure, and habitual SSG experience. The players had at least 10 years of soccer training experience and were recruited from the same team competing in the Turkish Professional Third Division. Participants were included if they met the following criteria: medical clearance for participation, no recent use of medications or supplements known to enhance aerobic or anaerobic performance, a Mini‐Mental State Examination (MMSE) score ≥ 26, and a Beck Depression Inventory (BDI) score of ≤ 9. The exclusion criteria comprised smoking, alcohol or drug use, use of psychoactive medication, a history of depression or neurological disease, any lower extremity injury within the previous 6 months resulting in ≥ 28 days of time loss or ongoing functional limitations, and the presence of neuromotor or musculoskeletal disorders.

Variables | Mean and SD
Age (year) | 21.7 ± 1.7
Height (cm) | 178 ± 4.0
Body mass (kg) | 76.3 ± 3.6
BMI (kg/m2) | 24.0 ± 1.0
VO2max (mL·kg−1 · min−1) | 50.7 ± 3.7
Body fat percentage (%) | 12.0 ± 6.0

The sample size was estimated using G*Power (version 3.1.9.7; University of Düsseldorf, Düsseldorf, Germany) for a repeated‐measures ANOVA (within–between interaction). Based on previously reported BDNF responses to acute exercise (effect size f = 0.30) reported by Devenney et al. (2019). The analysis was conducted with α = 0.05, power = 0.80, 2 conditions, and 2 measurements (ε = 1). The analysis indicated a required total sample size of 24 participants (Fcrit = 4.30, df1 = 1, and df2 = 22). However, to account for potential dropouts and increase statistical robustness, the study was conducted with 32 participants in total.

Before signing the informed consent form, the participants were informed of the research procedures, benefits, and risks. Ethical approval was obtained from the Tokat Gaziosmanpaşa University Research Ethics Committee (24.12.2024/21‐03), and the trial was retrospectively registered at ClinicalTrials. gov (NCT07395765), in accordance with the Declaration of Helsinki.

Participants were provided verbal information on the potential risks and benefits of the study during the initial briefing. Written informed consent was obtained before participation, and participants' willingness to continue was confirmed prior to each experimental session. The study design and reporting were conducted in accordance with the CONSORT 2025 statement (Hopewell et al. 2025) and relevant CONSORT extensions for randomized crossover trials (Dwan et al. 2019) and non‐pharmacological interventions (Boutron et al. 2017) (Figure 2). Participants were randomly assigned (1:1) to one of two counterbalanced condition orders (SSGt→SSGf or SSGf→SSGt) using a computer‐generated random sequence (www.randomizer.org). The allocated condition order was concealed until the completion of the baseline procedures. Given the within‐subject and non‐pharmacological nature of the SSG formats, blinding the participants and on‐field staff was not feasible. The participants were not informed of the specific study hypotheses. Laboratory analyses (serum BDNF and blood lactate) were performed using de‐identified data. Blinding of technical‐action coders was not feasible due to the obvious rule differences between formats; however, technical actions were coded offline from video recordings using predefined operational definitions and standardized coding criteria.

FIGURE 2: The consolidated standards of reporting trials (CONSORT) flowchart. SSG, Small‐sided games.

Procedures

During the first visit, body mass and body fat percentage were measured using bioelectrical impedance analysis (BC‐418, Tanita, Tokyo, Japan). Following the anthropometric assessments, resting heart rate (HRrest) was recorded for each participant. Subsequently, to mitigate potential learning effects, all participants underwent a familiarization session with Stroop Test and SSG formats. Finally, the participants completed a reliable assessment of team‐sport‐specific cardiorespiratory fitness using the 30–15 Intermittent Fitness Test (30–15 IFT), in accordance with the standardized procedure described by Buchheit (2008). The running velocity of the last fully completed stage was recorded as the maximum running speed (VIFT). Maximum oxygen consumption (VO2max) was then estimated using VIFT, age (A), gender (G, 2 for male), and body mass (BM), according to the equation proposed by Buchheit (2008):

VO2maxmL·kg−1·min−1=28.3−(2.15×G)−(0.71×A)−(0.0357×BM)+(0.0586×A×VIFT)+(1.03×VIFT)

During the two subsequent visits, each participant completed both SSGt and SSGf sessions, with one session (either SSGt or SSGf) completed during visit two and the other during visit three. The sequence of sessions was randomized and counterbalanced so that half of the participants performed the touch‐restricted condition first, whereas the remaining participants began with the free‐play condition. Each session started with a standardized 15‐min warm‐up, including light running, dynamic stretching, ball‐based technical drills, and acceleration, deceleration, and agility exercises (Jäger et al. 2025). This was followed by four 4‐min bouts of SSG interspersed with 3‐min passive recovery periods. The total duration of each session was 40 minutes, comprising a 15‐min warm‐up, 16 minutes of active play, and 9 minutes of passive recovery. Venous blood samples were obtained, and cognitive performance (Stroop task and mental effort) was assessed immediately before and after each session. Heart rate was continuously monitored, and all sessions were video recorded for subsequent technical analysis. In addition, psychophysiological responses were assessed after each session using ratings of perceived exertion (RPE) on the Borg 6–20 scale (Arney et al. 2019) and affective valence on the Feeling Scale (FS) (Figure 1).

All SSG sessions were performed between 10:00 and 12:00 to control for potential circadian influences and were conducted on a natural grass surface during the off‐season period. Sessions were supervised by certified strength and conditioning coaches holding UEFA licenses to ensure protocol consistency. Throughout the study, players continued their usual off‐season training routine. Participants were instructed to refrain from strenuous physical activity for 48 hours prior to testing, to avoid alcohol and caffeine for at least 24 hours, and to obtain a minimum of 7–8 hours of sleep before each session. They were also asked to maintain their habitual diet and daily lifestyle throughout the study period. Environmental conditions were consistent across sessions, with temperatures ranging from 18°C to 21°C and relative humidity between 37% and 40%.

Small‐Sided Games

Both experimental conditions used a 4v4 SSG format to provide a standardized game‐based training task across conditions (Ueda et al. 2023). All SSG were performed on a 30 × 24 m pitch (720 m2), corresponding to a relative area of 90 m2 per player. Following a standardized 15‐min warm‐up, participants completed four 4‐min bouts separated by 3‐min passive recovery intervals (Figure 1). To maintain continuous play and minimize interruptions, multiple FIFA‐approved size 5 balls, with a standard mass of 410–450 g, were positioned around the pitch, and two coaches ensured rapid ball replacement when necessary. The games were played without goalkeepers, using small goals measuring 1.2 × 0.8 m and positioned centrally on each end line. Players were instructed to perform at maximal effort throughout the sessions and received standardized verbal encouragement without technical or tactical feedback (Birinci, Pancar, and Soylu 2025). To promote competitiveness, participants were encouraged to score as many goals as possible in the opposing small goal. No additional scoring rules, such as a minimum number of passes or zone‐entry requirements, were applied. Players' primary playing positions were recorded and considered during team allocation. To minimize potential imbalance between teams, SSG teams were formed by the coaching staff based on players' usual playing positions and the coach's assessment of their technical, tactical, and physical abilities, in line with previous SSG research using coach‐based criteria to constitute balanced teams (Torrents et al. 2016).

The free‐play condition (SSGf) was performed without task constraints. In contrast, the touch restriction condition (SSGt) required players to release the ball within a maximum of two touches (e.g., pass or shot). Any violation of this rule (> 2 touches) resulted in immediate turnover of possession to the opposing team to maintain game flow. Standard football rules were applied, except that goalkeepers, the offside rule, and corner kicks were not used. When the ball left the playing area, coaches restarted play immediately using replacement balls positioned around the pitch. After goals, play was restarted from the scoring team's own goal line without allowing time for reorganization. Two assistant coaches acted as timekeepers and referees, ensuring adherence to the respective conditions during each bout.

Cognitive Performance

Cognitive performance was evaluated using the Stroop Color and Word Test—TBAG Form, standardized for the Turkish adult population (Karakaş et al. 1999). The test assesses key executive functions, particularly inhibitory control under conditions of cognitive interference, alongside selective attention, processing speed, and cognitive flexibility. The protocol comprised five standardized conditions presented on a computer screen, requiring verbal responses: (1) reading color words in black ink, (2) reading color words in colored ink, (3) naming the color of patches, (4) naming the ink color of neutral stimuli (e.g., “XXXX”), and (5) naming the ink color of incongruent words. Responses were recorded in real time by the examiner using the TBAG scoring sheet, including errors and self‐corrections. Participants were instructed to respond as quickly and accurately as possible and to correct errors immediately. Completion time (s), including correction time, was recorded for each condition, along with error and self‐correction counts. All five Stroop conditions were analyzed. In line with the TBAG Form scoring rationale, completion time in Condition 5 was treated as the main color‐word interference outcome because this section represents the critical interference condition of the test, whereas the other conditions serve as control conditions for reading speed and color naming (Karakaş et al. 1999). Testing was standardized in a quiet well‐lit environment, and a brief familiarization trial was provided before each session. Mental effort was quantified using the Rating Scale of Mental Effort (RSME), a 0–150 mm vertical visual analogue scale frequently used in sport psychology research (Fuster et al. 2021). Participants rated their perceived mental effort immediately after each SSG.

Technical Performance

Technical actions were recorded in 4K resolution using a drone‐mounted camera (DJI Mini 5 Pro, Shenzhen, China) operated by an experienced match analyst. The drone was positioned above the pitch to provide a stable overhead view that captured the entire playing area and all players throughout each bout. Technical performance was analyzed using E‐Analyze Digital Soccer Match Analysis Software (Espor Digital, Ankara, Türkiye). Technical actions were coded according to previously published procedures for technical performance analysis in small‐sided games (Owen et al. 2004; F. M. Clemente et al. 2019). Specifically, the following variables were recorded: successful passes, unsuccessful passes, interceptions, successful shots, unsuccessful shots, and lost balls. A successful pass was defined as a pass that reached and was controlled by a teammate without opponent intervention, whereas an unsuccessful pass failed to reach the intended teammate or resulted in loss of possession. An interception was recorded when a defender deliberately interrupted an opponent's pass. A successful shot was defined as an attempt resulting in a goal, whereas an unsuccessful shot included attempts that missed the target or were blocked by an opponent. A lost ball was recorded when possession was lost because of poor ball control, dispossession, or a technical error. During the 3‐min passive recovery periods between bouts, players were allowed to consume water ad libitum. To determine intra‐rater reliability, a randomly selected subset of the SSG bouts was re‐analyzed by the same analyst after a 14‐day interval. Reliability was evaluated using the intraclass correlation coefficient (ICC; two‐way random‐effects model, absolute agreement, single measures), with ICC values ranging from 0.744 to 0.987, indicating good‐to‐excellent reliability.

Internal/Perceptual Load

To objectively quantify internal load, heart rate (HR) was recorded continuously at 5‐s intervals throughout each session using Polar V800 units (Polar Electro Oy, Kempele, Finland) paired with Polar H10 (Polar Electro Oy, Kempele, Finland) chest straps. Mean HR values were subsequently computed for each repetition period during data processing. In parallel, blood lactate concentration and RPE were also assessed. Affective state was evaluated using the FS (Hardy and Rejeski 1989), where participants reported their immediate emotional status on a continuum ranging from −5 (very bad) to +5 (very good).

Biochemical Analysis

Venous blood (4 mL) was collected from the antecubital vein with participants seated, immediately prior to exercise and at each designated post‐exercise time point. Following collection, samples were transported to the laboratory under controlled conditions and processed on the same day. Samples were allowed to clot for 30 min at room temperature, then centrifuged at 3000 × g for 12 min. Serum was aliquoted and stored at −80°C until analysis. Biochemical analyses (BDNF and lactate) were subsequently conducted between 16:00 and 17:30 h using standard laboratory procedures. Serum BDNF was quantified using a commercially available sandwich ELISA kit (Cloud Clone, USCNK, Wuhan, China) in accordance with the manufacturer's instructions. Blood lactate was determined using a Super GL2 analyzer (Müller Gerätebau GmbH, Freital, Germany) based on an enzymatic–amperometric electrochemical method. The analyzer was calibrated before each measurement session according to the manufacturer's guidelines. All specimens were de‐identified during analysis, and sample codes were disclosed only after completion of the measurements. Biochemical analyses were performed by an experienced laboratory technician who was blinded to the experimental conditions.

Statistical Analysis

IBM SPSS Statistics (version 29.0, SPSS Inc., USA) was used for all analyses. Normality was evaluated using skewness and kurtosis values, with values between −1 and +1 considered acceptable for normal distribution. A two‐way repeated‐measures ANOVA (time × condition) was conducted to analyze BDNF, Stroop test performance, lactate, RPE, and FS scores. A one‐way repeated‐measures ANOVA (condition) was used to examine technical actions, HRmean, %HR, and mental effort. Bonferroni‐corrected tests were applied for post‐hoc comparisons. Partial eta squared (ηp2) values were reported as effect sizes (small: 0.01 ≤ ηp2 < 0.06; medium: 0.06 ≤ ηp2 < 0.14; large: ηp2 ≥ 0.14). Statistical significance was set at p < 0.05. Graphical visualization was performed using GraphPad Prism software.