Time-dependent comparison of serum BDNF responses following high-intensity interval exercise and moderate- and low-intensity continuous exercise in healthy young men

Abstract

Exercise serves as a potent physiological stimulus influencing brain-derived neurotrophic factor (BDNF), a key molecule involved in neuronal maintenance, synaptic plasticity, and cognitive regulation. However, the temporal dynamics of circulating BDNF in response to varying exercise intensities remain poorly understood. This study investigated serum BDNF kinetics following low- and moderate-intensity continuous exercise (LICE, MICE) and high-intensity interval exercise (HIIE) in young healthy men. Twelve participants completed all three exercise sessions in a randomized crossover design with seven-day washouts. Venous blood samples were collected at baseline, immediately post-exercise, and at 5, 15, 30, 45, and 60 min of recovery. Serum BDNF and blood lactate concentrations were measured, and heart rate was monitored continuously during each exercise session. BDNF levels were significantly higher in HIIE than in both LICE and MICE across all time points (p < 0.001), with no differences between LICE and MICE (p > 0.05). Within-group analysis revealed that HIIE induced a pronounced but transient rise in BDNF, peaking at 15 min post-exercise and returning to baseline by 60 min. These results suggest that HIIE elicits a distinct neurotrophic response pattern shaped by exercise intensity and interval structure. Our findings provide descriptive data on serum BDNF kinetics that may inform future mechanistic research. Trial registration: The study was registered on ClinicalTrials.gov (identifier: NCT07175831 https://clinicaltrials.gov/study/NCT07175831) on 15th of September 2025.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-026-37728-z.

Introduction

Brain-derived neurotrophic factor (BDNF) is a neurotrophin expressed both centrally in the brain and peripherally in tissues such as skeletal muscle. It plays a pivotal role in neurodevelopment, neuroprotection, and the regulation of synaptic plasticity, particularly within the hippocampus and prefrontal cortex¹. Experimental evidence demonstrates that BDNF enhances hippocampal neurogenesis, facilitating learning and memory² strengthens synaptic connectivity in the prefrontal cortex to support executive functions³, and contributes to mood regulation through amygdala-mediated pathways⁴. Beyond its neurotrophic functions, BDNF has been implicated in fundamental metabolic pathways, including glucose homeostasis and lipid oxidation⁵. Together, these roles position BDNF at the interface of cognitive function and energy metabolism.

Physical exercise represents a cost-effective and health-promoting intervention that can upregulate BDNF signaling and increase circulating BDNF (serum/plasma), with potential implications for synaptic plasticity and cognitive function⁶. Studies have reported a correlation between plasma BDNF levels and brain BDNF levels, and that changes in brain BDNF will also be reflected in plasma BDNF⁷. Regular physical activity has been associated with a reduced risk of neurodegenerative disorders, including dementia, multiple sclerosis, and Parkinson’s disease⁸. The upregulation of neurotrophic factors, particularly BDNF, is considered a key mechanism underlying exercise-induced neuroplasticity⁹. Notably, exercise intensity appears to modulate BDNF responses. Accordingly, high-intensity interval exercise (HIIE) has been shown to evoke greater BDNF responses than moderate- or low-intensity continuous exercise (MICE and LICE)¹⁰. Beyond intensity per se, a recent systematic review indicates that the superior BDNF response to high-intensity exercise may, at least in part, be mediated by the concurrent rise in lactate concentrations¹¹. Lactate is increasingly recognized not only as an energetic substrate but also as a signaling molecule capable of crossing the blood–brain barrier¹² and stimulating BDNF transcription¹³. Accumulating evidence from both human and animal studies suggests that higher lactate concentrations correlate with elevated plasma or serum BDNF levels^14,15. However, the time-dependent interplay between exercise intensity, lactate dynamics, and circulating BDNF remains incompletely characterized.

In line with this mechanistic perspective, a recent systematic review¹⁶ highlighted that HIIE has the potential to increase BDNF levels, but also emphasized that this effect is strongly influenced by exercise intensity, duration, and individual variability. Discrepancies across studies may further arise from methodological and physiological factors such as the use of serum versus plasma assays¹⁷, clotting time and centrifugation strategy¹⁸, variations in participants’ fitness levels and habitual activity¹⁹, and differences in nutritional or metabolic status^20,21. These factors can substantially influence circulating BDNF concentrations.

Sampling timing represents an additional major source of variability in exercise-induced BDNF responses. Accordingly, existing studies have reported inconsistent temporal patterns, with peak responses observed during exercise²², immediately post-exercise^10,23,24, or several minutes into recovery. Moreover, return-to-baseline kinetics have varied, ranging from rapid normalization within 20–30 min^10,23 to gradual declines extending up to one hour post-exercise²². These inconsistencies highlight the need for repeated blood sampling across recovery to more precisely delineate the temporal kinetics of exercise-induced BDNF secretion, particularly because BDNF secretion and clearance may occur rapidly and may not be adequately captured by single post-exercise measurements. Consequently, sampling time points can substantially influence the interpretation of exercise-induced neurotrophic release. Unlike previous studies that often relied on a single post-exercise measurement, our approach included multiple recovery time points and compared distinct exercise modalities, enabling a more detailed characterization of BDNF kinetics.

The purpose of this study was to investigate the temporal dynamics of serum BDNF by comparing LICE and MICE with HIIE in healthy young adult males. We hypothesized that MICE and HIIE would elicit greater serum BDNF increases compared to LICE. Additionally, due to the differing structures and intensities of the exercise protocols, we expected the peak timing and recovery course of BDNF to differ between conditions. However, given the limited evidence on the temporal dynamics of BDNF secretion, this expectation was considered exploratory.

Results

The demographic and anthropometric characteristics of the participants are presented in Table 1.

Table 1.

Features of the participants.

Variables	Mean and SD
Age (year)	22.30 ± 1.36
Height (cm)	177.10 ± 4.58
Weight (kg)	74.30 ± 6.15
BMI (kg/m2)	23.06 ± 1.42
VO2max (mL·kg⁻¹·min⁻¹)	46.78 ± 3.56
MAS (km/h)	15.08 ± 0.52

SD: Standard deviation; BMI: Body mass index; VO₂max: Maximum oxygen uptake was estimated from the Yo-Yo Intermittent Recovery Test Level 1 using a published prediction equation; MAS: Maximum aerobic speed.

In the following section of the present study, the statistical findings for heart rate, serum BDNF and blood lactate responses of the LICE, MICE and HIIE conditions are presented. Table 2 presents the values of serum BDNF and blood lactate responses across the experimental conditions.

Table 2.

Time-Dependent changes in serum BDNF and blood lactate following LICE, MICE, and HIIE.

Group	BDNF (ng/mL) mean and SD							F value
	Pre- Exercise	Post- Exercise	5 min.	15 min.	30 min.	45 min.	60 min.	Main effect (time)		Interaction effect (group x time)
								p		p
LICE	1.57 ± 0.60	1.69 ± 0.78	1.81 ± 0.83	1.76 ± 0.48	1.42 ± 0.67	1.35 ± 0.87	1.20 ± 0.83	< 0.001	0.594	< 0.001	0.504
MICE	1.68 ± 0.55	2.29 ± 1.00	2.18 ± 0.81	2.07 ± 0.85	2.05 ± 0.46	1.95 ± 0.69	1.98 ± 0.82
HIIE	2.44 ± 1.56	7.94 ± 1.64	8.73 ± 1.81	9.85 ± 2.51	7.30 ± 3.61	6.51 ± 2.69	6.35 ± 2.50
	Lactate (mmol/L) mean and SD							F value
Group	Pre- Exercise	Post- Exercise	5 min.	15 min.	30 min.	45 min.	60 min.	Main effect (time)		Interaction effect (group x time)
LICE	2.11 ± 0.20	2.85 ± 0.47	2.27 ± 0.34	2.60 ± 0.53	2.43 ± 0.25	2.30 ± 0.26	1.61 ± 0.32	p		p
MICE	1.97 ± 0.14	4.88 ± 0.53	4.00 ± 0.49	3.40 ± 0.70	3.24 ± 0.55	3.30 ± 0.35	3.01 ± 0.39	< 0.001	0.924	< 0.001	0.878
HIIE	2.03 ± 0.23	15.74 ± 2.80	14.88 ± 3.79	12.64 ± 3.20	8.78 ± 2.88	5.84 ± 2.31	5.36 ± 1.86

Values are presented as mean ± standard deviation (SD). The significance level was determined as p < 0.05. partial eta squared,.

Serum BDNF kinetics in response to LICE, MICE, and HIIE conditions were given in the Fig. 1. No statistically significant differences in BDNF levels were observed between the groups at pre-exercise (p > 0.05). A significant Condition × Time interaction was observed for serum BDNF levels (F_(12,132) = 11.2, p < 0.001, = 0.504). Post-hoc analyses revealed that immediately post-exercise, HIIE elicited significantly higher BDNF concentrations compared to LICE (Mean Difference = 6.26 ng/mL, 95% CI [5.18, 7.33], p < 0.001) and MICE (Mean Difference = 5.65 ng/mL, 95% CI [4.36, 6.93], p < 0.001). This pattern of HIIE superiority was consistent throughout the recovery period (detailed pairwise statistics are provided in Table S-1).

Fig. 1.

Serum BDNF kinetics in response to LICE, MICE, and HIIE conditions. Data are presented as mean ± SD. The horizontal dashed line represents the mean baseline value of the HIIE condition. BDNF levels were measured at baseline (pre-exercise), immediately post-exercise, and at 5, 15, 30, 45, and 60 min during recovery. HIIE: High-intensity interval exercise; LICE: Low-intensity continuous exercise; MICE: Moderate-intensity continuous exercise; ng/mL: Nanograms per millilitre.

In the HIIE group, within-group analysis revealed large effect sizes for time-dependent changes. BDNF levels showed significant increases compared to baseline immediately post-exercise (p < 0.001, d = 3.43), 5 min post-exercise (p < 0.001, d = 3.72), 15 min post-exercise (p < 0.001, d = 3.63), 30 min post-exercise (p = 0.01, d = 1.88) and 45 min post-exercise (p = 0.04, d = 1.91).

The individual trajectories of serum BDNF for all participants under all conditions are presented in Fig. 2 to demonstrate inter-individual variability in response patterns.

Fig. 2.

Inter-individual trajectories of serum BDNF. HIIE: High-intensity interval exercise; LICE: Low-intensity continuous exercise; MICE: Moderate-intensity continuous exercise; ng/mL: Nanograms per millilitre.

Blood lactate kinetics in response to LICE, MICE, and HIIE conditions were given in the Fig. 3. No statistically significant differences in blood lactate levels were observed between the groups at pre-exercise (p > 0.05). A significant Condition × Time interaction was observed for blood lactate concentrations (F_(12,132) = 79.1, p < 0.001, = 0.878). Post-hoc analyses revealed that immediately post-exercise, HIIE elicited significantly higher lactate concentrations compared to LICE (Mean Difference = 12.89 mmol/L, 95% CI [11.12, 14.66],) and MICE (Mean Difference = 10.86 mmol/L, 95% CI [9.11, 12.61]). Furthermore, MICE also induced significantly higher lactate levels than LICE immediately post-exercise (Mean Difference = 2.03 mmol/L, 95% CI [1.68, 2.38]).

Fig. 3.

Blood Lactate kinetics in response to LICE, MICE, and HIIE conditions. Data are presented as mean ± SD. Blood lactate leves were measured at baseline (pre-exercise), immediately post-exercise, and at 5, 15, 30, 45, and 60 min during recovery. The horizontal dashed line represents the mean baseline value of the HIIE condition. Symbols indicate significant differences: (*) HIIE is significantly higher than both LICE and MICE (p < 0.05); (#) MICE is significantly higher than LICE (p < 0.05). HIIE: High-intensity interval exercise; LICE: Low-intensity continuous exercise; MICE: Moderate-intensity continuous exercise; mmol/L: Millimoles per litre.

To further investigate the temporal dynamics of lactate responses, within-group comparisons were conducted to assess how lactate levels varied across different time points within each of the three exercise protocols. In the HIIE group, lactate levels were significantly elevated compared with baseline at immediately post-exercise (p < 0.001), 5 min post-exercise (p < 0.001), 15 min post-exercise (p < 0.001), 30 min post-exercise (p < 0.001), 45 min post-exercise (p = 0.002), and 60 min post-exercise (p = 0.001). In the MICE group, lactate levels were significantly elevated relative to baseline at immediately post-exercise (p < 0.001), 5 min post-exercise (p < 0.001), 15 min post-exercise (p < 0.001), 30 min post-exercise (p < 0.001), 45 min post-exercise (p < 0.001), and 60 min post-exercise (p < 0.001). In the LICE group, lactate levels were significantly elevated compared with baseline only at immediately post-exercise (p = 0.003). No significant differences were observed relative to baseline at 5 min (p = 0.057), 15 min (p = 1.00), 30 min (p = 1.00), and 45 min post-exercise (p = 1.00). However, lactate concentrations were significantly decreased below baseline values at 60 min post-exercise (p = 0.010) (Table S-2).

The individual trajectories of blood lactate for all participants under all conditions are presented in Fig. 4 to demonstrate inter-individual variability in response patterns.

Fig. 4.

Inter-individual trajectories of blood lactate.

HRmean,% HRR, %HRmax mean and total running distance in response to LICE, MICE, and HIIE conditions are presented in the Fig. 5. Figure 5 (a) presents the values of HRmean during the experimental conditions. HRmean values were 148.92 ± 6.13 bpm in LICE, 174.83 ± 3.59 bpm in MICE, and 190.33 ± 3.37 bpm in HIIE. Figure 5 (b) presents the values of % HRR mean during the experimental conditions. % HRR mean values were 61.41 ± 5.37 in LICE, 80.65 ± 2.96 in MICE, and 90.61 ± 2.74 in HIIE. Figure 5 (c) presents the values of HRmax % mean during the experimental conditions. HRmax % mean values were 73.56 ± 3.35 in LICE, 86.36 ± 2.35 in MICE, and 94.00 ± 2.08 in HIIE. Figure 5d presents the running distance (m) covered during the experimental conditions (N = 12). Mean distances were 5531.5 ± 361.59 m in MICE, 3951 ± 258.29 m in LICE, and 2755.83 ± 176.71 m in HIIE.

Fig. 5.

(a) presents the mean heart rate (HRmean), (b) presents the percentage heart rate reserve (% HRR) and c) presents the percentage heart rate maximum (%HRmax) and (d) total running distance for the LICE, MICE, and HIIE conditions. LICE: low-intensity continuous exercise; MICE: moderate-intensity continuous exercise; HIIE: high-intensity interval exercise; bpm: beats per minute.

Discussion

In this study, we tested whether LICE, MICE, and HIIE differentially shape the temporal profile of serum BDNF secretion in healthy young men, using repeated blood sampling across recovery to capture peak and decline patterns. Overall, HIIE elicited a greater post-exercise increase in serum BDNF, whereas MICE did not differ from LICE; therefore, our primary hypothesis was only partially supported. Specifically, 30-min HIIE [4 sets of 10 × 15-s runs at 105–110% maximum aerobic speed (MAS) with 15-s passive rest, separated by 2.5-min recovery] elicited a marked increase in serum BDNF, beginning immediately post-exercise and peaking at 15 min. Thereafter, BDNF levels showed a gradual decline but remained significantly elevated above baseline and the other exercise conditions at 5, 15, 30, and 45 min, before returning to values no longer statistically different from baseline at 60 min. Neither LICE nor MICE induced significant changes. In line with our exploratory expectation, HIIE exhibited a distinct temporal profile compared with both continuous conditions. This outcome likely reflects fundamental differences in the physiological stimuli imposed by these protocols. Although MICE involved higher intensity than LICE, the continuous structure and limited metabolic stress may have been insufficient to markedly enhance neurotrophic signaling. In contrast, HIIE combined supramaximal intensity (105–110% MAS) with a fragmented interval structure, which likely induced greater metabolic perturbation and neuroendocrine activation, key drivers of activity-dependent BDNF release.

Our findings align with previous evidence demonstrating that HIIE elicits robust elevations in BDNF^25–27,with several studies indicating that these increases are more pronounced than those induced by lower-intensity and/or continuous exercise^{10,24,28–30}, supporting a dose-dependent relationship between exercise intensity, interval structure and BDNF secretion. However, others reported contradictory findings^23,31,32. The delayed BDNF peak and prolonged rise observed at 15 min post-exercise can be attributed to the supramaximal intensity and intermittent nature of the HIIE protocol. Unlike continuous or submaximal exercise, the 105–110% MAS intensity and 15 s work/15 s rest format likely caused significant metabolic and neuroendocrine stress, extending lactate and catecholamine activity until early recovery. This may have shifted the peak of the BDNF response from the immediate post-exercise period to 15 min later. Therefore, our findings demonstrate that BDNF release follows a delayed temporal pattern under supramaximal and highly intermittent exercise conditions, highlighting the protocol-dependent nature of BDNF kinetics.

A plausible mechanism linking HIIE to BDNF involves lactate, which has been redefined from a metabolic by-product to a key signaling molecule³³. Prior work suggests that the brain may shift toward using lactate rather than glucose as a primary energy substrate, highlighting its dual role as both a metabolic fuel³⁴ and neuromodulator that facilitates neuroplastic adaptations during and after intense exercise³⁵.

Consistent with the parallel increases in lactate and BDNF following HIIE observed in our study, and based on previous mechanistic evidence suggesting a link between lactate signaling as mediator of BDNF upregulation^15,33, we speculate that lactate may contribute to the upregulation of circulating BDNF after HIIE. However, despite moderate elevations in lactate after MICE, no concomitant increases in serum BDNF were observed. This pattern suggests that lactate elevation alone may be insufficient, and that a threshold and/or additional intensity-dependent stimuli may be required to elicit a robust neurotrophic response. One possible explanation, requiring future verification, is that lactate concentrations achieved during MICE did not reach a level sufficient to engage central lactate signaling pathways to the same extent as HIIE.

Most previous studies^36,37, including our earlier work^31,38, have evaluated the effects of exercise on BDNF responses based on single blood sample collected immediately after exercise. However, many authors have acknowledged this as a limitation, since BDNF secretion is hypothesized to be a dynamic process that evolves during exercise and in the early recovery phase^39,40, To address this gap, we implemented repeated blood sampling across recovery to characterize the temporal profile of post-exercise BDNF release and identify the time point at which peak responses occur. Surprisingly, in contrast to prior reports^41,42, we observed no increase in serum BDNF at any sampling time following MICE or LICE. This may be due to the fact that BDNF release during low or moderate-intensity or shorter continuous exercise is not necessarily greater than at rest, and may be strongly influenced by exercise duration, intensity thresholds, and large interindividual variability¹⁶. Moreover, regulatory mechanisms governing BDNF uptake and release in the central nervous system or peripheral tissues may limit the detectability of acute increases under these conditions^22,43.

By contrast, HIIE elicited a marked increase in serum BDNF, evident immediately post-exercise, peaking at 15 min, and showing a gradual decline thereafter. Although concentrations had returned toward baseline by 60 min, BDNF levels remained significantly higher after HIIE than after LICE and MICE. In contrast to our findings, many studies reported that circulating BDNF concentrations rise progressively with exercise duration but return to baseline within minutes of recovery^23,39,40.

Several methodological differences across studies may help explain discrepancies in the temporal profile of circulating BDNF. Saucedo Marquez et al. [2015]¹⁰ examined an interval cycling protocol consisting of 1-min work at 90% maximal workload alternated with 1-min at 60 W (1:1 work-to-rest) for 20 min, and reported that serum BDNF increased progressively and peaked immediately post-exercise, but returned rapidly such that no difference was detected by 20 min of recovery. Rasmussen et al. [2009]²² reported a two- to threefold peak BDNF increase during exercise and returned toward baseline within approximately one hour. However, their prolonged low-intensity rowing protocol (4 h of rowing at ~ 15% below the lactate threshold) and limited recovery assessment constrain direct comparison with the present design. Ross et al. [2019]²³ used continuous cycling and prescribed intensity using heart-rate reserve (HRR), showing that ~ 70% HRR for 15 min elicited a transient serum BDNF increase peaking immediately post-exercise and returning to baseline within 15–30 min, whereas ~ 35% HRR induced only modest changes. Although these findings support an intensity-dependent response, the temporal pattern differs from the present study, where the maximum response occurred 15 min into recovery rather than immediately post-exercise. Such divergence may reflect differences in exercise modality (continuous cycling vs. interval outdoor running) and relative intensity (e.g., ~ 70% HRR vs. ~90% HRR in our HIIE), which can distinctly shape metabolic stress and the kinetics of neurotrophic release. Moreover, Li et al. [2022]²⁴ applied cycling-based HIIE with 1-min bouts at 85% VO₂peak alternated with 1-min active recovery at 25% VO₂peak (1:1). In their 20-minute protocol (10 cycles), serum BDNF increased immediately post-exercise and returned to baseline by 30 min, with a positive association between changes in lactate and BDNF. However, in their 30-minute protocol (15 cycles), post-exercise BDNF did not increase and was reduced at 30 min, highlighting that both intensity prescription and total duration can influence the direction and time course of the response. Notably, although our HIIE protocol lasted 30 min, the total accumulated work (10 min; 15-s × 10 reps × 4 sets) was comparable to that of Li et al.’s [2022]²⁴ 20-min intervention (1-min × 10 reps). Nevertheless, the intermittent supramaximal nature of our HIIE (90%HRR; 105–110% MAS) likely imposed a greater metabolic and neuromuscular load than continuous conditions. Notably, the observed discrepancies may also relate to intensity thresholds, as Bergersen et al., 2015⁴⁴ suggested that achieving ≥ 80% HRmax may be critical for lactate-mediated BDNF stimulation; accordingly, protocols that remain below this threshold may elicit more transient or absent BDNF responses. Therefore, between-study differences in work-to-rest structure, intensity metric, modality, and sampling schedule should be considered when interpreting discrepant BDNF kinetics across the literature.

Taken together, the delayed peak and prolonged recovery of BDNF observed in our study may be explained by methodological and physiological factors. Previous studies^16,24,39 have shown that exercise-induced BDNF secretion is highly sensitive to variations in duration, intensity, and interindividual characteristics, which likely contribute to discrepant temporal profiles across studies. In line with this, the relatively large standard deviations observed in our data, particularly in the HIIE condition (e.g., 9.85 ± 2.50 ng/mL at 15 min), indicate substantial inter-individual variability in the BDNF response. Such variability suggests that while HIIE consistently induces a robust group-level increase, individual responses may be strongly modulated by factors such as genetic predisposition, acute physiological state, or baseline fitness. Antunes et al. (2020)⁴⁵ reported that participants with VO₂max values below 49.7 mL·kg⁻¹·min⁻¹ exhibited greater BDNF responses to short-duration high-intensity exercise compared with those above this threshold. In this context, the mean VO₂max of our participants (46.78 ± 3.56 mL·kg⁻¹·min⁻¹), which falls below this threshold, may have contributed to an increased sensitivity to the HIIE stimulus and the robust BDNF elevations observed. Importantly, the HIIE protocol applied here involved higher relative intensity (105–110% MAS) and a more fragmented interval structure than those reported in most prior investigations. This design may have amplified the neurotrophic stimulus, shifting the peak response into early recovery (15 min post-exercise) and prolonging the return to baseline until 60 min. Thus, while our findings reinforce the shared view that BDNF secretion is an intensity- and interval-dependent dynamic process, they also highlight that the timing and duration of the response are contingent upon specific protocol characteristics.

The present study has several limitations. A major limitation of this study is the lack of standardized pre-exercise dietary control. Although participants were instructed to abstain from food and caffeine at least two hours before testing, no formal assessment or standardization of macronutrient intake was implemented. Because circulating BDNF is highly sensitive to metabolic state, glycemic fluctuations, and recent nutrient intake, the absence of dietary control likely contributed to inter-individual variability in post-exercise BDNF responses²⁰. Second, exercise intensity was prescribed individually based on MAS values; however, this approach may not reflect an equal physiological load across participants and should therefore be regarded as a limitation. As highlighted by Bok et al. [2023]⁴⁶, prescribing short-interval HIIE using anaerobic speed reserve (ASR) rather than MAS can reduce interindividual variability in physiological and perceptual responses and may thus provide a more objective method for intensity normalization However, maximal sprint speed was not assessed in the present study; therefore, ASR could not be calculated and was not used for exercise prescription. Third, although the exercise protocols were matched for time, this study did not control for overall energy expenditure (total work); therefore, differences in accumulated work and physiological load across exercise modalities cannot be excluded and should be considered when interpreting the observed BDNF responses. Fourth, the study sample consisted exclusively of healthy young men. In females, circulating BDNF is known to fluctuate across the menstrual cycle, with evidence suggesting higher concentrations during the luteal phase compared with the follicular phase⁴⁷. These hormone-related fluctuations may influence neurotrophic and metabolic responses to exercise, which was a key reason for restricting this initial investigation to healthy young men. This sampling choice also limits the generalizability of our findings to females, as well as to children, older adults, athletes, and clinical populations. In addition, although an a priori power analysis confirmed that the sample size exceeded the minimum required to detect moderate within-subject effects, the relatively small number of participants and the well-documented inter-individual variability in BDNF responses may have limited statistical power to detect smaller effects and further constrained the generalizability of the results. Fifth, BDNF was assessed only in serum, which is strongly influenced by platelet-stored BDNF release. Platelet count and activation indices were not evaluated; therefore, changes in platelet number or activation could have affected measured serum BDNF concentrations. In addition, plasma BDNF—which better reflects the free, bioavailable fraction—was not measured. Accordingly, our findings should be interpreted as systemic exercise-induced alterations in circulating BDNF rather than a direct index of central neuronal BDNF secretion. Sixth, plasma volume changes were not estimated because hemoglobin and hematocrit were not measured. Given that HIIE may induce greater hemoconcentration than LICE or MICE, the observed elevations in serum BDNF following HIIE may be partially artifactual and reflect plasma volume shifts rather than true increases in circulating BDNF mass. Finally, while we speculated that the concurrent increases in blood lactate and serum BDNF after HIIE may reflect a mechanistic link, only peripheral lactate was measured. As such, a causal relationship with BDNF secretion cannot be confirmed. Moreover, despite standardizing the timing of all exercise sessions using a crossover design, the absence of a non-exercise control condition limits the ability to fully account for individual circadian influences on BDNF.

Future research should employ direct or indirect assessments of cerebral lactate dynamics to validate this hypothesis. To establish functional relevance, studies should incorporate concurrent cognitive or behavioral outcomes alongside circulating biomarkers. Given that serum BDNF is strongly influenced by platelet-derived release, future work should also include plasma BDNF measurements to better capture the bioavailable fraction. In addition, broader and more diverse samples are needed, including female participants with appropriate menstrual cycle phase control, to improve generalizability. Methodologically, systematic assessment/standardization of pre-exercise dietary intake and the use of more precise prescription approaches (e.g., ASR-based intensity setting) may reduce inter-individual variability. Finally, recovery sampling beyond 60 min is needed to fully characterize return-to-baseline kinetics. Longitudinal follow-up would further clarify whether these acute responses translate into meaningful neuroplastic and cognitive adaptations.

Our data indicate that HIIE uniquely induces robust elevations in serum BDNF, with a distinct temporal profile characterized by a peak at 15 min and a gradual decline to baseline by 60 min. The current data extend the understanding of exercise-induced BDNF kinetics by suggesting that sufficiently intense, intermittent exercise is particularly effective in eliciting robust and sustained elevations during recovery. HIIE may represent an effective strategy in healthy young men to transiently boost prolonged neurotrophic responses, though repeated measurements at multiple time points post-exercise are recommended to more accurately characterize the temporal profile of serum BDNF.

Materials and methods

Sample size estimation

Sample size estimation was conducted via G*Power software (version 3.1.9.7; University of Düsseldorf, Düsseldorf, Germany) via repeated-measures ANOVA (within-subjects design). The effect size was calculated on the basis of level of serum BDNF levels between groups, with a partial eta squared of 0.202 for the group x time interaction in serum BDNF, as reported by Nishikawa et al. [2024]⁴⁸. This corresponds to an effect size (f) of 0.50. The analysis was performed with an alpha level of 0.05 and a statistical power of 0.95 (1-β = 0.95), assuming 1 groups and 7 measurement points. The output parameters indicated a critical F value of 2.364 and a total required sample size of 7. However, the study was planned to be conducted with 12 participants.

Participants

Participants were recruited through announcements distributed on social media platforms and posters displayed in publicly accessible areas, such as university campuses and sports centers. Individuals who expressed interest were invited to the Sport Science laboratory of the faculty for eligibility screening prior to enrollment. Eligible participants were healthy males aged 18 years or older, with valid medical clearance for exercise and no history of serious lower-limb injury within the previous 12 months. Exclusion criteria included smoking, regular alcohol consumption, recreational drug use, and the use of psychoactive substances, medications, or supplements known to enhance aerobic or anaerobic performance within the last three months. Participants were also excluded if they reported a history of depression, neurological disorders, or neuromotor/musculoskeletal impairments, or if they failed to provide written informed consent.

Participant recruitment and randomization

All participants received verbal information regarding the potential risks and benefits of the study during the initial briefing as well as immediately prior to each familiarization and intervention session. Written informed consent was obtained from every participant before the commencement of each session.

The study was designed and reported in accordance with the Consolidated Standards of Reporting Trials (CONSORT) guidelines⁴⁹. Specific CONSORT extensions for within-subject designs and nonpharmacological interventions⁵⁰ were also taken into account (Fig. 6). Owing to the within-subject, nonpharmacological nature of the protocol, blinding of participants and investigators was not feasible, as both were necessarily aware of the intervention conditions. To minimize bias, however, participants were not informed of the specific study hypotheses.

Fig. 6.

Consolidated Standards of Reporting Trials (CONSORT) flowchart of a three-arm randomized controlled trial and the number of participants lost to follow‐up.

Procedures

This randomized crossover trial was designed to compare the immediate effects of three experimental conditions—LICE, MICE, and HIIE—on the temporal patterns of serum BDNF in healthy young adult males.

Participants were assigned to different sequence orders of the three exercise conditions using a computer-generated randomization procedure. A Latin square design was applied to ensure that each condition appeared equally across sessions and that order effects were minimized through systematic rotation of the condition sequences. A 7-day washout period was implemented between sessions to minimize potential carryover effects⁵¹. At the end of the trial, all participants had completed each of the three experimental conditions in a fully counterbalanced manner (Fig. 7).

Fig. 7.

Study design. BDNF: Brain-derived neurotrophic factor; YYIRT-1: Yo-Yo Intermittent Recovery Level 1 Test; MAS: maximum aerobic speed.

All exercise sessions were conducted on the same outdoor synthetic track within the university’s sport science facility and supervised by the same certified strength and conditioning coach to ensure consistency in protocol delivery. To minimize environmental variability, all trials were performed at a fixed time of day (between 09:00 and 11:00) to control for circadian influences on BDNF and lactate responses. Ambient temperature (21–24 °C) and relative humidity (40–55%) were monitored using a portable digital thermo-hygrometer before each session, ensuring that environmental conditions remained within a narrow and comparable range across testing days. These precautions ensured that the experimental environment was standardized and that physiological responses could be attributed to the exercise protocols rather than environmental fluctuations. Participants were instructed to abstain from vigorous activity for 48 h and from alcohol or caffeine intake for 24 h prior to testing, to achieve at least 7–8 h of sleep the preceding night, and to maintain stable dietary and lifestyle habits throughout the study.

At the initial visit, participants’ body height, weight and body fat percentage were assessed. Resting HR (HRrest) was subsequently recorded, after which the Yo-Yo Intermittent Recovery Test Level 1 (YYIRT-1) was administered to determine MAS, aerobic capacity, and peak HR (HRmax).

During the subsequent three visits, each participant completed all of the assigned experimental conditions. Every exercise session began with a standardized 10-minute warm-up consisting of light jogging at 50% MAS, dynamic mobility drills, and preparatory running exercises. These included 3 × 20 m accelerations and 2 × 50 m flying sprints, during which participants gradually increased their running speed. All exercise protocols lasted approximately 40 min in total, including the warm-up (10 min) and the exercise bout (30 min).

Exercise intensity was individualized based on each participant’s MAS, determined from the YYIRT-1, with MAS defined as the final running speed successfully completed in accordance with the test protocol. Exercise protocols were matched for total session duration and each participant’s running distance was calculated using the formula time (s) × MAS (m·s⁻¹). Venous blood samples were collected seven times to assess blood lactate concentrations and serum BDNF levels. HR was continuously monitored throughout all exercise sessions.

Exercise protocols

The exercise conditions are presented in Fig. 7. In the HIIE condition, participants performed 15-second runs at 105–110% of their MAS, covering a pre-calculated distance. Each run was followed by 15 s of passive rest at the end point of the bout. During each 15-s passive rest period, participants were instructed to promptly return to the starting mark and remain standing still until the next running bout began. This constituted one repetition (interval protocol with a 1:1 work-to-rest ratio), and participants completed 10 repetitions per set, lasting 5 min. A total of four sets were performed, interspersed with 2.5-minute passive recovery periods. This protocol resulted in a total exercise duration of 30 min. In contrast, both the MICE and LICE protocols consisted of continuous 30-minute runs without rest, performed at 70% and 50% of MAS, respectively. To minimize variability in effort induced by external motivation, no standardized verbal encouragement was provided during the exercise bouts or recovery periods. Participants received only neutral, safety- and procedure-related instructions. Interval compliance was defined as either reaching the cone at the whistle or verbal countdown cue (≥ 95% of target distance) or maintaining running speed within ± 5% of the prescribed value.

BMI

A body composition analyzer (BC-418MA, Tanita Corp., Tokyo, Japan) was used to measure body mass and body fat percentage. This device employs bioelectrical impedance analysis with multiple frequencies (1–50 kHz) to obtain comprehensive body composition information. Participant height was determined using a stadiometer (SECA, Hamburg, Germany).

YYIRT-1 and estimation of maximum oxygen uptake (VO₂max)

Participants completed the YYIRT-1, a progressively demanding, audio-paced shuttle test designed to assess aerobic capacity, following the protocol described by Bangsbo et al.⁵². The test was terminated if a participant failed to reach the line on two consecutive shuttles, voluntary withdrawal, or if the assessor terminated it for safety reasons. VO₂max was not measured directly; instead, it was estimated from YYIRT-1 performance using the prediction equation reported by Bangsbo et al.⁵²:

HR monitoring

To objectively monitor exercise intensity, HR was continuously recorded at 5-second intervals during each session using Polar V800 monitors (Polar Electro Oy, Kempele, Finland). Participants additionally wore Polar H10 chest straps for HR acquisition, and mean HR values for each repetition period were calculated during subsequent data analysis. In parallel, blood lactate concentrations were also measured.

Blood samples and assessment

Venous blood samples (8 mL) were drawn from the antecubital vein of participants in a seated position immediately before exercise and at all designated post-exercise time points, with all samples analyzed within a fixed time window between 14:00 and 16:00. After allowing the samples to clot for 30 min at room temperature, they were centrifuged at 3000 × g for 15 min. The resulting serum aliquots were stored at − 80 °C until analysis. Serum BDNF concentrations were measured using a commercially available sandwich enzyme-linked immunosorbent assay (ELISA) kit (SunRed Bio Company, Cat. No: 201-12-1303, Shanghai, China) according to the manufacturer’s protocol. Prior to ELISA, samples and reagents were equilibrated to room temperature. All assays were conducted under controlled laboratory conditions at a constant room temperature of 22–24 °C and relative humidity of 40–60%. The reported analytical range of the assay is (0.1–10 ng/mL) and the lower limit of detection is (0.05 ng/mL). All samples were measured within the validated analytical range. In addition, all samples were processed under the same experimental conditions and analyzed within a single assay run to minimize inter-assay variability. Serum samples were used directly without dilution according to the manufacturer’s instructions. Due to the limited sample volume and repeated blood sampling design, each sample was measured only once according to the manufacturer’s approved protocol. Optical density was read at 450 nm, and BDNF concentrations were calculated from a standard curve generated using a four-parameter logistic regression model. The obtained serum BDNF values fall within the range reported in previous human exercise studies using ELISA-based measurements, supporting the physiological rationality and external validity of the results.

Blood lactate concentrations were determined using a Super GL2 analyzer (Müller Gerätebau GmbH, Freital, Germany), which employs an enzymatic–amperometric electrochemical method. The device was calibrated prior to each measurement session in accordance with the manufacturer’s guidelines. All blood samples were anonymized, and sample identifiers were revealed only after the completion of analyses.

Data analysis

Prior to statistical analysis, data verification was performed to ensure data integrity. Raw data entries were cross-referenced with the original recordings to identify outliers and errors, and corrections were applied to ensure physiological plausibility. Following this verification the data are presented as the mean ± standard deviation. IBM SPSS Statistics (version 29.0, SPSS Inc., USA) was used for all the analyses. Normality was evaluated using the Shapiro-Wilk test and the data were found to be normally distributed. Two-way repeated-measures ANOVA was used to analyze BDNF, lactate, HRmean and % HRR. The Mauchly test confirmed the assumption of sphericity. Bonferroni-corrected tests were used for post-hoc comparisons. Partial eta squared () values were reported as effect size (small: 0.01 ≤) < 0.06; medium: 0.06 ≤) < 0.14; large:) ≥ 0.14)⁵³. A post-hoc power analysis was performed using G*Power software (version 3.1.9.7; University of Düsseldorf, Düsseldorf, Germany) to verify the statistical power of the observed large effect sizes⁵⁴. Statistical significance was set at p < 0.05. Visualization was performed using JAMOVI (Version 2.3) and JASP 0.19.3 programs.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

Conceptualization, Y.Z.B.; Ş.Ş.; Y.S.; H.Ş.; and S.P. methodology, Y.Z.B.; Y.S.; H.Ş.; and S.P. software, S.P. validation, S.P.; and H.Ş. formal analysis, S.P. and Y.S. investigation, Y.Z.B.; K.K.; and S.P.; resources, Y.Z.B.; and S.P. data curation, S.P.; and Y.Z.B. writing—original draft preparation, Y.Z.B. and S.P. writing—review and editing, Y.S.; H.Ş.; and Ş.Ş. visualization, S.P.; K.K.; and Y.Z.B. supervision, Y.Z.B. and Ş.Ş. project administration, Y.Z.B. All the authors have read and agreed with the published version of the manuscript.

Funding

This research received no external funding.

Data availability

The data that support the findings of this study are available from the corresponding author on reasonable request.

Declarations

Competing interests

The authors declare no competing interests.

Ethics approval

The study protocol received ethical approval from the Research Ethics Committee of the local university (approval no. TOGÜ.FRM.185-07/10; approval date: 25 April 2025). The study was registered retrospectively at ClinicalTrials.gov (identifier: NCT07175831 https://clinicaltrials.gov/study/NCT07175831) on 15th of September 2025. All procedures were performed in accordance with the principles outlined in the Declaration of Helsinki.

Informed consent statement

Informed consent was obtained from all subjects involved in the study.