The combined effects of probiotic and high-intensity interval training on memory function in high fat diet-fed rats

Abstract

Introduction

This study aimed to investigate whether the probiotic Lactobacillus rhamnosus GG (LGG) (alone or combined with high-intensity interval training (HIIT)) could improve cognitive, electrophysiological changes, oxidative stress and metabolic parameters in HFD-fed rats.

Method

Rats were randomly divided into four groups (n = 8): HFD group, HFD + LGG group, HFD+ HIIT group, and HFD + LGG+ HIIT group. Rats were fed HFD daily for a period of 16 weeks, during which LGG (1 × 10¹⁰ colony forming unit (CFU)/ rats, orally), and HIIT protocol were administered four times a week on alternating days. At the end of study, assessment of social behavior, memory function, and Long-term potential (LTP) were performed using three-chambered apparatus, Y-maze task, and electrophysiology technique, respectively. Next, oxidative stress, lipid profiles, and liver enzymes were evaluated with routine kits.

Results

Both LGG and HIIT alone or in combination improved working memory, social memory, and LTP in HFD-fed rats. In addition, both LGG and HIIT alone or in combination increased the hippocampal levels of superoxide dismutase, catalase, and increased the serum levels of high-density lipoprotein (HDL), and decreased the serum levels of leptin, triglyceride, cholesterol, low-density lipoprotein (LDL), aspartate transaminase (AST), alanine transaminase (ALT), and alkaline phosphatase (ALP) in HFD-fed rats.

Conclusions

The combination of LGG and HIIT provides a multi-pathway intervention that improves HFD-induced memory impairments by concurrently targeting oxidative stress, dyslipidemia, and hippocampal synaptic function. This supports the potential of combined lifestyle and microbiome-based therapies for preventing metabolic and cognitive disorders.

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1186/s12993-026-00347-9.

Introduction

Research on how diet affects cognitive health has grown significantly, with notable emphasis on the adverse health consequences of high-fat diets (HFD) [1, 2]. Diets high in fat are linked to impaired cognitive function [1, 3]. The underlying mechanism by which HFD induces cognitive decline has yet to be fully characterized. Several studies reported that high-fat diets are associated with metabolic disorders that can negatively impact brain health through mechanisms like insulin resistance, oxidative stress, and inflammation in the brain [4]. HFD contributes to hippocampal injury-induced cognitive dysfunction by increasing blood-brain barrier permeability, microglial activation, and oxidative stress [5, 6]. The latter stimulates excessive ROS generation, which leads to increased neuroinflammation and ensuing cognitive dysfunction [7]. Recently, it has also been shown that HFD affects gut microbiota composition and is associated with cognitive dysfunction [8]. This dysbiosis often leads to an increased gut permeability (leaky gut), which promotes systemic inflammation and the circulation of bacterial lipopolysaccharides (LPS) that can penetrate the blood-brain barrier [9]. Furthermore, alterations in microbial metabolites, such as reduced levels of short-chain fatty acids (SCFAs), directly impair hippocampal neurogenesis and synaptic plasticity, critically contributing to cognitive impairment [10]. In this regard, modulation of gut microbiota might provide a therapeutic target for gut microbiota-related diseases [11].

The administration of live nonpathogenic microorganisms, named probiotics, improves cognition by modulating gut microbiota composition and regulating the microbiota–gut–brain axis [12–14]. For example, probiotics improve cognition in patients with obese-insulin resistance by modulating gut-brain axis, resulting in improving hippocampal plasticity and decreased microglial activation [15]. In addition to probiotics bacteria, exercise improves HFD-induced cognitive decline by targeting gut microbiota [16]. High-intensity interval training (HIIT), a physical exercise regimen characterized by short bursts of intense aerobic activity interspersed with brief periods of rest or low-intensity recovery, has recently been associated with improvements in both metabolic and cognitive disorders. Wang et al. [17] have shown that HIIT markedly decreased body weight and fat mass percentage, prevented liver lipid accumulation, and reduced adipose tissue inflammation in animal model of obesity induced by HFD. Furthermore, HIIT has been shown to improve memory by increasing antioxidants activity and BDNF levels in the hippocampus of ovariectomized rats [18]. In a study, exercise improved cognitive abilities related to object recognition and memory, with gut microbiota, specific microbial taxa (Lachnospiraceae, Coriobacteriaceae UCG-002) and intestinal barrier modifications serving as a potent modulator of lifestyle‑induced changes on brain and behavior [19]. Several studies have also reported that the combination of probiotics intake with HIIT had a more profound effect on decreasing memory disorder, anxiety-like behaviors, body weight, inflammation and insulin resistance in metabolic disorders [20–22]. Among the vast array of available probiotic strains, Lactobacillus rhamnosus GG (LGG) was specifically chosen for its well-documented “psychobiotic” properties and superior stability [23]. LGG is notably recognized for its ability to adhere strongly to the intestinal mucosa and enhance the gut-barrier integrity, thereby effectively reducing the systemic inflammation triggered by LPS leakage [24]. Importantly, LGG has demonstrated a unique capacity to modulate the neural components of the gut-brain axis, including the vagus nerve, resulting in significant anxiolytic and memory-enhancing effects that surpass those observed with many conventional Lactobacillus strains in preclinical models of cognitive impairment [25, 26]. This robust efficacy against both gut dysfunction and central nervous system deficits makes LGG the ideal candidate for multimodal interventions [27].

While individual and combined benefits of probiotics and exercise are established, a critical gap remains: no study has comprehensively evaluated the concurrent, multi-level combined impact of Lactobacillus rhamnosus GG and HIIT on both systemic metabolic health and cognitive function, alongside the underlying neurophysiological mechanisms. Specifically, the precise effects of this combination on fundamental mechanisms of memory, such as hippocampal long-term potentiation (LTP) (the electrophysiological correlate of learning and memory) are unknown. Therefore, our study was designed to test whether Lactobacillus rhamnosus GG (alone or combined with HIIT) could alleviate memory deficits, oxidative stress, and metabolic dysregulation in HFD-fed rats, with a particular focus on characterizing the combined improvement in electrophysiological endpoints alongside metabolic and cognitive markers.

Materials and methods

Animals

The study protocol was approved by the local ethic committee of Tehran University of Medical Sciences (registration no. 1400-3-233-56234) and was in accordance with the Helsinki Treat (Laboratory Animal Ethics). This study also was approved by the Research Ethics Committee of Tehran University of Medical Sciences (IR.TUMS.NI.REC.1401.007). This study was conducted in accordance with the ARRIVE guidelines 2.0. We used a total of 32 male Wistar rats weighing 250–280 g obtained from the Animal Breeding Facility of Tehran University of Medical Sciences (Tehran, Iran). The animals were housed in an animal lab under standard conditions (12 h light/dark cycle in a room at a temperature of 20–25 °C) with access to food and water ad libitum.

Experimental design

After one week of acclimatization, rats were randomly assigned to one of four experimental groups (n = 8 per group) using a computer-generated random number sequence (Microsoft Excel, RAND function). All behavioral tests, electrophysiological recordings, and biochemical assays were performed by experimenters blinded to group allocation. To ensure blinding, samples and animals were coded with numerical IDs that were not broken until after the completion of all measurements. The four groups were: (1) HFD (received HFD + PBS vehicle), (2) HFD+ Probiotic (received HFD + Lactobacillus rhamnosus GG), (3) HFD+HIIT (received HFD + PBS + HIIT), and (4) HFD+ Probiotic+ HIIT (received HFD + LGG + HIIT). A normal diet (20.6% kcal protein, 67.4% kcal carbohydrate, 12% kcal fat, 3.6 kcal/g) and a high-fat diet (20% kcal protein, 20% kcal carbohydrate, 60% kcal fat, 5.24 kcal/g) were used as described previously [27]. Rats were maintained on their respective diets for 16 weeks. Food consumption was measured daily in the evening immediately after the active (dark) cycle ended (Fig. 1).

Fig. 1.

Schematic diagram of the timeline for HFD and LGG administration, HIIT protocol, and behavioral testing. HFD, high fat diet; HIIT, high-intensity interval training; LGG, Lactobacillus rhamnosus GG

Lactobacillus rhamnosus GG was prepared from Department of Microbiology, Tehran University of Medical Sciences. Lactobacillus rhamnosus GG was suspended in sterile Phosphate-Buffered Saline (PBS) immediately prior to administration. The probiotic was administrated orally by gavage (10¹⁰ × 1 colony forming unit (CFU) per rat) once a day for 8 weeks (from the ninth to the sixteenth week). Rats in the non-probiotic groups (HFD and HFD+HIIT were given an equal volume of PBS through oral gavage daily as the vehicle control. This procedure ensured that the stress associated with the daily gavage did not differentially affect the behavioral and physiological outcomes between groups.

The exercise protocol was performed on a motorized treadmill. All rats in the exercise groups underwent a one‑week acclimatization period (15 min/day at 5–10 m/min) to familiarize them with the treadmill environment. Following acclimatization, each rat performed a maximal exercise capacity test to determine its individual peak running speed. The test started at 8 m/min, and the speed was increased by 1 m/min every 2 min until exhaustion. Exhaustion was defined as the point at which the rat remained on the shock grid at the rear of the treadmill for more than 10 s without attempting to resume running. The time and speed at exhaustion were recorded, and the maximal running speed (100% of capacity) was calculated for each animal. The HIIT protocol was then administered for 8 weeks (from week 9 to week 16), 4 days per week on non‑consecutive days. Each session consisted of 15 consecutive intervals: 2 min of running at 100% of the individual’s maximal speed, immediately followed by 2 min of rest (treadmill stopped). The total duration of each session was 60 min. To account for training adaptation, the maximal capacity test was repeated after 5 weeks of training, and running speeds were adjusted accordingly. Additionally, running speed was progressively increased by 1 m/min each week based on the animal’s updated maximal capacity [28]. The treatments (HIIT exercise and probiotics) lasted for 8 weeks (from the ninth to the sixteenth weeks). This intervention duration was selected based on established protocols in HFD rodent models that have demonstrated significant efficacy in improving metabolic parameters, modulating gut microbiota, and inducing beneficial neurological changes [29, 30].

The following a priori exclusion criteria were established: (1) Failure to complete at least 85% of the scheduled HIIT sessions (i.e., missing more than 4 sessions over 8 weeks). (2) Signs of illness or unexpected weight loss exceeding 20% of initial body weight. (3) Technical failure during electrophysiological recordings, including unstable baseline (< 10 min stability) or loss of fEPSP signal during LTP induction.

Behavioral testing apparatus

Y-maze task

The Y-maze test is used to assess working memory based on the spontaneous alternation behavior of rodents within a three-arm (Y-shaped) maze. In this test, rats are allowed to freely explore the maze without external reinforcement. Each trial begins with the rat positioned at the end of one arm, facing the central intersection. The animal is then given 5 min to explore. An arm entry is recorded when the rat’s hind paws fully enter an arm. A successful alternation is defined as consecutive entries into all three arms without repetition (i.e., visiting all three arms in sequence). Percentage alternation was defined as the ratio of actual to possible alternation (the total number of arm entries minus two), multiplied by 100, according to in the formula below:

The number of arm entries and the sequence of entries were scored manually by a trained observer who was blinded to the experimental groups. The percentage of spontaneous alternation was then calculated using the formula described above.

Social behavior

The three-chambered apparatus was used to evaluate social behavior. After habituation to the testing apparatus, each rat was kept in the central chamber. In this test, one side chamber contained a wire cage housing a non-manipulated (naïve) conspecific rat, while the opposite chamber held an identical but empty wire cage. The time spent by the experimental rat in each chamber was recorded to assess sociability, with increased time spent near the conspecific indicating a preference for social interaction. Animals with social behavior impairments spend more time in the empty chamber than with the conspecific [31]. The time spent by the experimental rat in each chamber (the chamber with the conspecific rat vs. the empty chamber) was measured manually by a blinded observer using a stopwatch or event-recording software.

Electrophysiology

Long-term potential

All surgical instruments and the stereotaxic frame were meticulously sterilized prior to use. Rats were deeply anesthetized with an intraperitoneal injection of urethane (Sigma, CAS Number: 51-79-6) at a dose of 1.5 g/kg, tailored to individual body weight. Throughout the procedure, animals were maintained in a temperature-controlled environment to minimize stress and movement artifacts. The scalp was shaved using an electric razor and disinfected with ethyl alcohol. The depth of anesthesia was confirmed by the absence of tail and toe pinch withdrawal reflexes. The animal was then securely positioned in the stereotaxic apparatus using ear bars inserted into the auditory canals, with correct placement verified by the corneal blink reflex. Following a midline scalp incision with fine scissors, the periosteal connective tissue was gently removed using a dental instrument to expose the skull. A bipolar stainless-steel stimulating electrode (0.125 mm diameter) was implanted into the medial perforant pathway using the stereotaxic coordinates 4.2 mm lateral to lambda and − 3.2 mm ventral from the brain surface. A second stainless-steel recording electrode was initially positioned in the dentate gyrus (DG) at coordinates 3.8 mm posterior and 2.2 mm lateral to the bregma. To ensure optimal placement, the final depth and position of the recording electrode were finely adjusted to locate the site yielding the maximal amplitude of both the field Excitatory Postsynaptic Potential (fEPSP) and the Population Spike (PS). This combination of precise stereotaxic targeting and functional mapping guaranteed a high degree of accuracy in targeting the DG granule cell layer. The recorded extracellular field potentials were amplified 1000x via a preamplifier headstage. The analog signals were then digitized at a sampling rate of 10 kHz using an A/D interface (Science Beam Co., Iran) and filtered with a band-pass of 1 Hz to 3 kHz. Data acquisition and analysis were performed using the Eprobe software suite. Stimulation was delivered as biphasic square-wave pulses, each 200 microseconds in width.

The correct positioning of the stimulating and recording electrodes was implemented based on previous studies [32, 33]. Furthermore, verification of the proper electrode placement was conducted using the occurrence of papulation spikes and spike waveform assessment, and the input/output (I/O) protocol was carried out to plot the corresponding curve [34].

The stimulation intensity was systematically adjusted to elicit 40% of the maximal fEPSP/PS response. Before inducing long-term potentiation (LTP), input-output relationships were assessed by generating stimulus-response curves across a range of intensities (100–1200 µA). After establishing a stable 30-minute baseline recording, LTP was induced by applying high-frequency stimulation (HFS), which consisted of 10 trains of 10 pulses at 400 Hz, with an inter-train interval of 10 s. Following HFS, baseline stimulation parameters were reinstated, and recordings were continued for at least 60 min. For LTP induction was considered to occur when the fEPSP changes (PS amplitude or EPSP slop) by at least 15%. Throughout the experiment, evoked responses were averaged over five consecutive trials at 10-second intervals for analysis. All recordings were conducted under stable and controlled conditions [35].

Paired-pulse response

To assess presynaptic neurotransmitter release dynamics and short-term synaptic plasticity, inter-pulse intervals (IPIs) were established following a 30-minute baseline recording. Paired electrical pulses were administered at 40% of the maximal stimulus intensity, with intervals of 20, 30, 50, 70, 100, 150, and 300 milliseconds tested. For each IPI, 10 consecutive responses were recorded and averaged. Two primary metrics were evaluated: (1) the percentage difference in the field excitatory postsynaptic potential (fEPSP) slope between the second and first pulse, and (2) the percentage change in population spike amplitude from the first to the second response. These comparisons were analyzed across all tested intervals to determine synaptic modulation patterns [35].

Biochemical parameters

At the end of experiment, blood was collected via cardiac puncture and centrifuged at 3000 r.p.m for 15 min, and serum lipid profiles and liver enzymes were analyzed. Included in the analyses were total serum levels of AST, ALT, ALP, total triglyceride, cholesterol and HDL-C using standard kits. In addition, LDL-C was calculated using the equation: LDL-C = [TC-(HDL-C)]-(triglyceride/5).

The serum levels of leptin were measured by ELISA kit (Cat No: RE2739R, Reed Biotech, China).

To assess antioxidant biomarkers in the hippocampus, frozen tissue samples were first digested and mixed in RIPA buffer (from Thermo Scientific, Waltham, MA, USA), which also contained protease inhibitors (complete and phospho STOP from Roche, Basel, Switzerland). Superoxide dismutase (SOD) and catalase (CAT) activities were analyzed with ready-made test kits (CA061, Canvax), following the manufacturer’s instructions. A POLARstar Omega plate reader was used to take absorbance readings—450 nm for SOD and 530/590 nm for CAT.

Statistical analysis

Statistical analyses were performed with Prism 8 software (Graph Pad Software, USA). Analyses of all the data from experimental groups were blindly conducted. Statistical comparisons for body weight gain, food intake, social recognition memory tests, and fEPSP recordings were carried out by one-way analysis of variance (ANOVA) followed by Tukey’s post hoc test. One-way ANOVA accompanied by Tukey’s post hoc test was used for comparisons involving more than two groups. The results are presented as mean ± standard error (S.E.), with differences regarded as statistically significant at a threshold of P < 0.05.

Results

Body weight gains and food intake

Figure 2a and b show the weekly body weight gains and food intake of the experimental groups from week 0 to week 16. Rats fed a high-fat diet HFD showed a steady rise in both body weight and food intake each week. Two-way ANOVA was performed, followed by Tukey’s post hoc test, and the weight gain was significantly different between the groups. Specifically, a significant increase in weight was found over time: F (2.53, 17.76) = 642.3, P < 0.0001, as well significant difference in animal weight between the groups: F (1.53, 10.72) = 7.52, P < 0.01. There was also a significant interaction effect: F (3.63, 25.42) = 14.44, P < 0.0001. Our findings indicate that the weight reduction was significant in the all-treatment groups compared to the HFD group after 12–16 weeks of HFD exposure (Fig. 2b). Our results also revealed significant difference in food intake over time: F(2.38, 11.92) = 109.6, P < 0.0001; between groups: F(1.83, 9.18) = 26.90, P < 0.001; and in interaction effects: F(3.58, 17.94) = 11.75, P < 0.0001. These results indicate that compared with animals of HFD group, food intake significantly decreased after 11 weeks in both the HIIT and combined HIIT and probiotic groups. In contrast, animals on either the HIIT diet alone or in combination with probiotics showed more stable food intake (Fig. 2b). These findings suggest that it is not only a high-fat diet that is associated with greater weight gain, but also one that promotes higher levels of food intake.

Fig. 2.

The effects of the probiotic Lactobacillus rhamnosus GG, administered either alone or in combination with high-intensity interval training (HIIT), on weight gain (n = 8) (A) and food intake (n = 6) (B) in rats subjected to a high-fat diet (HFD). *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001 HFD compared to HFD + HIT. #P < 0.05; ##P < 0.01; ###P < 0.001. HFD compared to HFD + LGG. $P < 0.05; $$P < 0.01; $$$P < 0.01 HFD compared to HFD + HIT+LGG. HFD, high fat diet; HIIT, high-intensity interval training; LGG, Lactobacillus rhamnosus GG

Behavioral tests

Y-maze test

The Y-maze test showed that both the probiotic and HIIT treatments, alone or combined, led to a significant effect on the alteration index and the total number of arm entries (F(3,19) = 20.06; P < 0.0001, Partial η² = 0.76, for total, and F(3, 22) = 16.76; P < 0.0001, Partial η² = 0.94 for alteration index). HFD led to a significant decrease versus all treatments on total arm entries (P < 0.0001, P < 0.001, and P < 0.05 for Probiotic, HIIT, and combination treatment, respectively) (Fig. 3a). We also observed a significant decrease in the alteration index induced by HFD in comparison with the treatment groups: P < 0.0001, P < 0.001, and P < 0.05, as compared to probiotic, HIIT, and combination treatments, respectively (Fig. 3b). Taken together, these results indicate that in the HFD group, cognitive impairment was significantly higher compared to the animals in the other groups. In addition, both Lactobacillus rhamnosus GG and HIIT alone and in combination show promise in enhancing working memory in animals on HFD.

Fig. 3.

Evaluation the effects of a high-fat diet (HFD) and subsequent treatments with the probiotic Lactobacillus rhamnosus GG or high-intensity interval training (HIIT), both individually and in combination on parameters related to working memory, entrance to arms (a), and alterations (%) in Y maze test (b), and parameters related to social interaction, sociability (c), and social memory (d) in three-chamber test. Data are presented as mean ± standard error of the mean (SEM) (n = 6 per group). ** P < 0.01 for HFD and. # P < 0.05 for HFD + HIT+LGG. HFD, high fat diet; HIIT, high-intensity interval training; LGG, Lactobacillus rhamnosus GG

Social behavior

The three-chamber test was used to assess sociability. In the sociability phase (Fig. 3c), a two-way repeated measures ANOVA revealed a significant interaction between row factor (empty cage vs. stranger rat) and column factor (experimental group) (F(3, 15) = 8.615, P = 0.0015, Partial η²: 0.89), indicating that the preference between the two chambers differed across groups. Post-hoc comparisons showed that the HFD group spent significantly more time investigating the empty cage than the stranger rat (P < 0.01), indicating impaired social motivation. In contrast, the combined HFD+HIIT + LGG group showed a significant preference toward the stranger rat over the empty cage (P < 0.05), indicating preserved sociability. The HIIT alone and LGG alone groups did not show a significant preference for either chamber (P > 0.05 for both), suggesting that the combined intervention was most effective in restoring social motivation in HFD-fed rats (Fig. 3c).

In the social recognition memory task, a two-way repeated-measures ANOVA revealed a significant main effect of stimulus (stranger 1 vs. stranger 2) on investigation time (F(1, 5) = 7.38, P = 0.04, Partial η²: 0.89). However, there was no significant main effect of group (F(1.82, 9.10) = 0.68, P = 0.51) and no significant stimulus × group interaction (F(1.42, 7.11) = 0.43, P = 0.60). These results indicate that, at the statistical level, the treatment groups did not differ significantly from the HFD group in their ability to discriminate between a familiar and a novel conspecific. Within-group post-hoc comparisons showed a non-significant trend toward longer investigation of the novel rat in the three treatment groups (all P > 0.05, Fig. 3d), but this trend did not reach statistical significance. Therefore, the present data do not allow a conclusion that the interventions improved social recognition memory; replication with a larger sample size is needed to clarify whether the observed trends reflect a genuine effect (Fig. 3d).

Electrophysiological changes

To investigate the long-term effects of the probiotic Lactobacillus rhamnosus GG, both alone or in combination with HII), we measured baseline synaptic activity and plasticity in the hippocampal perforant pathway across all groups. Input-output (I/O) curves evaluate basal synaptic transmission, while paired-pulse facilitation (PPF) and long-term potentiation (LTP) tests assess short- and long-term plasticity, respectively. Figure 4 illustrates the relationship between stimulus intensity and population spike (PS) amplitude and slope, represented as the input/output (I/O) curve from recordings conducted during 60-second periods at each stimulus intensity for each animal and group. Two-way ANOVA with Tukey’s post hoc test revealed significant differences in both PS amplitude and EPSP slope relative to stimulus intensity [F (_{2.38, 9.55}) = 81.11, P = 0.001 for PS amplitude; F (_{2.08, 8.34}) = 154.7, P = 0.001 for EPSP slope]. Additionally, significant differences were observed among the experimental groups [F (_{1.18, 4.75}) = 190.9, P = 0.001 for PS amplitude; F (_{1.78, 7.12}) = 58.46, P = 0.001 for EPSP slope], as well as an interaction effect between stimulus intensity and group [F (_{3.14, 12.00}) = 3.51, P = 0.05 for PS amplitude; F(_{3.18, 12.27}) = 3.66, P = 0.05 for EPSP slope].

Fig. 4.

The effect of probiotic as Lactobacillus rhamnosus GG, alone or in combination with HIIT, in electrophysiological properties of hippocampal neurons in rats (n = 5). The acquired input/output (I/O) curve from a 60 s recording and single traces of the DG neurons’ reaction to numerous stimulus intensities (100 to 1000 µA) (n = 5). In short intervals between two strictly paired stimuli, the amplitude of PS and fEPSP slope were prompted by a paired pulse in all groups. (E-F) Single traces were recorded before and after HFS. The amplitude of PS and fEPSP slope changes (%) and time following HFS). Data are expressed as mean ± SEM (n = 5 rat/group for B-E). HFD, high fat diet; HIIT, high-intensity interval training; LGG, Lactobacillus rhamnosus GG

Short-term plasticity via paired-pulse facilitation is shown in Fig. 4. In the HFD group, paired-pulse stimulation at short interstimulus intervals failed to enhance either the PS amplitude or the excitatory postsynaptic potential (EPSP) slope. Statistical analysis using two-way ANOVA followed by Tukey’s post hoc test confirmed a significant group-level disparity in the paired-pulse ratios for both PS amplitude and EPSP slope [F(_{1.45, 5.82}) = 35.83; P < 0.001 for PS; F(_{1.33, 5.32}) = 18.45; P < 0.05 for slope], but not across intervals [F(_{2.84, 11.37}) = 1.55; P = 0.25 for PS; F(_{1.88, 7.52}) = 0.68; P = 0.52 for slope] or in the interaction between groups and intervals [F(_{3.18, 12.75}) = 1.67; P < 0.22 for PS; F(_{3.15, 12.63}) = 1.59; P = 0.23 for slope]. Significant differences in the HFD group were noted compared to the probiotics or combination of HIIT and probiotics across all stimulus intervals and HIIT at intervals of 20 to 70 milliseconds.

Figure 5 illustrates the induction and maintenance of long-term potentiation (LTP) in the dentate gyrus across experimental groups (HFD vs. treatment groups). Statistical analysis demonstrated that both PS amplitude-LTP and EPSP slope-LTP following high-frequency stimulation (HFS) were significantly reduced in the HFD group compared to the HIIT group and the combined probiotics-HIIT group. Additionally, the analysis indicated a significant time-dependent variation in PS amplitude after HFS induction [F (_{2.62, 10.50}) = 9.09; P < 0.01], between groups [F(_{1.48, 5.92}) = 12.14; P < 0.01], and in their interaction [F(_{2.69, 10.79}) = 3.71; P < 0.05]. EPSP slope results also showed significant differences over time after HFS [F(1.82, 7.28) = 12.27; P < 0.01], but these differences were not observed between groups [F(1.29, 5.18) = 1.54; P = 0.28] or in their interaction [F(1.76, 7.06) = 1.49; P = 0.28]. Medial perforant path stimulation at 400 Hz induced sustained synaptic enhancement in the treatment group, with effects persisting for 60 min post-stimulation, while HFD groups showed significantly weaker potentiation (P < 0.001). Quantitative analysis revealed group-wise differences in LTP magnitude, whether assessed through population spike amplitude or synaptic response slope measurements taken at regular intervals following high-frequency stimulation (Fig. 5). As shown in Fig. 5, there is a significant difference in HFD PS amplitude means from zero to sixty minutes compared to HIIT and combination HIIT plus probiotic animal groups. Our results also indicated that compared to HFD animals, there is a significant increase in EPSP slope in the HIIT group.

Fig. 5.

The effect of HIIT and LGG, alone or in combination with HIIT on LTP induction and maintenance in the DG up to 60 min after high-frequency stimulation (HFS) (n = 5). Data were normalized against the baseline period (− 30 to + 60 min). Data are expressed as mean ± SEM (n = 5 rat/group for B-E). HFD, high fat diet; HIIT, high-intensity interval training; LGG, Lactobacillus rhamnosus GG

The correlation between behavioral test performance and electrophysiological recordings, specifically PS amplitude and fEPSP slope, measured 30–60 min after HFS stimulation across all experimental groups. Correlation analysis revealed no significant relationship between the alteration index in the Y-maze test and either PS amplitude or fEPSP slope. For PS amplitude, the correlation coefficient was r = 0.192 (95% CI: − 0.286 to 0.591, P = 0.41). For fEPSP slope, the correlation coefficient was r = 0.27 (95% CI: − 0.205 to 0.647, P = 0.24) (Fig S1D-E).

Biochemical study

Serum lipid factors

Since HFD can affect blood lipid levels, we studied how the probiotic Lactobacillus rhamnosus GG both by itself and combined with HIIT influences triglycerides, cholesterol, LDL, and HDL in HFD-fed rats. Evaluation of serum levels of this lipid revealed significant differences among the experimental groups (One-way ANOVA with Tukey’s post hoc analyses, F (_{3, 16}) = 68.52; P < 0.001, Partial η² = 0.93 for triglyceride; F (_{3, 16}) = 31.80; P < 0.001, Partial η² = 0.85 for cholesterol; F (_{3, 20}) = 11.32; P < 0.0001, Partial η² = 0.97 for LDL and F (_{3, 20}) = 33.20; P < 0.01, Partial η² = 0.62 for HDL). HFD led to increased triglyceride, cholesterol and LDL of serum and decreased serum HDL compared to the all-treatments groups. In Contrast, HIIT or probiotic as Lactobacillus rhamnosus GG, alone or in combination with HIIT led to decreased triglyceride, cholesterol and LDL levels and increased HDL levels in serum with a significant effect on serum factors content all treatment group (Fig. 6). As shown in Fig. 6, there was statistically significant effect between treatment groups on triglyceride levels in the HFD + HIT compared to combination HIT and Lactobacillus rhamnosus GG, as well as a significant effect between serum triglyceride, cholesterol, LDL and HDL level in HFD + LLC compared to combination HIT and Lactobacillus rhamnosus GG (P < 0.01, P < 0.05, P < 0.01 and P < 0.05 respectively).

Fig. 6.

The effects of the probiotic Lactobacillus rhamnosus GG, administered either alone or in combination with high-intensity interval training (HIIT), on serum lipid factors, including triglycerides (a), cholesterol (b), low-density lipoprotein (LDL) (c), and high-density lipoprotein (HDL) (d) in rats subjected to a high-fat diet (HFD) (n = 5). HFD, high fat diet; HIIT, high-intensity interval training; LGG, Lactobacillus rhamnosus GG

Oxidative stress

Figure 7 shows the changes in the activities of hippocampal antioxidant enzymes, catalase, and superoxide dismutase (SOD), in experimental groups. Significant decrease was noted in enzymes activity in the high-fat diet (HFD) group compared to the treatment groups (F (3, 16) = 7.86; P = 0.0019 for catalase, Partial η² = 0.59; F (3, 20) = 43.19; P < 0.0001, Partial η² = 0.86 for SOD. The analyses conducted subsequently revealed a significantly increase catalase activity between the combination of high-intensity interval training and Lactobacillus rhamnosus GG group as compared with the HFD and HIIT groups. Our results reveal that HIIT or Lactobacillus rhamnosus GG, either alone or in combination, significantly increased SOD activity as compared to HFD group (P < 0.05 HIIT, P < 0.05 LGG, and P < 0.001 of the combination between HIIT and LGG). Moreover, SOD activity was found to be different between the LGG group and both the HFD + HIIT and combination HIIT and LGG group. (Fig. 7a-b). Additionally, correlation analysis revealed positive significant relationship between the alteration index in the Y-maze test and oxidative stress marker including CAT and SOD. For CAT, the correlation coefficient was r = 0.424 (95% CI: − 0.036 to 0.736, P = 0.03). For SOD, the correlation coefficient was r = 0.709 (95% CI: − 0.428 to 0.865, P = 0.0001) (Fig S1A−B).

Fig. 7.

The effects of the probiotic Lactobacillus rhamnosus GG (LGG), administered either alone or in combination with high-intensity interval training (HIIT), on hippocampal oxidative stress factors including catalase and SOD (a, b), serum levels of leptin (c), and serum levels of liver enzymes (d–f) in rats subjected to a high-fat diet (HFD) (n = 5). *P < 0.05; ***P < 0.001; ****P < 0.0001 HFD compared to all groups. #P < 0.05; HFD + HIT compared to HFD + HIT+LGG. $P < 0.05; HFD + LGG compared to HFD + HIT+LGG. HFD, high fat diet; HIIT, high-intensity interval training; LGG, Lactobacillus rhamnosus GG

.

Leptin level

A high-fat diet significantly elevates serum leptin levels in rats, which corresponds with increased body fat deposition and obesity [36]. Levels of this hormone in the HFD group compared to treatment groups (F (_{3, 16}) = 81.65; P < 0.001, Partial η² = 0.93). Specifically, our findings showed that high-intensity interval training or the administration of the probiotic Lactobacillus rhamnosus GG, either alone or combined with HIIT, significantly lowered serum leptin levels in all treatment groups when compared to HFD group (P < 0.0001 vs. L. rhamnosus GG; P < 0.05 vs. HIIT; P < 0.0001 vs. combination of HIIT and L. rhamnosus GG). Furthermore, a significant difference was observed in serum leptin levels between the Lactobacillus rhamnosus GG group and both the HFD + HIIT (P < 0.05) and the combination group of HIIT and LGG (Fig. 7c). A correlation analysis revealed no significant relationship between the alteration index in the Y-maze test and leptin, the correlation coefficient was r = − 0.177 (95% CI: − 0.584 to 0.300, P = 0.45) (Fig S1C).

Changes in liver enzymes

Figure 7 (d-f) shows the changes in liver enzymes, AST, ALT and ALP in the experimental groups. Increase enzyme activity in the high-fat diet (HFD) group compared to the treatment groups (F (_{3, 20}) = 20.42; P < 0.01, Partial η² = 0.82 for AST; F (_{3, 20}) = 20.63; P < 0.0001, Partial η² = 0.95 for ALT and F (_{3, 20}) = 23.24; P < 0.01, Partial η² = 0.76 for ALP). Post hoc Tukey’s analyses conducted subsequently revealed a significant difference between high-intensity interval training and the combination of high-intensity interval training and Lactobacillus rhamnosus GG group as compared with the Lactobacillus rhamnosus GG group (P < 0.05 HIIT compaired to LGG for AST, ALT and ALP; P < 0.05 LGG compared to HIIT + LGG for and ALT; P < 0.01 compaired to HIIT + LGG for ALP), (Fig. 7d-f).

Discussion

The aim of present study was to determine whether the combination of Lactobacillus rhamnosus GG and HIIT could decrease spatial and sociability dysfunction in HFD-fed rats, and to evaluate its mechanisms of action, addressing synaptic plasticity, dyslipidemia and oxidative stress. Firstly, our results of Y-maze and social behavior tests showed that HFD induced spatial and social memory dysfunction. These results are consistent with previous studies [1, 2], where HFD accelerated the progression of cognitive impairment assessed by Y-maze in a transgenic Alzheimer’s disease model [37]. HFD also led to medial prefrontal cortex impairments, and social recognition dysfunction [38]. Then, main findings of our study demonstrated that the combined intervention of Lactobacillus rhamnosus GG supplementation and HIIT exerted combined treatment effects in HFD-fed rats, leading to improvements in memory function and social behavior. In line with our results, a previous research has also reported that the combination of L. rhamnosus GG and HIIT exerts combined treatment benefits on anxiety-like behaviors in HFD-fed mice, further supporting the potential of this dual intervention for mitigating neurobehavioral and metabolic disturbances induced by HFD [20]. The combination of HIIT with probiotics (Lactobacilli cocktail) significantly improved associative memory in ovariectomized rats [21, 29, 39].

Learning and memory are known as electrochemical signal processes of long-term potentiation (LTP) in the hippocampus [40]. The induction of LTP depends on NMDA receptor activation, a critical mechanism for synaptic plasticity [41, 42]. As referred in the method section, LTP induction occurred when the fEPSP (PS amplitude or EPSP slope) changed by at least 15%. Here, we found that HFD consumption impaired LTP in the hippocampus’s area, consistent with prior research demonstrating that HFD disrupts LTP in rodent models [43–45]. Electrophysiological assessments of LTP and paired-pulse stimulation further support the conclusion that L. rhamnosus GG in combination with HIIT, enhances both short- and long-term memory in HFD-fed animals. These cognitive improvements are likely linked to enhanced neuroplasticity the brain’s ability to adapt and reorganize in response to experience and environmental stimuli. The combined probiotic-HIIT intervention appears to induce beneficial neurobiological changes, particularly within the hippocampus, a brain region vulnerable to the deleterious effects of high-fat diets. Such changes may help counteract the cognitive deficits typically associated with dietary-induced metabolic dysfunction. Electrophysiological assessments confirmed that the combined intervention restored LTP in the hippocampus. This restoration is mechanistically linked to the observed reduction in oxidative stress and dyslipidemia, as excessive ROS and lipid peroxides are known to impair NMDA receptor function and destabilize LTP [46]. Based on our indirect findings (reduced oxidative stress and improved LTP), we hypothesize that the observed restoration of LTP may involve preserved NMDA receptor function. It is plausible that by mitigating oxidative disturbances, the intervention preserved synaptic calcium signaling and BDNF expression, thereby enabling neuroplasticity.

Dyslipidemia and oxidative stress are closely linked to the learning and memory impairments observed in HFD-fed rodents. HFD elevates LDL and triglycerides, which are highly susceptible to free radical attack, leading to lipid peroxidation and further ROS production while reducing antioxidant defenses [47, 48]. Hippocampal neurons are particularly vulnerable to HFD-induced lipid peroxidation and oxidative stress [45]. Under normal conditions, low levels of ROS facilitate LTP induction as part of physiological signaling; however, excessive oxidative stress disrupts LTP and contributes to cognitive dysfunction [47]. HFD also alters gut microbiota composition by increasing Gram-negative bacteria, resulting in elevated lipopolysaccharide (LPS) and impaired intestinal barrier integrity [49]. Increased circulating LPS promotes systemic inflammation, oxidative stress, dyslipidemia, and insulin resistance, all of which can compromise cognitive function [8, 50–52]. In the present study, administration of Lactobacillus rhamnosus GG significantly improved cognitive dysfunction and restored LTP in HFD-fed rats, consistent with growing evidence that probiotics exert neuroprotective effects in various neurological disorders [53–55] and ameliorate gut dysbiosis-related pathology.

Lactobacillus improved memory impairment in animal model of diabetes probably by reducing oxidative stress [56]. L. rhamnosus, L. plantarum and L. reuteri significantly mitigated LPS-induced learning and memory impairments, likely through the attenuation of neuroinflammatory responses [57]. Another study reported that probiotics exerted beneficial effects on learning capacity and LTP, potentially through their protective influence on serum lipid profiles [58]. Previous studies also have revealed positive effects of probiotics like Lactobacillus rhamnosus to decrease body weight, hyperlipidemia, the levels of liver enzymes and oxidative stress in HFD- fed rodents and obese subjects [59–62]. Probiotics appear to contribute to reductions in body weight and fat mass, likely through modulation of the gut microbiota, enhancement of insulin sensitivity, and promotion of satiety [12]. Probiotics decreased HFD-induced obesity by modulating gut microbiota by increasing Bifidobacterium, Lactococcus, and Akkermansia [63]. Lactobacillus rhamnosus reduced HFD-induced body weight by modulating the gut microbiota, increasing Ruminococcus, Enterobacteriaceae, Romboutsia, Lachnoclostridium and Lacticaseibacillus rhamnosus [64]. In addition, the administration of L. rhamnosus 86 and L. johnsonni 3121 prevented obesity and improved dyslipidemia in HFD diet fed obese mice by altering gut microbiota secondary to reduced Firmicutes/Bacteroidetes ratio and increased F. prausnitzii [65]. L. rhamnosus LS-8 and L. crustorum MN047 also significantly reduced body weight, inflammatory cytokines and insulin resistance. In addition, these strains modulated gut microbiota by reducing the abundance of Bacteroides and Desulfovibrio and enhancing Lactobacillus and Bifidobacterium, resulting in increased SCFAs and decreased LPS [62]. Finally, Kong et al. [61] demonstrated that Lactobacillus rhamnosus GG improved dyslipidemia in HFD-fed mice via modulation of cholesterol synthesis and hepatic triglyceride.

Our findings of reduced hippocampal oxidative stress and improved serum lipid profile are consistent with the hypothesized gut-brain axis mechanism, although direct evidence (e.g., microbiota or cytokine analysis) is lacking in this study. It has been previously proposed that HFD-induced dysbiosis increases Gram-negative bacteria, elevating circulating LPS [66]. LPS binds to TLR4 receptors, triggering systemic inflammation and oxidative stress, which in turn compromises blood-brain barrier integrity and promotes neuroinflammation [66]. Based on our indirect data, it is plausible that probiotics like LGG compete with pathobionts, strengthen tight junctions, and reduce LPS translocation, thereby potentially attenuating this cascade. However, as we did not directly measure gut barrier integrity markers or circulating LPS, this interpretation remains speculative and requires confirmation in future studies. HIIT may further enhance gut barrier function and reduce visceral fat, a source of pro-inflammatory cytokines. Thus, the combined intervention may exert neuroprotective effects, possibly by reducing the peripheral inflammatory drive; nevertheless, the absence of direct measurements of inflammatory cytokines (e.g., IL-6, TNF-α) and gut microbiota composition prevents us from establishing a causal link. In addition, we propose several hypothetical mechanisms by which L. rhamnosus GG might ameliorate oxidative stress, though none were directly tested in this study: (1) Direct antioxidant activity through the production of metabolites like glutathione, (2) Indirect modulation by increasing beneficial SCFA-producing bacteria like Lactobacillus, where SCFAs such as butyrate activate Nrf2 pathway to upregulate endogenous antioxidant enzymes, and (3) Reducing the production of pro-oxidant molecules like LPS from gut pathobionts. This multi-pronged antioxidant action can directly protect hippocampal neurons from lipid peroxidation and preserve LTP. These proposed pathways are based on previous literature and our observed reduction in hippocampal oxidative stress markers; however, direct measurements of glutathione, SCFAs, Nrf2 activity, or LPS are needed to confirm these mechanisms.

Current evidence indicates that high-intensity interval training (HIIT) alone also improves both behavioral outcomes and biochemical parameters. Specifically, HIIT has been shown to attenuate memory impairment by upregulating markers associated with synaptic plasticity in HFD-fed mice [52]. In a study by Yuan et al. [53], an 8-week HIIT intervention significantly reduced body weight, TG, and cholesterol levels in HFD-fed rats, likely through modulation of hepatic fatty acid β-oxidation and inhibition of de novo lipogenesis. In a study, exercise improved cognitive abilities related to object recognition and memory, with gut microbiota, specific microbial taxa (Lachnospiraceae, Coriobacteriaceae UCG-002) and intestinal barrier modifications serving as a potent modulator of lifestyle‑induced changes on brain and behavior [19]. In addition to the beneficial effects of Lactobacillus rhamnosus GG supplementation and HIIT alone, the combined intervention of Lactobacillus rhamnosus GG supplementation and HIIT exerted combined treatment effects in HFD-fed rats, leading reductions in body weight, dyslipidemia, hepatic enzyme levels, and oxidative stress markers. In line with our results, a previous research has also reported that the combination of L. rhamnosus GG and HIIT exerts combined treatment benefits on body weight reduction, and elevated corticosterone levels in HFD-fed mice, further supporting the potential of this dual intervention for mitigating neurobehavioral and metabolic disturbances induced by HFD [20]. The combination of HIIT with probiotics (Lactobacilli cocktail) significantly increased adiponectin levels and insulin sensitivity, decreased body weight and lipid profiles in ovariectomized rats [21, 29, 39]. This combination also synergically decreased inflammation and oxidative stress in animal model of hepatic steatosis [67]. The superior efficacy of the combined LGG and HIIT intervention, compared to either alone, suggests a combined interaction that operates through the gut-brain-liver axis via complementary and reinforcing mechanisms.

It is of note that much of the beneficial effect on systemic metabolic features including both improvements in dyslipidemia, decreases in liver enzymes and normalizing leptin levels might be due to profound weight loss from this combined intervention. Weight loss, per se, is known to be a major metabolic improvement through increased insulin sensitivity and decreased adipose tissue inflammation and hepatic lipid accumulation. Thus, despite our data showing a strong impact of the double treatment on these parameters, we could not rule out whether molecular effects promoted by LGG and HIIT per se versus those dependent on weight loss are dominant, and this question should be addressed in pair-fed or weight-matched studies.

We hypothesize that HIIT rapidly improves systemic metabolism (enhancing β-oxidation and insulin sensitivity) and upregulates brain-derived neurotrophic factor (BDNF), thereby directly fostering synaptic resilience. Concurrently, HIIT enhances systemic and splanchnic blood flow, which may improve gut mucosal perfusion and integrity. This exercise-induced improvement in the gut environment could facilitate the colonization, survival, and metabolic activity of the supplemented L. rhamnosus GG. In parallel, LGG directly restores gut microbiota homeostasis, fortifies the intestinal barrier to reduce endotoxemia (LPS leakage), and enhances hepatic lipid clearance. Together, these actions create a positive feedback loop: the reduction in gut permeability and systemic inflammation by LGG may improve the metabolic and neurotrophic responses to exercise, while the exercise-induced improvements in gut health and metabolism may amplify the probiotic’s benefits. This bidirectional crosstalk results in a compounded reduction in the peripheral drivers of oxidative stress (dyslipidemia and LPS) while simultaneously enhancing the brain’s intrinsic defense and plasticity capacity. Consequently, the combined intervention creates a systemic and neural milieu far more favorable for hippocampal LTP and memory consolidation than either intervention could achieve in isolation.

Our findings on a rodent model of diet-induced obesity may have translational implications for human populations. Obesity and high-fat dietary patterns are increasingly recognized as risk factors for cognitive decline and dementia in humans, including Alzheimer’s disease and vascular cognitive impairment [65, 66]. The combination of probiotics and structured exercise training, such as HIIT, represents two safe, low-cost, and accessible lifestyle interventions that could be readily tested in human clinical trials. Our study provides preclinical proof of concept that combining these two interventions may produce superior cognitive benefits compared with either alone. However, direct translation requires caution, as human gut microbiota composition, exercise tolerance, and dietary habits differ considerably from controlled rodent models. Nevertheless, our results support the rationale for future randomized controlled trials examining whether LGG supplementation combined with HIIT can mitigate obesity‑associated cognitive impairment and improve brain health markers in at‑risk human populations.

Several limitations of this study should be acknowledged. First, our focus was on hippocampal electrophysiology and behavior; future studies should investigate other critical brain regions implicated in social and metabolic regulation, such as the prefrontal cortex and striatum. Second, we used only male rats; however, this limits the generalizability of our findings, as sex differences in response to HFD, exercise, and probiotics are well-documented. Third, the absence of a standard chow diet control group for the intervention arms limits our ability to discern whether the combined treatment fully restores normal function or merely attenuates HFD-induced damage. Forth, further investigations into specific neuroinflammatory markers (e.g., microglial activation, cytokine levels in the brain) are needed to fully elucidate the neuroimmune component of the observed synergy. Finally, while we propose modulation of the gut-brain axis as a central mechanism, we did not directly analyze changes in gut microbiota composition or sequencing, microbial metabolites (e.g., SCFAs), gut barrier integrity markers (e.g., plasma zonulin, circulating LPS), or systemic inflammatory cytokines (e.g., IL-1β, IL-6, TNF-α). Consequently, our interpretation of the gut-brain axis involvement is inferential and based on measured outcomes (oxidative stress, lipid profile, and LTP). Direct mechanistic validation is essential and should be addressed in future investigations.

Conclusions

In summary, this study shows that a combined intervention of Lactobacillus rhamnosus GG (LGG) supplementation and high-intensity interval training (HIIT) is associated with attenuation of HFD-induced deficits in spatial and social memory, hippocampal long-term potentiation (LTP), hippocampal oxidative stress markers, and serum metabolic parameters in rats. The observed improvements are correlated with the combined treatment, and the data are consistent with the hypothesis that complementary actions on the gut–brain axis—possibly involving reduced oxidative stress—contribute to these effects. However, given the correlational nature of the findings, we cannot infer causality. The precise mechanisms linking the combined LGG+HIIT intervention to cognitive and electrophysiological benefits remain to be directly elucidated. This study provides a preliminary rationale for future research exploring combined lifestyle and microbiome-targeted therapies, but experimental validation—including microbiome sequencing, gut permeability assays, and cytokine profiling—is necessary to confirm the hypothesized pathways and establish causal relationships.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

FK, FS, SA, O-RT, MA, SH, JF-B and FN contributed in the conception or design of the work and drafting of the manuscript. All authors confirmed the final version for submission.

Funding

This study was supported by Tehran University of Medical Sciences fund (Grant no 1400-3-233-56234).

Data availability

The primary data for this study is available from the authors on direct request.

Declarations

Ethics approval and consent to participate

This study carried out according to guideline of the local ethics committee at the Tehran University of Medical Sciences, Tehran, Iran (IR.TUMS.NI.REC.1401.007). Consent to participate is not applicable.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.