Physical exercise as a therapeutic strategy in animal models of Alzheimer's disease: Molecular, behavioral, and histopathological evidence Authors: Helia Behrouzfar, Mahdi Gholami, Shokoufeh Hassani, Armin Salek Maghsoudi, Saeed Aghebat Bekheir Published: 2026 ABSTRACT AbstractBackgroundAlzheimer's disease (AD) is the most common neurodegenerative disease worldwide, imposing a substantial economic burden. As the population ages and cases rise, interest in early identification, prevention, and treatment has intensified. After establishing that pathology involves neurofibrillary tangles, neuroinflammation, and neuronal loss, interventions have been tested in laboratory animals to slow disease progression. One commonly tested intervention is physical exercise, an economical approach. Laboratory animal models of AD aim to demonstrate how different types and intensities of exercise affect disease pathology. This review aims to categorize and elucidate the mechanisms and their outcomes.Materials and methodsOriginal manuscripts concerning the impact of exercise, with or without medicinal intervention, on Alzheimer's dementia progression modeled in laboratory animals from 2020 onwards were included. The routes through which exercise exerts its impact at molecular, behavioral, and histomorphological levels were categorized.ResultsThe most frequent modeling approach was Aβ injection into the hippocampus of Wistar rats, followed by Streptozocin and other chemicals, and transgenic models. Exercise mainly consisted of treadmill or swimming. Interventions were categorized into short-, medium-, and long-term protocols, ranging from 3 to 24 weeks, with most at 4 weeks, 30–60 min per session, 5 days/week. Molecular tests primarily measured Reactive Oxygen Species (ROS), while the most common behavioral test was the Morris Water Maze (MWM), assessing learning and memory. Histopathological assessments focused on β-Amyloid plaque formation, dark cells, and glial activation. Most studies agreed on AD modeling methods, but exercise protocols varied in intensity, duration, and type. The combined use of molecular, behavioral, and histological tests to assess intervention effects was consistent.ConclusionPhysical exercise has been linked to lower oxidative stress and neuroinflammation, enhanced cognitive function, and fewer pathological changes in animal models of AD. However, variability in disease models and exercise protocols makes it difficult to identify the best exercise approach.Keywords: Alzheimer's, Dementia, Neurodegenerative, Exercise, Physical training, Animal models Introduction Alzheimer's is the most common type of dementia, accounting for 60-80% of cases, according to estimates provided by the Alzheimer's Association in 2024(1). As a progressive neurodegenerative disorder, AD causes approximately two-thirds of the estimated 57.4 million dementia cases in the world in 2019, expected to rise to 152.8 million cases in 2050(2). As the disease progresses, its effects on the patient and their environment deepen, causing severe financial and emotional distress. The patient lives for 8-10 years after receiving a clinical diagnosis until death, while losing the ability to perform the requirements of daily life and becoming wholly dependent [1,2]. According to global reports, estimated treatment and care for Alzheimer's disease (AD) patients in 2020 cost $305 billion [3]. The pathology of AD consists of the formation of tau neurofibrillary tangles and amyloid deposits, neuroinflammation, and neuronal loss, due to apoptosis and ferroptosis [4,5]. In spite of this, interventions that aim to reduce amyloid beta plaque formation and tau have all failed at the end of their clinical trials [[6], [7], [8]]. As of yet, no treatment that can modify the underlying cause of the disease has passed the final stages of trials [9]. Due to the difficulties in developing and procuring medicine for the treatment and symptom control of Alzheimer's, the use of alternative and cost-effective approaches is called for, which might slow the progression of the disease and increase quality of life [23,24]. Physical exercise is one of the alternative non-pharmacological treatment options widely studied to slow or reverse the progression of AD(14). Evidence suggests exercise can effectively improve memory and cognition, prevent and impede decay of white and gray matter in the brain, specifically in areas most affected by AD, such as the hippocampus [26]. Other mechanisms through which physical exercise exerts its impact include raising cerebral blood flow, modulating neuroinflammatory responses, and enhancing glucose metabolism [27,28]. Although there is growing evidence that exercise may affect animal models of Alzheimer's disease, findings vary widely across models, disease induction methods, exercise type and intensity, intervention duration, and measured outcomes. This variability makes it difficult to compare studies directly or identify which exercise features lead to molecular, behavioral, and pathological improvements. Additionally, it remains unclear whether early molecular and biochemical changes observed in short-term protocols translate into long-term behavioral and pathological benefits. This diversity and inconsistency underscore the need for a comprehensive synthesis of existing animal research. This paper attempts to answer the question: which of the exercise protocols used in vivo most efficiently alleviates AD symptoms [29]. Materials and Methods To gather data, PubMed, Web of Science, Scopus, and Google Scholar were searched in March 2026 using the following keywords: “Alzheimer's,” “AD,” or “Alzheimer's Disease”; “Dementia”; “rat model,” “animal model,” or “experimental model”; and “exercise,” “physical training,” “resistance training,” or “aerobic training.” The search focused on original research papers published between 2020 and 2025. Eligible studies were published in English, used laboratory animal models of Alzheimer's disease, and investigated physical exercise as a therapeutic strategy. They had to specify exercise duration and intensity and report histopathological results supported by molecular or behavioral data. Key details, including the animal model, method of disease induction, exercise type, intensity, duration, session frequency, and assessments, were extracted. The findings were then compared and summarized narratively, focusing on the Alzheimer's models, exercise protocols, and outcomes. Results and discussion General patterns in exercise protocols and Alzheimer's models Table 1 demonstrates the extracted data from all included manuscripts (Table 1). Most of the studies included in this review used intrahippocampal or intracerebroventricular injections of Aβ to induce reproducible AD-like phenotypes in rodents, whereas a smaller number relied on transgenic models such as APP/PS1, 3xTg, and TgF344-AD, or on metabolic/toxic models such as intracerebroventricular streptozotocin (STZ) and aluminum chloride to mimic sporadic AD. Aβ-based induction offers rapid onset and relatively homogeneous pathology and is therefore suitable for probing short-to medium-term effects of exercise interventions, while transgenic lines better capture the slow, progressive course of amyloid and tau pathology and are predominantly used in long-term paradigms. However, none of these models fully recapitulate the complexity of human AD, and interpretation of exercise effects must account for the specific limitations and biases inherent to each experimental system [10,11]. Across studies, the duration of exercise protocols ranged from 3 to 24 weeks, with training performed on 3-7 days per week, allowing the interventions to be broadly classified into short-term, moderate-length, and long-term regimens. The most frequently used configuration can be described as a moderate-length protocol: 30–60 min per session, 5 days per week, at approximately 50–90% of VO2max, corresponding to moderate-to-high-intensity aerobic or endurance training. In more than half of the experiments, initial VO2max or maximal running tests were employed to prescribe relative intensity and to enable progressive overload, whereas a minority used fixed speeds and durations, reflecting different levels of adherence to exercise physiology principles in animal models. Table 1Data extraction.Table 1Name of Author, Country, Year, RefAD InductionExercise TypeDuration and number of sessions per wkDuration per DayMolecular, histopathology, and behavioral testsMain FindingsSchimidt, Brazil, 2019 [10]Aβ Injection*Strength training#/Running via treadmill3 days/wk for 8 wks40 min, 60- 70% of VO2 maxROS, Short-term and long-term object recognitionStrength training prevented both object and social recognition memory deficits, showing superior neuroprotection to RunExStrength training ↓ lipid peroxidation and preserved AChE activityRunning protected only SR memory, ↓ ROS and lipid peroxidation, maintained total antioxidant capacityDare, Brazil, 2020 [11]Aβ Injection*Aerobic via treadmill4 days/wk for 4 wks30 min, 60-70% of the maximal indirect VO2ROS, Ellman method for AChE activity, H&E, OF, EPM, Object recognition test and social recognition testImproved object and social recognition memory, especially long-term object recognition↓ Hippocampal lipid peroxidationRestored AChE activityPartial tissue recoveryKlein, Brazil, 2020 [12]Aβ Injection*Swimming5 days/wk for 4 wks30 minGrip strength, Body Mass, MWM↓ ROS and nitric oxide synthase content↓ Tau phosphorylation↑ Mitochondrial fission via DRP1 oxidation (carbonylation)Vasconcelos, United Arab Emirates, 2020 [13]Aβ Injection*HIIT8 wks8 min incremental overload based on Maximal running test (MRT)ROS, BDNF/TrkB pathway, and neuritic plaques (NP), OF, MWM↑ Physical capacity, learning, and memory performance (p < 0.05)↑ CAT and SOD, ↓ lipid peroxidation↑ BDNF/TrkB activation, ↓ neuropathologyGholipour, Iran, 2022 [14]COMBAβ Injection*HIIT via treadmill5 days/wk for 8 wksIncremental time until 60 min. Speed at 90-50% of VO2 max.ROS, OF, Novel Object Recognition, EPM, Barnes Maze Test↑ Hippocampal oxidative balance,↓ Neuronal lossCombined intervention yielded additive neuroprotective effectsHegazy, Egypt, 2022 [15]ICV injection of STZSwimming5 days/wk for 4 wksIncremental swimming time until 60 min∼CSF glucose concentration, CSF irisin ELISA kit, Passive avoidance, Marble burying, Novel Object Recognition, MWM↓ Hippocampal FNDC5/irisin in ICV-STZ rats normalized by exercise or carnosine↓ Soluble Aβ and phosphorylated tau↑ BDNF and insulin signalingComparable neuroprotection and histological recovery in both treatmentsFakhraei, Iran, 2022 [16]Aβ Injection*Aerobic via treadmill4 wksIncremental up to 60 min at 15 m/min with breaksExpression of BDNF, TGF-b1, and TGF-b1 II, MWM↑ TGF-β1, TGF-β1II receptor, and BDNF expression in hippocampus↑ Spatial memory in Aβ + exercise vs. Aβ-only (p < 0.01)Exercise counteracted Aβ toxicity through BDNF and TGF-β1 pathwaysYang, China, 2022 [17]Transgenic rats called TgF344-ADAerobic via treadmill3 days/wk for 32 wks45 min at fixed 18 m/minROS, Aβ plaques, tau hyperphosphorylation neuroinflammation. Barnes maze task, Passive Avoidance task, Cued fear conditioning test, EPM, OF, Sucrose Preference, Forced Swim TestLong-term exercise pretreatment ↓ behavioral and pathological AD markersPrevented disease progression and cognitive declinePanahzadeh, Iran, 2022 [18]COMBAβ Injection*Endurance Training via Swimming7 days/wk for 3 wks30 min with the fixed speed of 18 m/minPRDX6 and GCN5L1 mRNAs and protein levels, Apoptosis, Shuttle box test spatial memorySwimming + clove improved oxidative balance↑ Mitochondrial homeostasisLiu, China, 2022 [19]APP/PS1 MiceHIIT via treadmill5 days/wk for 10 wks40 min, 10 sets of 23 m/min with 1 min restsCongo Red, GFAP, OF, EPM↑ Mitochondrial dynamics in both normal and AD mice↑ Cognitive and exploratory behaviorGodinho, Brazil, 2022 [20]Aβ Injection*HIIT via treadmillN/A for 8 wksIncremental based on MRTFNDC5, PPARGC1A, ITGB5-R, MWM↓ Aβ [[1], [2], [3], [4], [5], [6], [7], [8], [9], [10], [11], [12], [13], [14], [15], [16], [17], [18], [19], [20], [21], [22], [23], [24], [25], [26], [27], [28], [29], [30], [31], [32], [33], [34], [35], [36], [37], [38], [39], [40], [41], [42]]-induced cognitive deficits↑ FNDC5 protein levelsModulated FNDC5, ITGB5, PPARGC1A gene expression↑ Muscle fiber adaptation and neuroprotective molecular changesPark, Korea, 2022 [30]COMB3XTg mouse model of ADTreadmill6 days/wk for 12 wksIncremental time and incremental pace until 50 min of the main exercise at 13 m/minIba-1, GFAP, Mitochondrial Ca2þ Retention Capacity, H2O2 Emission, MWM, Step AvoidanceCombined exercise + 40 Hz light flicker improved mitochondrial function and neuroplasticity in early-stage AD↓ Aβ and tau accumulation, ↓ neuroinflammation, ↑ hippocampal protectionNoninvasive dual intervention effective for preventing cognitive decline through enhanced mitochondrial and synaptic integrityFeng, China, 2023 [31]STZ injected into the hippocampusHIIT via treadmill5 days/wk for 8 wksIncremental until 60 min a dayApoptosis, Aβ and p-tau, Branes maze test↑ HIIT enhanced Aβ and tau clearanceModulated astrocyte AQP4 polarizationIndicated glial phenotype regulation in neuroprotectionKhairy, Egypt, 2023 [32]IP aluminum chloride for ADAerobic via Swimming5 days/wk for 4 wks60 minSerum hypothalamic–pituitary–gonadal hormones, Beta amyloid levels, MWM↑ Testosterone↓ GnRH, LH, FSH, and β-amyloidModulated HPG axis dysregulation in ADKaraji, Iran, 2023 [33]Aβ Injection*Endurance Training via Swimming7 days/wk for 3 wks30 minExpression of NLRP1 and α7nAChR, MWM↓ α7nAChR↑ NLRP1Dark cell activationBasirat Dehkordi, Iran, 2024 [34]COMBAβ Injection*Aerobic via treadmill5 days/wk for 8 wksIncremental pace until 16 m/min for 60 minGDNF gene expression, Passive AvoidanceAIT + GB supplementation ↑ GDNF expressionImproved passive avoidance and long-term memorySynergistic neuroprotective effect in AD ratsRadfar, Iran, 2024 [35]COMBICV injection of AβTreadmill5 days/wk for 4 wksIncremental time and incremental pace until 60 min of the main exercise at 15 m/minBlood glucose levels, weight and rat head circumference, tail length and circumference. MWMExercise + intranasal insulin improved glucose metabolismPrevented spatial memory and brain growth deficitsDemonstrated additive benefit of combined interventionRadfar, Iran, 2024 [35]COMBICV injection of AβTreadmill5 days/wk for 4 wksIncremental time and incremental pace until 60 min of the main exercise at 15 m/minBlood glucose, expression of IGF1, BDNF, and GLUT4 in hypothalamus, Thioflavin-S staining, FTIC imaging for Insulin delivery and distribution in the body, MWMPretreatment exercise + intranasal insulin improved learning and memory↑ IGF1, BDNF, and GLUT4 pathway activationExercise amplified insulin's neuroprotective effectsChen, China, 2025 [36]COMBAD in APP/PS1 Transgenic MiceModerate intensity aerobic exercise via treadmill5 days/wk for 12 wks30 min with incremental pace, 60-73% of VO2 maxBcl-2 and BACE1 in the hippocampus, nuclear translocation of SUMO1 and IGF1R in the hippocampus, Apoptosis, ROS, IL-1b, IL-6, and TNF-a inflammatory markers, H&E, Nissl Staining, MWMPhysical exercise + IGF1R SUMOylation modulation improved metabolismEnhanced learning and memoryHall, USA, 2025 [26]TgF344-AD RatModerate intensity aerobic exercise via treadmill5 days/wk for 24 wksIncreasing, then decreasing $Grip strength, Body Mass, MWMExercise preserved skeletal muscle strength and improved memory in female TgF344-AD ratsMarked sex differences in body mass, motor coordination, and metabolic aging (♀ gain vs. ♂ loss)Highlights exercise as a sex-specific, nonpharmacological strategy for AD prevention and therapyAbbreviations: ICV (Intracerebroventricular), STZ (Streptozotocin), VO2 (Maximal Oxygen Consumption, H&E (Hematoxylin and Eosin), OF (Open Field), EPM(Elevated Plus Maze), MWM (Morris Water Maze), ORT (Object Recognition Test), SRT (Social Recognition Test), PAT (Passive Avoidance Task), EPM (Elevated Plus Maze), GFAP (Glial Fibrillary Acidic Protein), Iba-1 (Ionized Calcium-Binding Adaptor Protein-1), FNDC5 (Gastrocnemius Fibronectin Type III Domain Containing 5), PPARGC1A (PPARG Coactivator 1 Alpha), ITGB5-R (Integrin Subunit Beta 5), $ (Training was completed during the animal's dark cycle and began at 15 m/min for 15 min, progressing up to 24 m/min for 60 min. At 16 months of age, the intensity was decreased to 18 m/min, and the run time declined to 40 min), # (Vertical ladder with added weight), * (Aβ infusion in the dorsal hippocampus), & (With a workload of 1% body weight attached to tail), ↓ (Decreased), ↑ (Increased), and COMB (Combination therapy with exercise). Data extraction. Abbreviations: ICV (Intracerebroventricular), STZ (Streptozotocin), VO2 (Maximal Oxygen Consumption, H&E (Hematoxylin and Eosin), OF (Open Field), EPM(Elevated Plus Maze), MWM (Morris Water Maze), ORT (Object Recognition Test), SRT (Social Recognition Test), PAT (Passive Avoidance Task), EPM (Elevated Plus Maze), GFAP (Glial Fibrillary Acidic Protein), Iba-1 (Ionized Calcium-Binding Adaptor Protein-1), FNDC5 (Gastrocnemius Fibronectin Type III Domain Containing 5), PPARGC1A (PPARG Coactivator 1 Alpha), ITGB5-R (Integrin Subunit Beta 5), $ (Training was completed during the animal's dark cycle and began at 15 m/min for 15 min, progressing up to 24 m/min for 60 min. At 16 months of age, the intensity was decreased to 18 m/min, and the run time declined to 40 min), # (Vertical ladder with added weight), * (Aβ infusion in the dorsal hippocampus), & (With a workload of 1% body weight attached to tail), ↓ (Decreased), ↑ (Increased), and COMB (Combination therapy with exercise). Duration and intensity: from molecular changes to behavioral rescue A consistent pattern emerging from the included studies is a time-dependent gradient of exercise effects. Short-term protocols (≈3-4 weeks) primarily produced improvements at the molecular level, such as reduced oxidative stress markers, improved mitochondrial function [37,38], or increased neurotrophic signaling, with less robust or less consistently reported changes in behavior or histopathology. In contrast, moderate-to long-term training (≥8 weeks) more frequently translated these molecular benefits into measurable improvements in spatial learning and memory [39], reduced amyloid and tau pathology, and decreased neuronal loss on histological examination. This supports a “dose–response in time” concept in which early molecular adaptations precede and likely enable the later emergence of structural and behavioral protection. High-intensity interval training (HIIT) paradigms, even with relatively brief daily sessions (≈8-30 min), were able to increase aerobic capacity, improve redox balance, and activate BDNF/TrkB signaling, as well as through neurochemical such as insulin-like growth factor-I(IGF-1), insulin-like growth factor binding proteins, vascular endothelial growth factor (VEGF), homocysteine, tumor necrosis factor (TNF-α), and interferon-γ [40] and in several reports, these changes coincided with reduced neuritic plaques and better performance in the Morris water maze. Continuous moderate-intensity treadmill running and endurance swimming over 4-12 weeks similarly improved oxidative and neurotrophic profiles and were consistently associated with shorter escape latencies, increased time spent in the target quadrant during probe trials, and partial normalization of emotional and exploratory behaviors in tasks such as the open field and elevated plus maze. Taken together, both HIIT and continuous aerobic training appear capable of affecting the full cascade of outcomes, from biochemical markers to histopathology and behavior, provided that total training volume and duration are sufficient. Convergence between exercise-induced pathways and AD pathology The canonical AD-like pathology in these models includes extracellular Aβ deposits [41], hyperphosphorylated tau, oxidative damage, neuroinflammation and apoptotic or ferroptotic neuronal loss, which manifest histologically as increased “dark neurons,” reduced neuronal density in CA1 and CA3 [42], and intense staining for amyloid and glial markers such as GFAP and Iba-1 [43]. Behaviorally, these changes correspond to impaired acquisition and retention in the Morris water maze and Barnes maze, reduced time in target quadrants during probe trials, decreased exploratory activity in the open field, and deficits in passive avoidance and other memory tasks. A key observation from this review is that the very pathways driving AD pathology—oxidative stress, inflammatory cascades, and synaptic and neurotrophic dysfunction—are also the primary targets through which exercise exerts its neuroprotective and disease-modifying effects. Regarding oxidative stress, multiple studies reported that exercise reduced hippocampal reactive oxygen species and lipid peroxidation, while enhancing antioxidant defenses such as superoxide dismutase (SOD) and catalase, leading to a more favorable redox balance. These biochemical adaptations were paralleled by a lower proportion of dark neurons, preserved mitochondrial integrity, and reduced structural damage on hematoxylin–eosin staining. At the behavioral level, improved redox status co-occurred with better spatial learning and memory, reduced anxiety-like behavior in elevated plus maze tests, and restored exploratory patterns in the open field, suggesting that mitigation of oxidative damage is a central mechanistic bridge between exercise and functional rescue. Inflammation-related outcomes provide a similar picture: exercise attenuated pro-inflammatory cytokines such as interleukin (IL)- 1β, IL- 18, IL-1β, IL-6, and TNF-α and modulated downstream pathways including NF-κB and p38 MAPK, often together with a reduction in GFAP and Iba-1 immunoreactivity [42,44,45]. Several studies suggested a shift of microglial activation from a detrimental M1 phenotype toward a more reparative M2 profile, accompanied by increased anti-inflammatory mediators such as TGF-β. This immunomodulation was associated with reduced amyloid burden, improved synaptic organization, and consistent gains in spatial learning and memory performance [[46], [47], [48]]. In parallel, upregulation of hippocampal BDNF and activation of TrkB, often with downstream engagement of CREB and PI3K/Akt or MAPK signaling, were recurrent findings and aligned with enhanced neurogenesis, synaptic plasticity, and maintenance of neuronal density in hippocampal subfields [46,[49], [50], [51], [52], [53], [54], [55]]. Regarding neurotrophins, including nerve growth factor (BDNF), neurotrophin-3 (NT- 3), NT-4, NT-5, NT-6, and glial-derived neurotrophic factor (GDNF) all effect growth and the differentiation of neuron development in CNS and PNS as well as the survival of neuronal cells in response to stress [56], and they are all identified as drivers of brain elasticity as a result of exercise, albeit dependent on factors including age, genetics, and intensity of exercise [57]. Glial and metabolic contributions to the exercise response Beyond neurons, glial cells appear to be pivotal mediators of exercise benefits in experimental AD. By dampening excessive astrocytic hypertrophy and modifying the expression of GFAP and aquaporin-4, exercise may improve interstitial fluid dynamics and glymphatic clearance, thereby facilitating removal of Aβ and stabilizing synaptic microenvironments. Microglial responses to training, characterized by reduced nitric oxide production and a shift toward M2-like profiles, further support a scenario in which exercise “recalibrates” glial function from a chronically activated, neurotoxic state to a more homeostatic, pro-repair mode [58]. Metabolic and endocrine mechanisms also contribute significantly to the observed neuroprotection. A subset of studies combined exercise with interventions targeting insulin and IGF-1 signaling or with nutraceuticals such as Ginkgo biloba or Syzygium aromaticum, and showed additive or synergistic effects on glucose utilization, insulin sensitivity, and activation of IGF-1 and GLUT4 pathways in hippocampus and hypothalamus. These combined approaches often produced larger improvements in spatial and avoidance memory, reduced growth retardation, and more pronounced attenuation of AD-like pathology compared with either intervention alone, suggesting that exercise is a potent amplifier of beneficial metabolic and trophic signals in the AD brain (Fig. 1). Overview of the effect of Alzheimer's and Exercise on the brain. Overall synthesis of findings Overall, evidence from the reviewed studies indicates that physical exercise in animal models of AD confers benefits across molecular, behavioral, and pathological levels. At the molecular level, studies report reduced oxidative stress and lipid peroxidation, decreased markers of neuroinflammation and Tau phosphorylation, and improved neurotrophic signaling and mitochondrial health. Behaviorally, exercise primarily enhances learning and memory, as evidenced by better performance in object recognition tests and the Morris Water Maze. Some studies also report reduced tissue damage and pathological features linked to the disease. These findings suggest that exercise affects multiple pathways involved in AD pathophysiology. Nevertheless, variations in animal models, exercise type, intensity, duration, and outcome measures pose challenges for direct comparisons of results and for identifying an optimal protocol. Limitations and implications for future protocol design Despite broadly positive findings, there is considerable heterogeneity across studies regarding animal species and strains, age and sex, type of AD model, and the exact prescription of exercise intensity and progression, which complicates direct comparison and meta-analytic synthesis. Many experiments focused on a limited set of biomarkers, most commonly BDNF and general oxidative stress indices, while neglecting a more integrated assessment of oxidative, inflammatory, neurotrophic, and metabolic pathways within the same animals. As a result, our ability to map the full mechanistic network linking exercise “dose” to structural and behavioral outcomes remains incomplete. A key limitation is the timing of starting exercise in relation to Alzheimer's pathology development. Exercising before pathology appears might offer more preventive and neuroprotective benefits, while exercising afterward could provide therapeutic or disease-modifying effects. Due to these timing differences and inconsistent reporting across studies, direct comparisons are challenging. Future research should precisely document when exercise begins and directly compare preventive versus therapeutic effects. According to our current accepted terms, high-intensity aerobic training is characterized by 76-95% of maximal heart rate, 80% of maximal oxygen consumption, or 70-90% of cardiac output [40]. Future work would benefit from moving toward more standardized and transparently reported exercise protocols, for example: defining intensity as a percentage of VO2max or maximal running speed, using 30–60 min sessions on approximately 5 days per week, and extending interventions to at least 8 weeks to capture molecular, histopathological, and behavioral endpoints in a unified design. Incorporating broader behavioral batteries combining the Morris water maze with tests of anxiety, depression-like behavior, and different memory domains, and pairing these with comprehensive panels of oxidative, inflammatory, neurotrophic, and metabolic markers would yield a more holistic view of how exercise remodels the AD brain. Finally, the relatively limited use of long-term transgenic models and the under-representation of females and aged animals highlight the need for studies that systematically address sex, age, and disease stage, which are essential for translating preclinical exercise paradigms into precise, patient-tailored recommendations in clinical settings. Conclusion Animal studies indicate that physical exercise offers multiple neuroprotective effects in various Alzheimer's disease models. Exercise has been linked to decreased oxidative stress and neuroinflammation, enhanced mitochondrial function, and regulation of neurotrophic and metabolic pathways. Such changes correlate with better learning, memory, and fewer pathological features in many studies. Results vary depending on the exercise protocol's characteristics, like intensity and duration; shorter protocols tend to induce more molecular and biochemical changes, while longer ones (at least 8 weeks) show greater behavioral and pathological improvements. However, heterogeneity in animal models, age, sex, disease induction methods, exercise type and intensity, and intervention duration complicates identifying an optimal protocol and applying findings broadly. Current evidence supports exercise as a promising multi-targeted approach to modulate Alzheimer's-related processes in preclinical trials. Future research should aim for standardized designs and combined molecular, behavioral, and pathological assessments to identify the most effective exercise strategies. These insights could inform more rigorous preclinical studies and guide clinical translation. CRediT authorship contribution statement Helia Behrouzfar: Writing – original draft, Formal analysis, Data curation. Mahdi Gholami: Methodology. Shokoufeh Hassani: Validation. Armin Salek Maghsoudi: Validation, Supervision. Saeed Aghebat Bekheir: Writing – review & editing, Visualization, Supervision. Ethical statement Not applicable. Funding This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors. Declaration of competing interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. Acknowledgments Not applicable. Contributor Information Helia Behrouzfar, Email: h.behrouzfar111@gmail.com. Mahdi Gholami, Email: gholami.mahdi.1360@gmail.com. Shokoufeh Hassani, Email: shokoufehasani@gmail.com. Armin Salek Maghsoudi, Email: Armin.salek@gmail.com. Saeed Aghebat Bekheir, Email: saeedaghebat@gmail.com. Data availability No data was used for the research described in the article. References Associated Data Data Availability Statement No data was used for the research described in the article.