How Exercise May Slow Brain Aging by Improving Mitochondrial Quality Control

The most prominent manifestation of brain aging at the macro-structural level is the progressive reduction in brain tissue volume and weight. Research indicates that between the ages of 20 and 60, brain weight decreases by approximately 0.1% annually on average. After the age of 70, the rate of atrophy may accelerate, with an average annual weight reduction of 2 to 5% [11]. Cognitive decline accompanies atrophy across the entire brain or in specific regions, manifesting in reduced processing speed, working memory, reasoning, and executive function [12]. Crucially, this atrophy is not uniformly distributed but follows specific patterns. Through deep learning analysis of nearly 50,000 brain scans, researchers have identified five distinct patterns of brain atrophy. These patterns correlate with specific aging trajectories and disease risks. Different combinations of atrophy patterns can predict the risk of diseases such as Alzheimer’s and Parkinson’s, as well as the transition from normal cognition to mild cognitive impairment [13]. Among these, the frontal and temporal lobes (particularly the hippocampus) exhibit the most pronounced atrophy and are closely associated with higher cognitive functions such as memory and executive function [14,15].

At the microscopic level, aging triggers a series of complex cellular and molecular events.

At the systemic level, neural network activity in the aging brain exhibits characteristic alterations. A magnetoencephalography (MEG) study spanning ages 18 to 88 revealed that aging alters the microstate patterns of spontaneous neural oscillations. This manifests as shorter durations but increased frequency of alpha rhythms (8–13 Hz) associated with sensory processing microstates, potentially reflecting diminished sensory information processing efficiency and compensatory attempts. Alterations in theta and beta rhythms within specific microstates may correlate with age-related declines in motor function [27].

In summary, brain aging constitutes a non-linear, phased dynamic process exhibiting distinct characteristics across different life stages. The following section discusses how dysregulation of the mitochondrial quality control system underlies these age-related changes.

Mitochondrial biogenesis is the process of generating new mitochondria, primarily regulated by the peroxisome proliferator-activated receptor gamma coactivator-1α (PGC-1α) and its upstream modulators AMP-activated protein kinase (AMPK)/sirtuin 1, alongside downstream effectors nuclear respiratory factor 1/2 (NRF1/2), the estrogen-related receptor alpha (ERRα), and mitochondrial transcription factor A (TFAM) [29,30]. Under physiological conditions, AMPK and SIRT1 coordinately activate PGC-1α, which then promotes the transcription of mitochondrial genes through NRF1/2, ERRα, and TFAM, thereby supporting mitochondrial biogenesis. In the aging brain, impaired AMPK responsiveness, reduced sirtuin activity, weakened TFAM–mtDNA binding, and cumulative mtDNA damage collectively compromise mitochondrial biogenesis, thereby contributing to neuronal bioenergetic failure [31,32,33,34]. Although TFAM expression may increase in aged cerebral cortex, its binding to mitochondrial DNA (mtDNA) is markedly reduced, suggesting that compensatory upregulation is insufficient to preserve effective mitochondrial biogenesis [35]. In parallel, age-related accumulation of mtDNA mutations and deletions further compromises mtDNA replication and oxidative phosphorylation capacity [6,36,37]. Together, impairment of the AMPK/SIRT1/PGC-1α signaling axis, reduced TFAM–mtDNA interaction, and progressive mtDNA damage constitute a major molecular basis for defective mitochondrial biogenesis in the aging brain, thereby contributing to neuronal bioenergetic failure and functional decline [38]. In addition, the age-associated reduction in NRF2 may further aggravate mitochondrial vulnerability, highlighting NRF2-related pathways as potential targets for delaying brain aging [39]. In addition to these canonical pathways, recent evidence suggests that serum response factor (SRF) may modulate age-related mitochondrial dysfunction, while earlier studies have shown that the SRF–cofilin–actin signaling axis regulates mitochondrial dynamics [40,41]. However, its role in brain aging, particularly in the context of exercise, remains to be determined.

Mitochondrial dynamics constitute a vital process for maintaining mitochondrial morphological and functional homeostasis, encompassing two opposing yet coordinated mechanisms: fusion and fission. Fusion is primarily mediated by mitofusin 1/2 (Mfn1/2) and optic atrophy protein 1 (OPA1), whereas fission depends on dynamin-related protein 1 (Drp1) and its outer mitochondrial membrane adaptors, including mitochondrial fission factor (Mff), mitochondrial fission protein 1 (Fis1), and MiD49/51 [42,43]. Areas of mitochondrial kinetic disorder are prevalent in the aging brain, characterised by reduced expression of fusion proteins such as Mfn2 and OPA1, alongside significantly elevated levels of the fission protein Drp1 (particularly its activated form p-Drp1 S616) and Fis1. This leads to excessive mitochondrial fission and fragmentation [44,45,46]. Concurrently, PUM2, which regulates Drp1 mitochondrial translocation, is upregulated during aging. This further amplifies fission abnormalities by inhibiting Mff protein synthesis [46]. This fragmented mitochondrial network exhibits impaired function, reduced membrane potential (ΔΨm), and suppressed respiratory chain activity. Consequently, synaptic energy supply is compromised, accelerating neuronal aging and cognitive decline [47,48,49,50]. Corresponding to enhanced fragmentation, aging brains exhibit markedly reduced expression of fusion proteins Mfn1, Mfn2, and OPA1. This diminishes mitochondrial network integration capacity, elevates mitochondrial and cytoplasmic ROS levels [48,51,52,53,54], and consequently impairs the maintenance and differentiation of neural stem cells (Sox2+) and newborn neurons (DCX+). This leads to impaired hippocampal neurogenesis and reduced synaptic density [54]. Consequently, the disruption of mitochondrial dynamic equilibrium represents one of the key molecular underpinnings of mitochondrial dysfunction.

Mitophagy constitutes a specialised form of autophagy responsible for selectively eliminating aged or damaged mitochondria, serving as a crucial quality control mechanism for removing impaired mitochondria. This process involves the encapsulation of damaged mitochondria within autophagosomes to form double-membrane vesicles, which subsequently fuse with lysosomes for degradation, thereby maintaining cellular energy and metabolic homeostasis [55]. This process is commonly assessed using markers such as LC3, BNIP3L/NIX, Parkin, p62, and Beclin-1 [56,57]. In the aging brain, mitophagy flux is broadly impaired. Compared with young animals, aged brains show reduced levels of Beclin-1 and LC3-II, abnormal accumulation of p62, and attenuated mitophagic activity in regions such as the hippocampal dentate gyrus [58,59,60]. Although the PINK1/Parkin pathway may be up-regulated in the aging brain, its function is severely impaired, leading to the ineffective clearance and progressive accumulation of damaged mitochondria [61,62]. Mitophagy failure may permit the accumulation of damaged mitochondria and mtDNA-driven inflammatory signaling, including cGAS–STING activation, thereby aggravating neuroinflammation [63,64]. It also compromises blood–brain barrier (BBB) integrity and diminishes cerebral vascular health [65]. Thus, the decline in mitochondrial autophagy levels within the brain during aging represents a key mechanism underpinning mitochondrial dysfunction and neuronal functional decline.

Because neurons have exceptionally high energy demands, mitochondrial energy metabolism is indispensable for maintaining membrane excitability, neurotransmitter turnover, and synaptic plasticity [66]. However, this intricate energy production system undergoes systemic decline with advancing age (Table 1), with its dysfunction recognised as a key cellular biological basis driving brain aging and cognitive decline [67]. One major manifestation is the impairment of tricarboxylic acid (TCA) cycle function. The activity or expression of several rate-limiting enzymes, including α-ketoglutarate dehydrogenase, citrate synthase, and malate dehydrogenase, is reduced in the aging brain, thereby limiting the generation of reducing equivalents required for oxidative phosphorylation [68,69]. At the same time, dysfunction of the electron transport chain, reduced mitochondrial membrane potential, and increased electron leakage collectively diminish ATP production while promoting excessive reactive oxygen species generation [66,67]. This creates a vicious cycle of oxidative damage and energetic insufficiency [70,71]. As oxidative phosphorylation becomes less efficient, aging neurons increasingly rely on glycolysis to meet basic energetic demands, reflecting a loss of metabolic flexibility [67,68]. Synapses are particularly vulnerable to this shift because of their high and spatially restricted energy requirements. Local ATP deficiency at axon terminals can impair neurotransmitter release, postsynaptic responsiveness, and synaptic plasticity, ultimately contributing to deficits in learning and memory [69]. Recent evidence further suggests that the coupling between neuronal activity and mitochondrial gene transcription becomes attenuated with age. In young brains, neuronal activity can rapidly stimulate mitochondrial transcriptional responses to match local energy demands, whereas this adaptive mechanism is weakened in aging [22]. This age-related disruption of activity-dependent mitochondrial regulation may represent an additional mechanism linking mitochondrial dysfunction to cognitive decline.

Thus, the energy crisis in brain aging constitutes a multi-layered, systemic failure originating from the decline in core enzyme function within the TCA cycle and oxidative phosphorylation. This leads to insufficient ATP synthesis, excessive ROS production, and ultimately triggers synaptic energy starvation alongside the failure of the “on-demand energy supply” mechanism. Reversing this process requires not only supportive strategies such as supplementing metabolic substrates and employing antioxidants but also focusing on reactivating and enhancing mitochondrial function and adaptability through physical activity (such as exercise) and cognitive engagement. This approach rebuilds a stable and efficient energy supply system for the aging brain.

Exercise serves as a potent stimulus for activating mitochondrial biogenesis in the brain (Table 2). Animal studies indicate that prolonged exercise significantly upregulates the expression of SIRT1, PGC-1α, NRF1/2, and TFAM in brain regions such as the hippocampus [79,80,81]. Activation of PGC-1α may further promote the synthesis of brain-derived neurotrophic factor (BDNF), which plays a pivotal role in neural plasticity and neuronal survival [82,83,84]. Through the core signaling axis of AMPK/SIRT1/PGC-1α/BDNF, experimental studies suggest that exercise enhances mitochondrial biogenesis, replenishing the aging brain with functionally competent mitochondria and elevating energy metabolism potential [31,85].

Among diverse strategies countering cerebral aging, exercise training demonstrates irreplaceable advantages, with one core mechanism being its capacity to efficiently activate mitochondrial biogenesis within the brain. For neurons with exceptionally high energy demands, maintaining the vigour of this process underpins their synaptic plasticity, signal transduction, and long-term survival. The age-related decline in mitochondrial biogenesis constitutes a pivotal factor in causing neuronal energy insufficiency and diminished metabolic plasticity (i.e., the cell’s capacity to adapt to fluctuations in energy requirements) [31]. Extensive research indicates that regular exercise represents one of the most effective non-pharmacological interventions for reversing this age-related decline, primarily through the integrated signaling axis of energy sensing–transcriptional regulation–neurotrophic support.

Exercise-induced alterations in cellular energy status (such as elevated AMP/ATP ratios) primarily activate the energy sensor AMP-activated protein kinase (AMPK). Activated AMPK directly phosphorylates or enhances NAD⁺ levels to activate the deacetylase SIRT1, thereby jointly regulating the master regulatory factor peroxisome proliferator-activated receptor gamma coactivator-1α (PGC-1α) [31]. The deacetylation and activation of PGC-1α are crucial steps for its functional engagement. It translocates to the nucleus, where it cooperates with transcription factors such as nuclear respiratory factor 1/2 (NRF1/2), ultimately upregulating the expression of mitochondrial transcription factor A (TFAM). TFAM enters mitochondria, where it governs mtDNA replication, maintenance, and transcription, thereby driving the synthesis of new mitochondria [29,30]. Crucially, exercise-induced PGC-1α activation also directly promotes brain-derived neurotrophic factor (BDNF) synthesis [83]. BDNF serves not only as a key mediator of synaptic plasticity and neuronal survival, but its signaling feedback further supports mitochondrial biogenesis and function, forming a positive feedback loop [82]. Thus, through the AMPK/SIRT1/PGC-1α/BDNF axis, exercise achieves comprehensive remodelling spanning energy sensing, gene expression, functional structures (new mitochondria), and supportive environments (neurotrophic factors). This provides the aging brain with a robust mitochondrial reserve, excelling in both quality and quantity, fundamentally enhancing its metabolic resilience and stress resistance.

In summary, exercise powerfully drives mitochondrial biogenesis in the aging brain by activating the highly integrated AMPK/SIRT1/PGC-1α/BDNF signaling network. This process not only directly increases the number of functionally competent mitochondria, compensating for age-related depletion, but more importantly optimises the neuronal survival and plasticity microenvironment by elevating BDNF levels. Ultimately, exercise intervention endows the brain with enhanced metabolic resilience—the capacity to mobilise and utilise energy more efficiently when confronted with increased energy demands or metabolic stress. This establishes a robust cellular energetics foundation for delaying cognitive decline and sustaining brain health.

In the aging brain, both core pillars of the mitochondrial quality control system—mitochondrial dynamics (the balance between fusion and fission) and mitochondrial autophagy (selective clearance of damaged mitochondria)—exhibit significant dysfunction. Dysregulated dynamics lead to mitochondrial network fragmentation and functional loss, while impaired autophagic flux causes accumulation of dysfunctional mitochondria. Together, these exacerbate oxidative stress and energy crises, driving neuronal degeneration [6,44,55]. Experimental studies suggest that exercise intervention may partially reverse these alterations (Table 3). As a potent physiological stimulus, it synergistically reconfigures the dynamic equilibrium and renewal capacity of the mitochondrial network by activating multiple interconnected signaling pathways, thereby preserving neuronal metabolic health and cognitive function [10].

Mitochondrial morphology and function are highly dependent on the dynamic equilibrium between fusion and fission. Fusion (mediated by Mfn1/2 and OPA1) facilitates exchange of contents, compensates for defects, and maintains network integrity and membrane potential; fission (mediated by Drp1 and its receptors) promotes the isolation, distribution, and subsequent clearance of damaged segments [42,43]. Aging brains commonly exhibit excessive fission (increased Drp1 activity) coupled with insufficient fusion (reduced expression of Mfn1/2, etc.), leading to the accumulation of numerous small, dysfunctional mitochondrial fragments [45,46].

Exercise represents a key non-pharmacological means of restoring this equilibrium. One core mechanism involves activating the cyclic adenosine monophosphate (cAMP)/protein kinase A (PKA) signaling pathway. Exercise may restrain excessive mitochondrial fission partly through cAMP/PKA-dependent modulation of Drp1 phosphorylation, although direct evidence in the aging brain remains limited [86]. This phosphorylation inhibits Drp1 recruitment to mitochondria and its GTPase activity, effectively suppressing excessive mitochondrial fission [87,88]. Concurrently, exercise also upregulates the expression of mitochondrial fusion proteins, such as mitochondrial fusion protein 2 (Mfn2) [89]. This dual regulation—inhibiting fission while promoting fusion—collectively drives the transformation of mitochondria in aging neurons from a fragmented state towards a healthier, interconnected network morphology. This facilitates the stabilisation of mitochondrial membrane potential, optimises oxidative phosphorylation efficiency, and provides a sustained, stable energy supply to synapses [50].

Mitochondrial autophagy represents the terminal quality control step for eliminating dysfunctional mitochondria arising from kinetic imbalances or accumulated damage. The age-associated decline in mitochondrial autophagy correlates with defects across multiple stages of the process, including initiation (e.g., reduced expression of Beclin-1 and LC3-II) and execution (e.g., impaired PINK1/Parkin pathway function and p62 accumulation) [59,61]. Exercise may enhance this mitochondrial “scavenger” function.

The energy expenditure effect of exercise activates the cellular energy sensor AMPK. Activated AMPK initiates autophagy through a dual mechanism: firstly, it directly phosphorylates and activates key components within the autophagy initiation complex; secondly, and more critically, it inhibits the activity of the mammalian target of rapamycin (mTOR) [90]. mTOR serves as the primary regulator of cellular growth and anabolism, whilst also acting as a potent inhibitor of autophagy. Exercise, via AMPK-mediated mTOR inhibition, releases this braking effect on autophagy [91]. This process is particularly evident in the suppression of activity in mTOR’s downstream effector, the ribosomal protein S6 kinase 1 (S6K1) [92]. Furthermore, exercise upregulates the expression of receptor proteins specifically mediating mitochondrial autophagy, such as BNIP3L and Parkin [93,94]. Collectively, these alterations facilitate the recognition and engulfment of damaged mitochondria by autophagosomes, culminating in their degradation through fusion with lysosomes. This restores normal mitochondrial autophagy flux, thereby preventing the vicious cycle of toxic protein accumulation and oxidative stress.

Thus, exercise remodels the mitochondrial quality control system not through a single pathway, but via a synergistic network. On one hand, it precisely regulates Drp1 activity via the cAMP/PKA signaling axis, inhibiting excessive fission while promoting fusion protein expression to restore a healthy mitochondrial network architecture. On the other hand, it releases autophagy inhibition through the AMPK/mTOR signaling axis and upregulates specific autophagy receptors, thereby accelerating the recognition and clearance of damaged mitochondria. These two processes complement each other: healthy dynamics provide autophagy with correct targets for clearance (i.e., damaged segments isolated by moderate fission), while efficient autophagy clears damaged mitochondrial components for the dynamic process, allowing the establishment of a newly fused, healthy mitochondrial network. This systemic optimisation of the mitochondrial “morphology–function–renewal” cycle constitutes the indispensable cellular and molecular basis for exercise’s capacity to enhance energy metabolism and synaptic function in the aging brain, ultimately delaying cognitive decline.

The neuroprotective effects of exercise extend far beyond direct stimulation of the brain. A growing body of evidence indicates that regular physical activity induces robust systemic adaptations, indirectly and persistently supporting brain health and cognitive function through a dynamic bidirectional communication system termed the “muscle–brain axis” [97,98]. In this process, skeletal muscle during exercise functions as a crucial endocrine organ, releasing a series of factors into the circulatory system. These factors either directly cross the blood–brain barrier or transmit signals via receptors on the barrier surface, thereby promoting neurogenesis, enhancing synaptic plasticity, and optimising mitochondrial function in the distant brain (Table 4).

Moderate-intensity continuous training (e.g., brisk walking, jogging, swimming) represents the most thoroughly researched and safest foundational programme. It has been shown to improve cardiovascular health, increase cerebral blood flow, and enhance mitochondrial function while promoting neuroplasticity through multiple pathways including AMPK/PGC-1α and BDNF [103,104]. Its core mechanism lies in the sustained, stable enhancement of cardiovascular function and cerebral blood perfusion, thereby supplying the brain with ample oxygen and nutritional substrates [7]. At the molecular level, it effectively activates the AMPK/PGC-1α signaling axis, promotes mitochondrial biogenesis, and upregulates brain-derived neurotrophic factor (BDNF) levels, consequently enhancing synaptic plasticity and neuronal survival [7]. A recent study further suggested that moderate-intensity aerobic training may reduce neuronal apoptosis and improve learning and memory, potentially through miR-21-3p-related mechanisms [97]. A preliminary 2024 study comparing exercise patterns among community-dwelling older adults found that six months of moderate-intensity aerobic exercise significantly improved executive function, an effect correlated with increased surface area in the right dorsolateral prefrontal cortex [105].

High-Intensity Interval Training (HIIT) is characterised by its time efficiency and profound disruption of the metabolic system. HIIT rapidly induces substantial lactate production in skeletal muscle. Lactate serves not only as an energy substrate but also as a crucial signaling molecule capable of traversing the blood–brain barrier to directly influence neurons. It may promote neuroplasticity and angiogenesis by activating pathways such as PGC-1α, potentially offering unique advantages in improving executive function and cerebral metabolism in older individuals [97,106]. A recent animal study proposed a potentially novel cross-organ mechanism underlying the neuroprotective effects of HIIT, suggesting that mitochondria originating from exercised skeletal muscle may be transported via circulating platelets to the brain, where they could contribute to white matter repair and functional recovery following cerebral ischemia [107]. However, this hypothesis is currently based on a single mouse study, and it remains unclear whether similar mechanisms occur under physiological aging conditions or in humans. Therefore, the concept of exercise-induced “mitochondrial migration” should be regarded as preliminary and requires further experimental and translational validation.

The contribution of resistance training (strength training) to brain health is frequently underestimated. Its most direct effect lies in counteracting age-related muscle loss (sarcopenia). Muscle tissue functions not only as a motor organ but also as a significant endocrine organ. Resistance training significantly increases muscle mass and strength while elevating circulating insulin-like growth factor-1 (IGF-1) levels. IGF-1 crosses the blood–brain barrier, activating the IGF-1R/PI3K/Akt survival signaling pathway within neurons. This pathway is crucial for sustaining neuronal survival and inhibiting apoptosis [98,108,109,110]. Moreover, greater muscle mass correlates with elevated basal metabolic rates and improved systemic glucose-lipid metabolism regulation, indirectly fostering a healthier systemic environment for the brain [110]. Thus, resistance training constitutes a pivotal peripheral support mechanism for central nervous system health. Nevertheless, evidence regarding the direct effects of resistance training on brain mitochondrial quality control remains relatively limited compared with aerobic exercise.

Mind–body practices such as Tai Chi and yoga, alongside combined training protocols integrating aerobic, resistance, and balance exercises, represent more holistic intervention approaches. These approaches not only confer combined physiological benefits but also improve psychological well-being through enhanced mind–body awareness, breathing regulation, and mindfulness [111,112]. This effectively reduces chronic stress hormone levels (such as cortisol), as prolonged exposure to elevated cortisol exerts clear neurotoxic effects on hippocampal neurons and mitochondria [113,114,115]. By alleviating stress, improving sleep, and elevating mood, mind–body exercises cultivate an optimised internal microenvironment conducive to neuroplasticity and neuronal health. Their mechanisms may involve coordinated physiological and psychological adaptations. However, empirical research in this area remains relatively scarce and warrants further strengthening [116]. Moreover, the molecular mechanisms underlying these benefits remain insufficiently characterised.

Nevertheless, findings regarding the optimal exercise modality and intensity for delaying brain aging remain inconsistent. While moderate-intensity aerobic exercise is generally considered safe and effective, some studies suggest that high-intensity exercise may induce excessive oxidative stress or inflammatory responses in vulnerable older individuals [117]. In contrast, other studies report superior metabolic and cognitive benefits following high-intensity interval training [97,106]. These discrepancies may be related to differences in age, sex, genetic background, baseline fitness, and disease status. Furthermore, variations in exercise protocols and outcome measures across studies complicate direct comparisons. In addition, some individuals may exhibit limited responsiveness to exercise interventions (exercise non-responders) [118], although the mechanisms underlying this phenomenon in brain aging remain poorly understood. Therefore, a “one-size-fits-all” exercise prescription for brain aging is unlikely to be appropriate. Taken together, these findings highlight the importance of developing personalised and adaptable exercise strategies based on individual physiological characteristics and aging trajectories.

No new data were created or analyzed in this study. Data sharing is not applicable to this article.