Psychiatric and neurological disorders represent a major and growing global health challenge, accounting for a substantial proportion of disability, reduced quality of life (QoL), and increased healthcare expenditure worldwide [1–3]. Despite decades of pharmacological innovation, current treatments may remain insufficient for many patients, with high rates of partial response, relapse, and treatment resistance. These limitations reflect the complex, multisystem nature of brain disorders and highlight the need for interventions capable of targeting multiple biological and functional domains simultaneously.

In this context, exercise is increasingly recognized as a potentially effective and still underused intervention for brain health [4–6]. Accumulating evidence supports its role not only as an adjunctive treatment, but also as a preventive and potential disease-modifying strategy [7–10] across psychiatric and neurological conditions. Unlike pharmacological approaches that typically target single molecular pathways, exercise induces broad, system-wide adaptations. It modulates neuroplasticity, neuroinflammation, mitochondrial function, cerebrovascular integrity, stress physiology, and immune–metabolic signaling pathways involved in both psychiatric and neurological disorders [11–13].

Clinically, exercise interventions have demonstrated efficacy in improving symptoms and functional outcomes in major depressive disorder (MDD), anxiety disorders, bipolar disorder (BD), schizophrenia, and post-traumatic stress disorder (PTSD). Parallel evidence in neurological diseases, including Alzheimer’s disease (AD) and other dementias, Parkinson’s disease (PD), multiple sclerosis (MS), stroke, and epilepsy, indicates potential beneficial effects on cognition, motor function, fatigue, and overall QoL. Exercise is scalable, cost-effective, and adaptable to individual capacities. It may therefore serve as a valuable intervention for long-term prevention and management, rather than simply a lifestyle add-on.

The relevance of exercise for brain health has been further amplified by the emergence of long COVID [14, 15], a condition characterized by persistent neuropsychiatric and neurological symptoms such as cognitive impairment, fatigue, mood disturbances, and autonomic dysfunction. These manifestations share overlapping biological mechanisms with established brain disorders, including chronic inflammation, endothelial dysfunction, dysregulated stress responses, and impaired energy metabolism [16]. While exercise represents a promising therapeutic avenue in long COVID, inappropriate prescription may exacerbate symptoms [17, 18]. This underscores the need for mechanistically informed, individualized approaches.

Despite rapidly expanding evidence, the literature on exercise in psychiatric and neurological disorders remains fragmented across disciplines, limiting translation into clinical practice. An integrated framework that bridges mechanistic insights with clinical evidence is urgently needed. This narrative review aims to synthesize current knowledge on exercise as medicine in long COVID patients with psychiatric and neurological diseases, highlighting shared mechanisms, clinical efficacy, and key challenges for implementation. By doing so, it seeks to reposition exercise from an optional lifestyle recommendation to a core component of brain healthcare.

This study was conducted as a narrative review aiming to provide an integrative and clinically oriented synthesis of the role of exercise in long COVID patients with psychiatric and neurological disorders. The objective was to synthesize mechanistic, clinical, and rehabilitation evidence rather than to perform a systematic or umbrella review.

Relevant literature was identified through targeted searches in PubMed and Scopus for studies published up to January 2026, using combinations of keywords including “long COVID”, “post-acute sequelae of SARS-CoV-2”, “exercise”, “physical activity”, “psychiatric disorders”, “neurological disorders”, and “rehabilitation”. Additional articles were identified through manual screening of reference lists of key publications. Studies were included if they were original investigations, including randomized controlled trials (RCTs), pragmatic trials, cohort studies, and mechanistic human studies, or systematic reviews (SRs) and meta-analyses (MAs) addressing exercise interventions relevant to psychiatric or neurological outcomes. Eligible studies involved either long COVID or post-COVID conditions, or established psychiatric and neurological disorders considered relevant for mechanistic or clinical extrapolation. Only studies evaluating exercise or structured physical activity as a primary intervention or exposure and reporting at least one relevant clinical, cognitive, neurological, psychiatric, or functional outcome were considered. Animal studies, non-peer-reviewed reports, editorials, and commentaries without original data, as well as studies not involving exercise interventions or not relevant to brain-related outcomes, were excluded.

Priority was given to high-quality evidence, including MAs, SRs, RCTs, and key mechanistic studies. When direct evidence in long COVID populations was limited, findings from non-COVID psychiatric and neurological populations were included and explicitly identified as extrapolated. Importantly, all statements derived from non-COVID evidence were clearly indicated throughout the manuscript to distinguish them from findings directly obtained in long COVID populations.

No formal systematic search strategy, predefined study selection protocol, or formal quality assessment tool was applied, which is consistent with the narrative nature of the review. Study inclusion was guided by relevance to the scope of the manuscript, methodological rigor, and contribution to the conceptual framework.

This approach allows a flexible and hypothesis-generating synthesis of a rapidly evolving and multidisciplinary field.

Long COVID, also referred to as post-acute sequelae of SARS-CoV-2 infection (PASC), is defined by the World Health Organization as the persistence of symptoms occurring during or after acute COVID-19 that last for at least two months, extend beyond three months from the initial infection, and cannot be explained by an alternative diagnosis [19]. Definitions from major health bodies vary in minimum duration (e.g., 4–12 weeks), but all emphasize persistent, multisystem symptoms following confirmed or probable SARS‑CoV‑2 infection [20].

Long COVID is a heterogeneous, multisystem condition that may present with both neuropsychiatric and internal medicine manifestations, often overlapping and fluctuating over time [21]. From an internal medicine perspective, it has been associated with persistent cardiopulmonary symptoms (e.g., dyspnea, chest pain, and palpitations), reduced exercise capacity, and post-exertional symptom exacerbation [22]. Cardiovascular sequelae may include myocarditis, pericarditis, microvascular dysfunction, and arrhythmias [23], while pulmonary involvement may manifest as impaired diffusion capacity or residual interstitial changes [24]. Endocrine and metabolic alterations (e.g., new-onset diabetes or thyroid dysfunction), gastrointestinal symptoms, chronic inflammation, and coagulation abnormalities have also been reported [25–29].

Common neuropsychiatric and neurological manifestations include cognitive impairment (“brain fog”), fatigue, sleep disturbances, mood changes, autonomic dysfunction, headache, and exercise intolerance [30–32]. Diagnosis is primarily clinical, supported by targeted investigations such as blood tests, neuroimaging, neurocognitive testing, and autonomic assessments [33–38]. Routine biomarkers lack diagnostic specificity, reinforcing the importance of clinical and multidisciplinary evaluation [39].

Management is multidisciplinary and focuses on symptomatic relief and functional recovery. Evidence supports structured exercise, cognitive rehabilitation, sleep optimization, and psychological interventions, while pharmacological treatments remain largely supportive [40].

The overlap of pathophysiological mechanisms between long COVID and other psychiatric and neurological disorders, including chronic inflammation, endothelial dysfunction, mitochondrial impairment, and dysregulation of the hypothalamic–pituitary–adrenal (HPA) axis, suggests that interventions targeting these systemic processes may confer broad benefits, although this remains a mechanistic hypothesis. However, among these interventions, structured physical activity and exercise appear particularly promising, given their ability to simultaneously modulate immune, vascular, metabolic, and neural pathways. Understanding these mechanisms is essential to inform multidisciplinary rehabilitation strategies aimed at alleviating the cognitive, emotional, and physical sequelae of long COVID. This mechanistic insight is particularly important before considering the clinical implications of exercise in individuals with neurological and psychiatric disorders, which will be addressed in the following sections.

Exercise is associated with a range of biological adaptations relevant to brain health through a network of interconnected biological mechanisms that operate across molecular, cellular, and systemic levels. Rather than acting via a single pathway, physical activity induces coordinated adaptations across neural, immune, vascular, metabolic, and endocrine systems, as illustrated in Figure 1, providing a plausible biological rationale for its efficacy in both psychiatric and neurological disorders.

In long COVID patients with pre-existing or newly developed neurological disorders, exercise may represent a potentially valuable adjunctive intervention to support cognitive, motor, and functional recovery [98–100]. High-quality RCTs specifically targeting well-defined neurological diagnoses in long COVID remain scarce; however, indirect evidence from neurological rehabilitation and emerging post-COVID studies suggests clinically meaningful benefits when exercise is appropriately tailored [96, 101]. A summary of the available evidence is reported in Table 3.

Long COVID may exacerbate pre-existing neurological disorders or unmask latent vulnerabilities, leading to worsening of motor symptoms, cognitive decline, increased fatigue, autonomic instability, and reduced functional reserve [32, 102]. Taquet et al. [102] analyzed over 1.28 million patients and showed that SARS-CoV-2 infection is associated with a long-term increased risk of neurological and psychiatric disorders compared with other respiratory infections. While mood and anxiety disorders tend to normalize over time, risks of cognitive impairment, dementia, epilepsy, and seizures remain elevated for up to two years, indicating persistent neurological sequelae.

In patients with neurodegenerative diseases, MS, epilepsy, or prior cerebrovascular events, persistent systemic inflammation, immune dysregulation, and autonomic impairment associated with post-COVID condition may contribute to amplifying baseline neurological deficits and accelerating functional deconditioning.

Also in neurological populations, the effects of exercise depend on modality, intensity, frequency, and duration. Moderate-intensity aerobic exercise is generally well-tolerated and consistently associated with functional benefits, whereas higher-intensity protocols may induce greater neuroplastic and neurotrophic adaptations but are often less well-tolerated, particularly in individuals with post-exertional symptom exacerbation. Resistance and neuromotor training may further improve muscle strength, balance, and functional independence.

In long COVID individuals reporting persistent cognitive dysfunction (“brain fog”), moderate-intensity aerobic exercise delivered with gradual progression may contribute to improvements in attention, processing speed, and working memory [14, 40, 103], although findings remain inconsistent and not reproducible. For example, a recent RCT [103] evaluated an 8-week cardiopulmonary rehabilitation program in individuals with long COVID (n = 40; rehabilitation group vs. control). The intervention consisted of an individualized program delivered three times per week, combining light-to-moderate aerobic exercise, resistance training, and respiratory exercises. Although the rehabilitation group showed significant reductions in perceived stress and depressive symptoms compared with controls, no significant between-group differences were observed in neuropsychological test performance. Furthermore, a recent SR including 9 studies and approximately 672 patients with long COVID reported improvements in fatigue and dyspnea following structured exercise interventions. However, the findings suggest limited exercise-related effects on cognitive performance. In addition, cognitive fatigability and symptom fluctuation require careful pacing strategies, shorter sessions when needed, and close monitoring to avoid post-exertional worsening [104–106].

Based largely on evidence from non-COVID literature, evidence from stroke rehabilitation trials indicates that repetitive, task-oriented motor training combined with low-to-moderate intensity aerobic exercise (generally 3–5 sessions per week over 8–12 weeks) can improve multiple motor and non-motor endpoints, including gait speed, functional mobility, and activities of daily living [107, 108]. These findings are further supported by SR evidence, which reports consistent improvements in upper extremity function following task-oriented rehabilitation approaches [109]. These findings support the potential role of combining aerobic conditioning with task-specific training to enhance functional recovery, although their applicability to long COVID populations remains to be established. In this context, programs should be individualized and symptom-contingent, taking into account fatigue, autonomic instability, and fluctuating tolerance frequently observed in long COVID [38, 110, 111].

Similarly, in individuals with long COVID who experience peripheral neuropathic symptoms, balance impairment, or sensory disturbances, neuromotor and strength-based interventions may enhance proprioception, reduce fall risk, and support overall physical autonomy [98, 112, 113], although direct evidence in long COVID populations remains limited. Autonomic dysfunction, including orthostatic intolerance and exercise intolerance, is common in long COVID and necessitates specific adaptations [110]. Recumbent or semi-recumbent aerobic modalities, low-impact strengthening, and gradual verticalization protocols are often better tolerated in the early phases. Progression should be individualized, based on daily functional capacity and symptom variability.

In patients with epilepsy, demyelinating disorders, or significant neurological instability, exercise programs should be supervised and coordinated with neurological care to ensure safety and optimize adherence. As reported in some reviews and MAs [114–117], structured, supported interventions have been associated with better functional outcomes compared with unsupervised activity.

Overall, available evidence from neurological rehabilitation studies suggests that structured exercise interventions of approximately 6–24 weeks, depending on condition severity and modality, can produce clinically meaningful improvements in motor and cognitive outcomes. However, direct evidence specifically in long COVID neurological populations remains limited and heterogeneous and further disease-specific studies are needed to confirm these preliminary observations.

Exercise may represent a multimodal, mechanistically grounded intervention for both psychiatric and neurological sequelae of long COVID. Its effects span cognitive, motor, emotional, and autonomic domains. These broad system-wide adaptations position exercise as a potential adjunctive strategy that complements pharmacological and rehabilitative approaches.

In individuals with pre-existing or newly developed psychiatric or neurological disorders, long COVID may exacerbate symptoms, unmask latent vulnerabilities, and reduce functional reserve. Evidence, though still limited in long COVID populations, indicates that structured, individualized, and supervised exercise programs may improve cognition, mobility, mood, endurance, and QoL. The greatest benefits are observed when interventions are tailored to symptom severity, comorbidities, and daily functional capacity, with careful pacing to avoid post-exertional symptom exacerbation.

Overall, while mechanistic and indirect clinical evidence is relatively strong, the lack of high-quality, disease-specific RCTs in long COVID remains a major limitation. Future research should focus on disease-specific protocols, optimal intensity and frequency, long-term adherence, and mechanistic correlations to further refine recommendations and facilitate clinical translation.