Two-year patterns of thinking decline and brain-related problems in COVID-19 survivors with early neurological symptoms

The COVID-19 pandemic caused by SARS-CoV-2 has profound implications for global public health. Recently, COVID-19 has caused approximately 780 million infections and over 7 million deaths (World Health Organization, 2025). Initially characterized by respiratory symptoms, its impact on multiple organ systems—especially the nervous system which was first reported by our team (Mao et al., 2020)—has become increasingly evident (Yang et al., 2021; Xie et al., 2022). These patients experienced neurological symptoms during the acute phase of infection and beyond, range from headache, dizziness, and anosmia to more severe conditions such as encephalitis and acute vascular events (Mao et al., 2020; Cho et al., 2023; Giussani et al., 2024).

Some symptoms that occur during the acute infection, along with newly emerged symptoms, may persist for a long time and have attracted a great deal of attention. Together, these were collectively referred to as post-acute sequelae of SARS-CoV-2 infection (PASC), also known as long COVID. Cognitive dysfunction—particularly manifestations like “brain fog” involving memory deficits, attention impairment, executive dysfunction, and slowed processing speed—is a predominant feature across diverse patient populations (McWhirter et al., 2023; Zhao et al., 2023). Upto 10.7–19.1% of COVID-19 survivors reported subjective cognitive decline after a 1–2-year follow-up (Liu et al., 2022; Araujo et al., 2025). Identified risk factors included age, severity of acute phase infection, dysregulation of glucose and lipid metabolism. The underlying mechanisms are thought to involve direct viral invasion of the central nervous system (CNS), systemic inflammation, and microvascular damage (Zubair et al., 2020). These findings suggest that neurological symptoms during acute SARS-CoV-2 infection may contribute to subsequent cognitive decline—particularly long-term impairment—and may serve as potential prognostic indicators. Although some studies have reported long-term follow-up results of COVID-19 patients (Giussani et al., 2024; Zhang et al., 2024), evidence remains limited regarding the specific COVID-19-related neurological manifestations and risk factors strongly associated with cognitive impairment. In addition, the long-term persistence of neurological symptoms following COVID-19, as well as the emergence of new-onset symptoms in COVID-19 survivors after recovery, warrant serious attention.

This study conducted a 2-year longitudinal follow-up of COVID-19 patients to explore the associations between acute-phase neurological symptoms and subsequent cognitive changes and neurological sequelae. By analyzing the epidemiological patterns of COVID-19-related neurological symptoms and their cognitive functional impacts, this research aims to identify key risk factors of COVID-19 infection associated with long-term neurological sequelae.

Although previous studies have described the prevalence of cognitive impairment following COVID-19, our study systematically evaluated a broad spectrum of acute neurological manifestations—including both central and peripheral nervous system symptoms. Their associations with cognitive decline and impairment were quantified over a 2-year period, with propensity score matching applied to minimize potential confounding. It was demonstrated that acute CNS manifestations (e.g., delirium and brain fog) remained independent risk factors for cognitive decline (aOR 2.16) and cognitive impairment (aOR 2.75), even after adjustment for comorbidities. Furthermore, although the majority of acute symptoms resolved over time, relatively high residual rates of fatigue (8.66%) and brain fog (5.99%) were observed at 2-years’ follow-up. Although most new-onset symptoms after SARS-CoV-2 infection were relatively rare (< 5%), the proportion of insomnia, tinnitus, blurred vision, movement disorder, palpitation, muscle weakness, and respiratory symptoms were much higher in the neurological symptom group, with a highest proportion in insomnia (9.99%), compared with the non-neurological symptom group. Collectively, these results underscore the prognostic value of acute CNS symptoms for targeted intervention strategies and address a critical gap in evidence regarding the need for sustained long-term monitoring of COVID-19 patients with neurological involvement.

Our findings add to the growing evidence that SARS-CoV-2 infection is a significant risk factor for cognitive impairment. A previous study reported that the prevalence of cognitive impairment was 19.1% among hospitalized COVID-19 patients—nearly twice that of non-hospitalized patients (10.7%) and significantly higher than that observed in uninfected controls (3.2%) (Araujo et al., 2025). Community-based cohorts also revealed that individuals hospitalized with COVID-19 demonstrated poorer performance in executive function and memory compared with non-hospitalized counterparts (Demmer et al., 2025). Longitudinal analyses from the UK PROTECT study further indicated that COVID-19 infection was identified as one of the major contributing factors to accelerate cognitive decline in older adults during the pandemic (Corbett et al., 2023). Among patients with long COVID, 27% exhibited global cognitive impairment, with particularly pronounced deficits in attention, working memory, processing speed, and verbal fluency (Serrano Del Pueblo et al., 2024).

Neurological symptoms during acute SARS-CoV-2 infection serve as potential prognostic indicators of subsequent cognitive decline—particularly long-term impairment. For example, previous studies have shown that delirium during hospitalization for COVID-19 was associated with functional impairment and cognitive decline within 6 months after discharge (OR, 2.48; 95% CI, 1.38–4.82) (Kaushik et al., 2024). One cross-sectional study reported that patients who experienced neurological symptoms such as headache, anosmia, or dysgeusia during infection performed worse on cognitive testing 10–34 days after discharge (Almeria et al., 2023). However, these studies did not specifically examine the association between multiple neurological symptoms and long-term cognition. Our study not only assessed the overall relationship between the presence of neurological symptoms and cognitive outcomes but also further disentangled the impact of individual symptoms on long-term cognitive trajectories. Also, individual neurological symptoms during acute infection—including delirium, brain fog, headache, stroke, numbness, and facial paralysis—were all identified as risk factors for cognitive decline at 2 years post-infection, may be linked to that SARS-CoV-2 has already exerted neuroinvasive effects. This suggests that the occurrence of these symptoms during acute COVID-19 infection warrants particular attention from neurologists, with timely cognitive assessments to enable early detection of cognitive changes and to guide appropriate intervention strategies.

It is important that in our findings, the CNS-symptoms—brain fog—were not only significantly associated with cognitive decline and cognitive impairment at 2-years’ follow-up, but also still maintained at a high level. Characterized by ongoing subjective cognitive complaints, brain fog can persist long after the initial illness has resolved, frequently impairing patients’ daily functioning and overall wellbeing (Hampshire et al., 2024; Qin et al., 2024). Its underlying mechanisms may be linked to persistent neuroinflammation, immune system activation, and disruption of the blood-brain barrier following COVID-19 infection, all of which can accelerate cognitive deterioration (Greene et al., 2024).

Systematic reviews reported that neurological symptoms are common sequelae of COVID-19, often persisting over extended trajectories and exerting substantial impact on quality of life (Giussani et al., 2024). A prospective cohort study from Germany reported that neurological symptoms alone accounted for as much as 61.5% of long COVID sequelae (Bahmer et al., 2022). Another bidirectional, longitudinal cohort study demonstrated that 55% of patients continued to exhibit at least one sequela at 2-year follow-up (Huang et al., 2022), a prevalence inconsistent with our findings. However, in our cohort, the proportion of sequelae was much lower at 2-year follow-up. This discrepancy likely stems from differences in viral pathogenicity. Our study focused exclusively on Omicron cases (December 2022–March 2023), contrasting with UK findings on the original strain, suggesting that Omicron is associated with a lower risk of long-term sequelae.

It is interesting that some patients who experienced neurological symptoms during acute SARS-CoV-2 infection continued to exhibit sequelae throughout the 2-year follow-up period, warranting further attention. Then, a further comparative analysis between patients with persistent symptoms and those without revealed that a higher proportion of women and fewer vaccinations are the risk factors rather than comorbidities. This is consistent with a previous case-control study comparing long COVID patients with predominant neurological symptoms to fully recovered individuals, in which a markedly higher proportion of women in the long COVID group (71% vs. 36%) was reported (Serrano Del Pueblo et al., 2024). The overrepresentation of women in the persistent-symptom group raises the possibility of sex-specific biological mechanisms, such as differences in immune responses, hormonal regulation, or susceptibility to neuroinflammatory processes following viral infection. The protective effect observed with higher vaccination doses may reflect reduced viral persistence or attenuation of immune-mediated damage. Alternatively, vaccination could mitigate the severity of acute infection, thereby lowering the risk of subsequent neurological sequelae. These observations underscore the importance of close monitoring in vulnerable subgroups and highlight the need for further research into underlying mechanisms, as well as validation in larger and more diverse cohorts.

Moreover, new-onset symptoms following SARS-CoV-2 infection is a key point worthy of particular attention. A previous Chinese cohort study of 125 patients reported that 26.45% and 9.92% developed insomnia and depressive symptoms, respectively, shortly after recovery, but it is limited by small sample size and assessing only immediate post-recovery outcomes (Xu et al., 2022). In our study, we found patients with acute-phase neurological symptoms were significantly more likely to develop new neurological symptoms at 2-year follow-up (40.84%), and depression, insomnia, and stroke are three most common newly developed neuropsychiatric symptoms. In addition, our findings revealed distinct risk profiles for new-onset sequelae: (1) patients with new stroke were generally older, predominantly male, and frequently had comorbid hypertension and hyperlipidemia, consistent with traditional vascular risk factors; (2) new-onset depression was more common among patients with hyperlipidemia and lower educational attainment, suggesting an interplay of metabolic and psychosocial determinants; (3) new-onset insomnia was more frequent among older patients, women, and those with fewer vaccine doses, implying that demographic characteristics and immunization status may influence susceptibility to sleep disturbances. These results highlight the diverse manifestations of long COVID and necessitate symptom-specific risk stratification. Future research should prioritize elucidating the mechanisms underlying these multi-system interactions.

Nonetheless, our study has several limitations. First, the epidemiological context during our enrollment period (December 2022–March 2023) coincided with a nationwide Omicron wave in China, during which universal testing policies revealed exceptionally high infection rates, leaving few truly uninfected individuals available for recruitment as controls. Consequently, cognitive changes in healthy individuals over the same 2-year period could not be directly assessed and could only be inferred from previous cohort studies. Second, the IQCODE relies on retrospective informant reports, which may introduce recall bias. Relatives’ or caregivers’ judgments of cognitive decline could be influenced by their emotional state, illness perceptions, or media exposure. Thus, the inherent subjectivity of this instrument should be considered when interpreting our findings. Third, despite the use of propensity score matching, multivariate logistic regression and subgroup analysis to control confounding factors, residual confounding from unmeasured variables (e.g., genetic susceptibility and socioeconomic status) may persist and potentially bias our estimates. Forth, our study focused on hospitalized Omicron patients during a specific epidemic wave. The generalizability of our findings to pre-Omicron variants or other geographic regions and healthcare systems requires further investigation through larger, multicenter studies with diverse populations and viral variants. Finally, given the close relationship between cognition and aging, extended follow-up over the next 5–10 years is essential to fully capture long-term outcomes.

Acute neurological symptoms, particularly CNS symptoms, like delirium and brain fog, are significantly associated with long-term cognitive decline in COVID-19 survivors. Patients with these symptoms are at higher risk for either ongoing cognitive impairment or decline over time and more likely to have a higher incidence of new-onset symptoms at the 2-year follow-up, highlighting the importance of early intervention and long-term monitoring in this subgroup.

The author(s) declared that financial support was received for this work and/or its publication. This work was supported by the National Key Research and Development Program of China (No. 2024YFC3044800 to BH), the National Natural Science Foundation of China (No. 82330040 to BH, No. 82090044 to BH, No. 81820108010 to BH), the National Science and Technology Major Project (No. 2024ZD0527901 to BH), the Hubei Province Key R&D Program (No. 2022BCA008 to BH), and the Core medical technology project of union hospital (No. 2024JBGS2008 to BH). This work was also supported by the TCM Scientific Research Project under the Hubei Provincial Administration of Traditional Chinese Medicine (No. ZY2025M015).

The original contributions presented in the study are included in the article/, further inquiries can be directed to the corresponding authors. Supplementary material

The studies involving humans were approved by the Ethics Committee of Union Hospital, Tongji Medical College, Huazhong University of Science and Technology. The studies were conducted in accordance with the local legislation and institutional requirements. The ethics committee/institutional review board waived the requirement of written informed consent for participation from the participants or the participants’ legal guardians/next of kin because given that the study was conducted via telephone interviews, the requirement for written informed consent was waived. Instead, verbal informed consent was obtained from all participants or their legal guardians.

JL: Methodology, Data curation, Project administration, Writing – original draft, Supervision, Investigation, Writing – review & editing. SC: Writing – review & editing, Supervision, Conceptualization, Methodology, Investigation. LZ: Supervision, Methodology, Investigation, Writing – review & editing, Data curation. MQ: Data curation, Methodology, Investigation, Writing – original draft. WZ: Data curation, Writing – original draft, Investigation. LW: Writing – review & editing, Investigation, Data curation. QH: Writing – review & editing, Investigation, Data curation. YL: Data curation, Writing – review & editing, Investigation. HJ: Investigation, Writing – review & editing, Data curation. ST: Investigation, Data curation, Writing – review & editing. MH: Writing – review & editing, Investigation. CZ: Writing – review & editing. QJ: Writing – original draft. MW: Data curation, Methodology, Conceptualization, Investigation, Writing – review & editing, Writing – original draft. YW: Methodology, Writing – review & editing, Project administration, Data curation, Conceptualization, Investigation. BH: Supervision, Methodology, Resources, Conceptualization, Funding acquisition, Writing – review & editing, Project administration.

The author(s) declared that this work was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

The author(s) declared that generative AI was not used in the creation of this manuscript.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fnagi.2026.1724803/full#supplementary-material↗