Fig. 1: Exploration cohort overview: study design and neurologic sequelae across PASC subgroups.

a A total of 278 individuals were enrolled between November 2022 and April 2025, of whom 269 with prior COVID-19 infection were included in the present analysis after excluding nine uninfected participants. All participants underwent cognitive evaluations using the Montreal Cognitive Assessment (MoCA), the Fatigue Assessment Scale (FAS), and a symptom-based Post-Acute Sequelae of COVID-19 (PASC) questionnaire designed to assess cognitive impairment. Based on the time of enrollment, participants were assigned to the exploration cohort (n = 160; November 2022–October 2023) and the validation cohort (n = 109; November 2023–April 2025). Each cohort was further stratified into three clinical subgroups according to symptom profiles: NS-PASC (exploration: n = 77, validation: n = 46), Other-PASC (exploration: n = 55, validation: n = 47), and Cog-PASC (exploration: n = 28, validation: n = 16). The study encompassed three main analytical approaches: PASC symptom questionnaires for cognitive function assessment (including 32 symptoms covering brain fog, fatigue, sleep disturbances, dizziness, headache, myalgia, and anxiety/depression), blood analysis using proteomic techniques (Proximity Extension Assay [PEA] with Olink 3072 Explore platform and Single-Molecule Array [Simoa] measuring GFAP, NfL, Aβ42/40 ratio, p-Tau181), and brain Magnetic Resonance Imaging (MRI) analysis focusing on cortical parcellation, paramagnetic susceptibility maps, choroid plexus segmentation, and DTI-ALPS processing. b, c Comparison of long COVID-related neurologic symptoms between NS-PASC vs. PASC and Other-PASC vs. Cog-PASC subgroups. Two-sided Pearson’s Chi-square tests were used to analyze categorical variables. Significance levels are indicated as *p < 0.05, **p < 0.01, and ***p < 0.001. χ² statistics, degrees of freedom, and exact p-values are provided in the Source Data file.