Characterizing predictors and chronicity of brain fog in long COVID

Long COVID is a chronic illness that persists following COVID-19 infection, with fatigue and brain fog as the most frequent complaints. However, there is no objective case definition for brain fog in long COVID and chronicity of symptoms remains unclear. This study recruited two waves of participants: those with a history of COVID-19, who participated in 2023 (N = 793, age = 38.5 ± 13.2, 44% female, 35.1% long COVID) and a follow up cohort collected in 2024 of participants who qualified as having long COVID at time point 1 (N=119, 61 female, 58 male). Participants completed questionnaires and cognitive tasks from home. We trained a binary classification model for long COVID diagnosis (73% accuracy) and a linear regression model for cognitive complaints (RMS error 5.8). A long COVID diagnosis at timepoint 1 was classified by biopsychosocial variables including stress, social support, and sex (women more likely). Symptom clustering revealed that phenotypes with both mental and physical health symptoms were predictive of brain fog, but not phenotypes with only sleep-related or physical symptoms. Markers of brain fog included slower reading and typing, slower reaction times in cognitive tasks, and changes in information processing speed and thresholds for making choices. Timepoint 2 data showed that the majority (82.4%) of participants did not remit from long COVID . These findings highlight the complex biopsychosocial factors that predict having long COVID with brain fog, and the need for interventions to improve remission rates.