Plasma neurofilament light chain (NfL) concentrations in children with malaria. Longitudinal differences in concentrations of plasma NfL from admission to 84 hours by severity group (uncomplicated malaria [UM], severe malaria [SM] without neurological manifestations, or SM with neurological manifestations) (A) or from admission to day 14 by presence of each clinical neurological trait in children with SM (B–D). Longitudinal regression lines depict linear mixed-effects (LME) model–predicted, back-transformed values and 95% confidence intervals of NfL over time stratified by severity group (A) or clinical trait (B–D), while holding other fixed effects constant. All LME models included time postadmission, child sex, child age, and treatment arm (rosiglitazone vs placebo) as fixed effects, and a by-participant intercept as a random effect. An interaction term between severity group (A) or each clinical trait (B–D) and time postadmission was included in each model to account for variation in NfL over time. P values overlaid on graphs indicate the statistical significance of the interaction term assessed using a Wald test. A total of 30 children had UM, 55 had SM without neurological manifestations, and 82 had SM with neurological manifestations (15 with coma, 40 with impaired consciousness, and 69 with a history of ≥2 seizures in the preceding 24 hours). For the selected timepoints, 71.1% (788/1109) of samples were available for NfL quantification.