Integrating inflammation-oriented brain injury mechanisms in acute COVID-19 using multi-modality MRI and serum proteomics: a longitudinal study

Background
Neurological sequelae have been frequently reported among COVID-19 patients, which are still lacking of definite therapeutic targets. Previous studies linked structural and functional brain alterations on MRI to the brain injury and neuropsychiatric symptoms in COVID-19, but few of them provided molecular underpinnings of the brain changes driven by infection.

Methods
Pre- and post-infection MRI including 3D-T1WI, fMRI (function MRI), and dMRI (diffusion MRI) and neuropsychiatric assessments were collected both about one month from infection (interval for 70 days) on patients who were first infected with COVID-19. Longitudinal data were compared by paired t-test and adjusted by multiple comparison correction to reveal the specific brain changes after infection. Generalized linear mixed models and multiple linear regression were used to detecte interactions between the brain changes on MRI and neuropsychiatric assessments. Serum samples were collected in the follow-up scanning for proteomics analysis to uncover the molecular mechanisms of MRI-visible brain injury.

Results
Brain changes after COVID-19 infection include gray matter volume decrease, white matter microstructure integrity decrease, and brain function alterations, particularly in the limbic system. The hyperfunction of IPL (inferior parietal lobe) and MOFG (medial orbitofrontal gyrus) correlated with the atrophy in the corresponding regions of the contralateral hemisphere. The decreased functional index of amplitude of low-frequency fluctuation (ALFF) on frontal lobe was further correlated with sleeping disorders (P = 0.003). The proteomic indicated that the MRI-visible brain changes represented neural injury (NEFL [neurofilament light chain]: FDR-corrected p < 0.046) and were driven by the COVID-19-specific inflammation (P adj = 0.00015).ConclusionsThis study investigated the relationships between structural and functional brain changes and neuropsychiatric sequelae in the acute recovery phase of COVID-19, and identified the proteomics significance of MRI-visible brain changes. These findings provided more explicit MRI biomarkers and risk factors for COVID-19 infection, which may be generalizable to other coronavirus studies.