Date of Award

January 2026

Document Type

Thesis

Degree Name

Medical Doctor (MD)

Department

Medicine

First Advisor

Lindsay S. McAlpine

Abstract

Background: Neuropsychiatric Long COVID (LC) is a heterogeneous and debilitating disease spanning multiple mental health symptom domains including depression, anxiety, cognitive dysfunction, fatigue, and post-traumatic stress disorder (PTSD). Despite the prevalence of neuropsychiatric LC, significant gaps exist in the current body of research as to the pathogenesis of the disease. Methods: We recruited 32 individuals with neuropsychiatric LC (new self-reported neurologic and/or psychiatric symptoms >3 months after COVID) and 17 recovered controls (COVID >3 months prior and no LC) who were prospectively enrolled for neuropsychological testing, mental health questionnaires, and an MRI scan of the brain on a single Siemens 3T scanner. We further recruited 31 participants with neuropsychiatric LC as well as 22 recovered controls who consented to a large-volume lumbar puncture for cerebrospinal fluid (CSF) proteomic analysis in addition to neuropsychological testing and mental health questionnaires. Results: Participants with LC in the MRI cohort had significantly worse neuropsychological assessment scores across all tested domains compared to controls. In the proteomics cohort, participants with LC still had worse neuropsychological assessment scores than controls, but these differences only reached statistical significance for the PHQ-9, GAD-7, PSS, and GNA processing speed. Cortical gray matter (CGM) and cerebral blood flow (CBF) were measured using MRI and correlations between these values and mental health questionnaire scores were assessed for all cortical regions. Variations in CGM volume in the left precuneus were associated with PTSD severity in the LC but not control groups. Decreased cerebral blood flow (CBF) to the left amygdala was associated with worse depression and apathy symptoms in the LC but not control groups and decreased CBF of several key sensory processing regions was associated with worse symptoms of depression and apathy in the LC group. The association between decreased CBF and mental health symptom severity was unique to the LC group and not found in the control group. Furthermore, CSF protein expression levels were examined for all immune-related proteins and correlated with neuropsychiatric assessment scores. Many immune-related proteins showed significant associations with symptoms of depression, apathy, anxiety, and PTSD in the LC, but not control groups. The most significant proteins were involved in the interleukin-6 (IL-6), nitric oxide synthase (NOS), and arginase (ARG) pathways and were strongly associated with symptoms of apathy and depression. Conclusions: Neuropsychiatric symptoms in LC have a unique neurobiologic signature which corresponds with several key features of LC pathogenesis. Scores of apathy and depression were strongly associated with decreased CBF in several key brain regions and also corresponded to changes in immunologic proteins involved in the IL-6/NOS/ARG pathway which may correspond to persistent neurovascular dysfunction in LC. Furthermore, apathy and depression symptoms were associated with changes to immunologic proteins involved in T-cell exhaustion and innate immune activation, which parallels previous research into the origins of LC. None of these findings carried over to the control group. Taken together, our study suggests that the unique features of persistent neurovascular dysfunction and immune dysregulation play a strong role in the pathogenesis of neuropsychiatric symptoms in LC, particularly in relationship to scales of apathy and depression.

Comments

This thesis is restricted to Yale network users only. It will be made publicly available on 07/14/2027

Download

Share

COinS