Abstract
<jats:p>Abstract Background: Long COVID frequently includes post-exertional malaise, neuromuscular fatigue, internal vibrations, tremor-like symptoms, orthostatic intolerance, sleep disturbance, cognitive dysfunction, and autonomic instability. Neuroimaging studies in Long COVID and post-infectious ME/CFS have identified abnormalities across limbic, thalamic, mesiotemporal, brainstem, cerebellar, and white-matter systems, but no single established circuit accounts for this clinical phenotype. We examined whether abnormalities cluster at the mammillary body-fornix-superior tuberal hypothalamic interface and within connected brainstem-cerebellar pathways. Methods: Structural MRI and diffusion tensor imaging were analysed in 88 participants with clinician-diagnosed Long COVID and 34 healthy controls (total N = 122). Thirty-three patients were classified as bedridden. Volumetric and diffusion analyses focused on the mammillary bodies, superior tuberal hypothalamic region, fornix, dorsal and median raphe, midbrain reticular formation, and cerebellar peduncles. Fractional anisotropy was used as an index of white-matter microstructural organisation. Exploratory clinical associations included motor impairment, proprioceptive dysfunction, autonomic symptoms, fatigue, internal vibrations, and tremor-like symptoms. Results: Three mammillary-body volume phenotypes were identified: reduced, enlarged, and control-range. Left and right mammillary-body volumes differed across groups, with large effect estimates. Segmentation showed narrowing or loss of a visible internal passage at the superior tuberal-mammillary interface, while the fornix showed altered diffusion measures and reduced tract coherence in the hypothesised gate region. Additional findings included lower superior cerebellar peduncle volume, lower middle cerebellar peduncle fractional anisotropy, and lower dorsal raphe and midbrain reticular formation volumes in Long COVID. Conclusions: The findings support an exploratory mammillary body-fornix gate model in which a vulnerable periventricular hypothalamic-limbic interface may contribute to network dysfunction in a subgroup of patients with severe Long COVID. The data do not establish direct viral invasion, a coronavirus entry route, axonal destruction, or a single causal pathway. Prospective replication with standardised acquisition, preregistered regions of interest, correction for multiple comparisons, objective movement and autonomic measures, and longitudinal follow-up is required. Keywords: Long COVID; post-COVID condition; ME/CFS; mammillary body; fornix; hypothalamus; diffusion tensor imaging; internal vibrations; tremor-like symptoms; neuromuscular fatigue; dysautonomia; brainstem</jats:p>