Objective Neuropathology caused by COVID-19 has been widely reported, and the characterisation of the spatial distribution of these pathology remains critical to assess long and short-term neurological sequelae.

Methods We performed a systematic review of the literature to quantify the locations of small neurological events identified with magnetic resonance imaging (MRI) among COVID-19 patients. Neurological events were localised into the Desikan-Killiany grey and white matter atlases. A mathematical network diffusion model was then used to test whether the spatial distribution of neurological events could be explained via a linear spread through the structural connectome of the brain.

Results The highest proportions (26%) of white matter events were observed within the bilateral corticospinal tracts. The highest proportions (~10%) of grey matter events were observed in areas including the bilateral superior temporal, precentral, and lateral occipital cortices. Sub-cortical events were most frequently identified in the Pallidum. The application of a mathematical network diffusion model suggested that the spatial pattern of the small neurological events in COVID-19 can be modelled with a linear diffusion of spread from epicentres in the bilateral cerebellum and basal ganglia (Pearson’s r=0.41, p<0.001, corrected).

Conclusions To our knowledge, this is the first study to systematically characterise the spatial distribution of small neurological events in COVID-19 patients and test whether the spatial distribution of these events can be explained by a linear diffusion spread model. As such, initial sub-cortical events which manifest as altered consciousness could be expected to be followed by later cortical events manifesting as altered sensorimotor functioning.