Magnetic resonance imaging (MRI) provides an unprecedented ability to investigate brain health and function. However, the high cost and high variability in the population limits its use in understanding complex diseases, which requires new methodologies. As such, the focus of this project was to define and identify a novel reorganisation of long Covid brain MRI data into network graphs and subgraphs based on functional, rather than spatial, connections between voxels. We define a physiological connectome: a graph in which the nodes are voxels, MRI metrics are node attributes, and edges are formed according to physiological similarity. From these graphs we define local neighborhood subgraphs containing each voxel’s set of nearest neighbours, which we examine for their properties. MRI features most strongly associated with high connectivity included low values of axial diffusivity (AD), mean diffusivity (MD), radial diffusivity (RD), and isotropic volume fraction (ISOVF); and mean values of intracellular volume fraction (ICVF). Orientation dispersion index (ODI) and fractional anisotropy (FA) values were highly variable. The general data trends outside of highly connected voxels include an inverse relationship between FA and ODI, RD and FA, AD and ODI, ICVF and RD, and ICVF and MD; a positive correlation between AD and MD, ODI and RD, and RD and MD. RD and AD, and ISOVF and ICVF do not demonstrate a clear trend. In future, these subgraphs will form the basis for a generalisable data augmentation and analysis method, to identify underlying patterns responsible for large-scale functional differences between brains.