MIST: A multi-resolution parcellation of functional brain networks

Abstract

Functional brain connectomics investigates functional connectivity between distinct brain parcels. There is an increasing interest to investigate connectivity across several levels of spatial resolution, from networks down to localized areas. Here we present the Multiresolution Intrinsic Segmentation Template (MIST), a multi-resolution parcellation of the cortical, subcortical and cerebellar gray matter. We provide annotated functional parcellations at nine resolutions from 7 to 444 functional parcels. The MIST parcellations compare well with prior work in terms of homogeneity and generalizability. We found that parcels at higher resolutions largely fell within the boundaries of larger parcels at lower resolutions. This allowed us to provide an overlap based pseudo-hierarchical decomposition tree that relates parcels across resolutions in a meaningful way. We provide to explore the MIST parcellations and an interactive web interface also made it accessible in the neuroimaging library nilearn. We believe that the MIST parcellation will facilitate future investigations of the multiresolution basis of brain function.

Neuropsychiatric mutations delineate functional brain connectivity dimensions contributing to autism and schizophrenia

“…We segmented the brain into 64 functional seed regions defined by the multi-resolution MIST brain parcellation 51 . FC was computed as the temporal pairwise Pearson's correlation between the average time series of the 64 seed regions, after Fisher transformation.…”

Multivariate consistency of resting-state fMRI connectivity maps acquired on a single individual over 2.5 years, 13 sites and 3 vendors

“…) . The networks we used came from the Multiresolution Intrinsic Segmentation Template ( MIST) parcellation, which overlaps substantially at this resolution with the Yeo-Krienen atlas(Urchs et al, 2017) . The MIST atlas was generated from 200 healthy subjects and consists of nine functional parcellations capturing successively finer levels of spatial detail, of which we used parcellations from resolution seven, consisting of seven commonly used large-scale networks: cerebellar (CER), default-mode (DMN), frontoparietal (FPN), limbic (LIM), motor (MOT), salience (SAL), and visual (VIS).…”

Unraveling reproducible dynamic states of individual brain functional parcellation

“…To generate seed-based parcellations, we studied seed voxels from three regions of the MIST parcellation. We picked MNI coordinates (0, -76, 10), (0, 20, 28) and (3, -43, 37) as the respective medoids of the ROIs 90, 6 and 42 corresponding to the posterior medial visual network (PM-VIS), the dorsal anterior cingulate cortex (dACC) and the posterior cingulate cortex (PCC) networks in the MIST atlas [9] . The choice of seeds is driven by the properties of the networks in the literature.…”

“…This interest dates back to at least a century, with the Broadman atlas [7] . Many popular functional parcellations have been generated at the group level, by averaging the connectivity patterns over many subjects, such as the Yeo-Krienen [8] , MIST [9] , Glasser [10] and Schaffer atlases [11] , amongst many others [12] . Researchers have proposed competing approaches to generate such group parcellations, with improvement measured in terms of reproducibility and homogeneity of the parcels across datasets and subjects, e.g.…”