Glasser52: A Brain Parcellation for MEG Analysis

A neuroimaging dataset providing a 52-parcel brain atlas derived from the Human Connectome Project Multimodal Parcellation (HCP-MMP). The parcellation was created by adapting the anatomical HCP-MMP atlas using functional signals from magnetoencephalography (MEG) recordings of healthy controls and Parkinson's disease volunteers, following the method of Tait et al. (2021). It includes parcellation files at multiple voxel resolutions, label files, and documentation detailing the merging rationale and parcel coordinates.

Use Cases

• Source localization in MEG studies based on a functionally-informed anatomical atlas.
• Region-of-interest analysis for MEG sensor signals using the provided 52-parcel labels.
• Comparing brain activity patterns between healthy controls and Parkinson's disease patients using the derived parcellation.
• Mapping MEG data to standard MNI152 space using the provided NIfTI files at 1mm, 2mm, and 8mm resolutions.

Strengths

• Provides a specific 52-parcel scheme designed for MEG analysis, offering a compromise between purely anatomical and functional parcellations.
• Includes multiple practical file formats: NIfTI parcellation images at three standard resolutions (1mm, 2mm, 8mm) and supporting label files.
• Documents the methodological rationale with an Excel sheet detailing why specific HCP-MMP regions were merged or divided.

Limitations

• Key metadata such as dataset size (row count), specific column names, and last update date are not provided across any source.
• The dataset's creation is based on MEG data from a specific cohort (healthy and PD volunteers), which may limit generalizability to other populations.
• Relies on inference for data types; the core 'dataset' is a set of parcellation maps and labels rather than a traditional tabular or time-series record.

Provenance

Source

University of Oxford (author Oliver Kohl), based on the HCP-MMP atlas (Glasser et al., 2016).

Collection Method

Created by calculating the influence of 22 HCP-MMP regions on MEG recordings, then merging and dividing regions to obtain 52 parcels with similar MEG sensor signal influence.

Geography

Brain atlas in standard MNI152 stereotactic space.