MR-TIM

MR-TIM produces detailed head tissue segmentations from structural magnetic resonance (MR) images in MATLAB to improve EEG source reconstruction by incorporating electrode positions and accurate head tissue distributions.

Key Features:

• MATLAB implementation: Implemented in MATLAB for processing structural MR images.
• Image pre-processing (denoising): Performs denoising of raw MR images to reduce artifacts prior to further analysis.
• Tissue Probability Mapping (TPM): Generates template tissue probability maps in individual subject space from pre-processed MR images.
• Tissue segmentation: Integrates TPMs to assign each voxel to a tissue class and produces realistic 3D tissue masks.
• Twelve tissue classes: Segments twelve tissue types: brain grey matter, cerebellar grey matter, brain white matter, cerebellar white matter, brainstem, cerebrospinal fluid, spongy bone, compact bone, eyes, muscle, fat, and skin.
• Electrode position integration: Incorporates detailed electrode positions alongside head tissue distributions into the modeling.

Scientific Applications:

• EEG source reconstruction: Provides realistic head models and tissue segmentations to improve localization accuracy in EEG source reconstruction.
• Validation of segmentation accuracy: Demonstrated superior whole-head tissue segmentation accuracy on MR images from healthy volunteers, patients, and an MR template image from an open repository.
• Neuroimaging investigations: Enables generation of head models used to investigate neural activity and connectivity in EEG studies.

Methodology:

MR-TIM applies three computational modules: denoising of raw MR images; generation of tissue probability maps in individual subject space from pre-processed MR images; and integration of TPMs to assign voxels to tissue classes, producing 3D masks for twelve tissue types.