BRAHMA: Population specific T1, T2, and FLAIR weighted brain templates and their impact in structural and functional imaging studies

Highlights

• BRAHMA, a collection of T1w, T2w, and FLAIR templates for the Indian population

• A full set of TPMs is provided for CSF, GM, WM, DGM regions and ROI labels.

• Template compared with ICBM-152 for validation in structural and functional studies.

• T2 and FLAIR template provide accurate segmentation of ROIs like substantia nigra.

• BRAHMA template and construction pipeline are made available for open access.

Abstract

Differences in brain morphology across population groups necessitate creation of population-specific Magnetic Resonance Imaging (MRI) brain templates for interpretation of neuroimaging data. Variations in the neuroanatomy in a genetically heterogeneous population make the development of a population-specific brain template for the Indian subcontinent imperative. A dataset of high-resolution 3D T1, T2-weighted, and FLAIR images acquired from a group of 113 volunteers (M/F - 56/57, mean age-28.96 ± 7.80 years) are used to construct T1, T2-weighted, and FLAIR templates, collectively referred to as Indian Brain Template, “BRAHMA”. A processing pipeline is developed and implemented in a MATLAB based toolbox for template construction and generation of tissue probability maps and segmentation atlases, with additional labels for deep brain regions such as the Substantia Nigra generated from the T2-weighted and FLAIR templates. The use of BRAHMA template for analysis of structural and functional neuroimaging data obtained from Indian participants, provides improved accuracy with statistically significant results over that obtained using the ICBM-152 (International Consortium for Brain Mapping) template. Our results indicate that segmentations generated on structural images are closer in volume to those obtained from registration to the BRAHMA template than to the ICBM-152. Furthermore, functional MRI data obtained for Working Memory and Finger Tapping paradigms processed using the BRAHMA template show a significantly higher percentage of the activation area than ICBM-152 in relevant brain regions, i.e. the left middle frontal gyrus, and the left and right precentral gyri, respectively. The availability of different image contrasts, tissue maps, and segmentation atlases makes the BRAHMA template a comprehensive tool for multi-modal image analysis in laboratory and clinical settings.

Introduction

A brain template provides a standard reference coordinate system for the processing and analysis of multi-modal neuroimaging data collected from an individual or a study group. They are widely used in neuroscience research as registration targets for the alignment of images obtained from different sources, and are also essential for performing pre-processing steps such as spatial normalization, that are necessary to compensate anatomical variations in study populations [1]. Templates are also used with atlases containing labels corresponding to different brain regions for segmentation of Magnetic Resonance Imaging (MRI) data [2], with the segmentation volumes being used for structural studies through Voxel-Based Morphometry (VBM) [3]. These atlas segmentations are also utilized for functional studies using functional magnetic resonance imaging (fMRI) [4] or Magnetoencephalography (MEG) data analysis [5]. Templates are increasingly being utilized in analysis of large imaging datasets [6,7] and translated to clinical practice as reference images in differentiating normal neurological development and aging in individuals from abnormal neuroanatomy that might be due to neurodegenerative disorders [[8], [9], [10], [11]].

A template enables the combination of structural and functional data from multiple subjects to create statistically significant group-averages, making them vital in applications such as mapping of functional activations in fMRI experiments [4]. They are also applied to images acquired from other MR sequences such as Quantitative Susceptibility Mapping (QSM) [[12], [13], [14]] to isolate the specific brain tissue or regions, and in Diffusion Tensor Imaging (DTI) [[15], [16], [17], [18]]. Templates are also significant in multi-modality image registration, as in case of Computed Tomography with Positron Emission Tomography (PET) or Single Photon Emission Computed Tomography (SPECT). Brain templates are also utilized along with individual's brain scan in clinical settings, such as in the planning and evaluation of Transcranial Magnetic Stimulation (TMS) [[19], [20], [21]] and Deep-Brain Stimulation (DBS) [22,23].

The earliest atlases used stereotactic procedures to localize brain structures with reference to external landmarks such as the auditory meatus and orbital boundary, or internal landmarks such as the Anterior Commissure (AC) and Posterior Commissure (PC). The Talairach and Tournoux atlas [24] created from postmortem sections of a single individual brain formed the basis for modern digital templates, with templates such as the Montreal Neurological Institute (MNI) template (MNI-305) [25] and International Consortium for Brain Mapping (ICBM) templates (ICBM-152 [26] and ICBM-452 [27,28]) being widely used for neuroimaging data analysis. These templates are derived through a combination of several structural scans taken from a population group, with the individual images being registered to each other and averaged to arrive at a statistically optimal image that is representative of the population group. The template images are primarily used to normalize individual scans obtained from a cohort of participants to generalize the results of study sample to the entire population. Such spatial normalization introduces errors during the forward or inverse transformations or deformations, which may lead to flaws in the processed data and yield inaccurate results. Furthermore, the structure of the human brain varies with age and race of individuals, making atlases and labels associated with the template subjective to the scans being used for the construction of the template. Hence, the reliability and utility of the templates and associated segmentation atlases are highly dependent on the closeness between the template and the clinical population on which they are to be applied, and templates used for any analysis should be obtained from individuals having a similar traits to the study participants.

Presently, used brain templates from MNI and ICBM have been created from MRI scans obtained from a cohort of Caucasian descent. The differences in brain structures between such groups used to generate the templates and the cohorts being studied lead to erroneous results in clinical and research usage depending upon the degree of variation between the template and the data being studied [[29], [30], [31], [32], [33], [34], [35]]. Fig. 1 represents some of the adult brain templates currently available for neuroimaging studies along with associated tissue maps. The clear differences between the average templates generated over the Caucasian templates (MNI-305, ICBM-152, and ICBM-452) and the Chinese templates (Chinese Brain Atlas (CBA-56) and Statistical Chinese Brain Template (SCBT-2020)) populations highlight the need for an Indian population specific template which is different from each of the present groups. These variations have necessitated the development of population-specific brain templates for different age groups and races [[35], [36], [37], [38], [39], [40], [41], [42]]. A comprehensive list of templates and atlases for neuroimaging analysis is also available in various reviews articles [[43], [44], [45]]. India, with a population of nearly 1.33 billion, forms approximately one-fifth of the world's population share, and variations in the overall brain structure and morphology between Indian and other ethnic groups necessitates the development of an Indian population-specific brain template [44,46]. While some efforts have been made in this direction, the size of the study cohort and possible homogeneity between the participants of the study severely restricts its generalizablity to the entire population [46]. Furthermore, most population-specific templates only provide a T1-weighted template image without any Tissue Probability Maps (TPMs) or atlases necessary for the segmentation of different brain tissue regions and structures.

This manuscript presents the construction of an Indian brain template using a dataset of high-resolution T1, T2-weighted, and FLAIR images collected on participants hailing from all over India. A pipeline for the construction of template images, probability maps, and segmentation atlases and its implementation through a MATLAB-based toolbox is also described. A set of T1, T2-weighted, and FLAIR templates constructed from the collected dataset is provided along with TPMs and segmentation atlases to facilitate robust and accurate neuroimaging analysis from different studies performed on Indian population. The T2-weighted and FLAIR templates provide a clear view of deep brain structures that are not readily distinguishable on T1-weighted template images. Furthermore, a set of segmentation atlases and a package for the segmentation of structures such as the Substantia Nigra (SN) are also incorporated as part of the toolbox. The complete BRAHMA package consisting of the template, atlases, and the template construction toolbox is available for download (http://www.nbrc.ac.in/newweb/research/groups/PM).