Creation of structural MRI brain templates for middle to late adulthood in the Thai population using single and multi-scanner data.

BACKGROUND

The global increase in elderly populations, particularly in low- and middle-income countries, underscores the rising prevalence of age-related conditions, such as Alzheimer's disease. Structural magnetic resonance imaging (MRI) plays a crucial role in Alzheimer's disease research and early diagnosis; however, manual brain segmentation is labor-intensive, time-consuming, and prone to inter-rater variability. Automated segmentation using existing brain templates from Mongoloid, Caucasoid, and Negroid populations may not be suitable for Thai individuals due to anatomical differences, emphasizing the need for population-specific templates. This study aimed to develop structural MRI brain templates tailored for cognitively normal Thai adults aged 50-70 years, using both single- and multi-scanner data.

METHODS

This study developed structural brain templates tailored for cognitively normal Thai adults aged 50-70 years. Participants underwent 3D T1-weighted MRI scans across three scanners. The image processing pipeline included bias field correction, rigid registration to standard space, and iterative non-rigid registration to ensure anatomical consistency. Two templates were created: TH150, based on data from a single scanner site, and TH240, constructed from multi-site data collected across three scanners.

RESULTS

Two templates were created: TH150 and TH240, which exhibited nearly identical brain spatial dimensions (width, length, and height). These features closely align with those obtained from individual brains through manual measurement, confirming the reliability of the template construction pipeline. Compared to CN200 and US200, the Thai templates showed 16-22% less gray matter, 2-9% less white matter, and 27-68% more cerebrospinal fluid. These differences likely reflect age-related changes, as the Thai templates were constructed from individuals aged 50-70 years, whereas the CN200 and US200 templates represent younger populations.

CONCLUSIONS

These Thai adults aged 50-70 years population-specific templates reveal morphological differences from Chinese and Western templates. The high consistency between TH150 and TH240 supports methodological robustness, providing a foundation for future neuroimaging research and applications.