PREEMACS: Pipeline for preprocessing and extraction of the macaque brain surface

Area of Science:

• Neuroimaging
• Comparative Neuroscience
• Medical Image Analysis

Background:

• Cortical surface extraction is vital for human brain studies but challenging in non-human primates (NHPs) due to unique artifacts.
• Existing automated methods for NHP cortical surface extraction are limited and not tested on diverse data.

Purpose of the Study:

• To present PREEMACS, a novel pipeline for standardizing structural MRI preprocessing and automating macaque brain surface extraction.
• To evaluate the generalizability and robustness of PREEMACS across multiple acquisition sites.

Main Methods:

• PREEMACS integrates volume orientation, cropping, intensity correction, averaging, and CNN-based skull-stripping.
• Quality control is performed using an adapted MRIqc method to ensure reliable surface estimation.
• White matter and pial surfaces are extracted using a customized FreeSurfer version, leveraging both T1w and T2w MRI data.

Main Results:

• PREEMACS demonstrated accurate and robust automatic brain surface extraction on NHP MRI data.
• The pipeline was successfully validated on 57 NHP MRI volumes from 11 different sites (PRIME-DE dataset).
• The quality control module effectively identified images suitable for reliable surface extraction.

Conclusions:

• PREEMACS provides a robust, efficient, and generalizable solution for automated MRI surface analysis in macaques.
• This pipeline facilitates standardized neuroimaging analysis and cortical thickness estimation in NHP research.
• The developed method addresses key challenges in NHP neuroimaging, promoting wider adoption of advanced analysis techniques.