SEGMA: An Automatic SEGMentation Approach for Human Brain MRI Using Sliding Window and Random Forests.

Area of Science:

• Neuroimaging
• Developmental Neuroscience
• Computational Neuroscience

Background:

• Quantitative brain magnetic resonance imaging (MRI) data across the lifespan is crucial for understanding brain development and aging.
• Existing segmentation tools often lack the adaptability for diverse age groups and large datasets.
• Automated segmentation is needed to efficiently analyze brain structure changes throughout life.

Purpose of the Study:

• To develop an automated segmentation method for human brain MRI applicable across the entire life course.
• To enable accurate analysis of brain structure in newborns, children, adolescents, and adults.
• To create a tool that can efficiently process large-scale neuroimaging datasets.

Main Methods:

• Developed an automatic segmentation method for human brain MRI.
• Utilized a sliding window approach combined with a multi-class random forest classifier.
• Applied the method to high-dimensional feature vectors for precise segmentation.

Main Results:

• The method demonstrated strong performance on brain MRI data from 179 individuals across three age groups: newborns, children/adolescents, and adults.
• The automated segmentation accurately captured brain volumes throughout the lifespan.
• The approach is capable of learning from partially labeled datasets, enhancing efficiency.

Conclusions:

• The developed automated segmentation method is effective for analyzing human brain MRI across all life stages.
• This tool facilitates efficient segmentation of large-scale datasets and can be applied to diverse populations and imaging modalities.
• It supports research into long-term effects on brain development, healthy aging, and early life determinants of adult brain structure.