AI tool Detects Gender-Based Differences in Brain Structure
A recent study reveals that artificial intelligence (AI) software analyzing MRI scans can detect differences in the cellular organization of the brains of men and women. Researchers observed these distinctions in the white matter, which primarily resides in the innermost layer of the human brain and facilitates communication between different brain regions.
Men and women exhibit differences in the rates and symptoms of various brain-related conditions such as multiple sclerosis, autism spectrum disorder, and migraines.
Understanding how biological sex influences the brain is seen as crucial for enhancing diagnostic methods and treatments. However, while aspects like brain size, shape, and weight have been studied, researchers have only a limited understanding of the brain’s cellular layout.
A study led by researchers at NYU Langone Health utilized machine learning, an AI technique, to analyze thousands of MRI brain scans from 471 men and 560 women.
The results revealed that computer programs could accurately distinguish between male and female brains by identifying patterns in structure and complexity that were not visible to the naked eye.
Validation by Diverse AI Models
Three different AI models confirmed these findings, each emphasizing different strengths: pinpointing small portions of white matter or analyzing relationships across larger brain regions.
Dr. Yvonne Lui, the study’s senior author and neuroradiologist, noted, “Our findings provide a clearer picture of how a living, human brain is structured, which may in turn offer new insight into how many psychiatric and neurological disorders develop and why they can present differently in men and women.”
Lui, a professor and vice chair for research in the Department of Radiology at NYU Grossman School of Medicine, highlights that prior examinations of brain microstructure heavily relied on animal models and human tissue samples.
Challenges in Statistically Analyzing Manually Drawn Regions
Furthermore, some past findings have faced credibility challenges because they relied on statistical analyses of manually drawn regions of interest, requiring numerous subjective decisions about the shape, size, and location of these regions, potentially biasing the results.
To circumvent this issue, the recent study employed machine learning to analyze entire sets of images without specifying any particular spot for inspection, thereby minimizing human biases, as noted by the authors.
The research began by training AI models with existing data comprising brain scans from healthy individuals of both sexes, and indicating the biological sex of each scan.
As they accumulated more data, these models were designed to progressively enhance their capability to independently distinguish biological sex. Crucially, Lui prohibited the models from utilizing overall brain size and shape for their determinations.
Features Influencing Machine Determinations
The outcomes revealed that all models correctly identified the sex of scanned subjects between 92% and 98% of the time. Certain features, including the ease and direction of water movement through brain tissue, were particularly influential in the machines’ determinations.
Junbo Chen, MS, a doctoral candidate at NYU Tandon School of Engineering and one of the study’s co-lead authors, emphasized, “These results highlight the importance of diversity when studying diseases that arise in the human brain.”
Meanwhile, Vara Lakshmi Bayanagari, MS, a graduate research assistant at NYU Tandon School of Engineering and another co-lead author, cautioned against using men as the standard model for various disorders, as this approach may overlook critical insights.
Inability to Attribute Features to a Specific Gender
Bayanagari cautioned that while the AI tools could identify differences in brain-cell organization, they could not specify which sex was more likely to exhibit certain features. She added that the study categorized sex based on genetic information and exclusively included MRIs from cis-gendered men and women.
The authors stated that their future research will delve into understanding the evolution of sex-related differences in brain structure over time to better grasp the potential influence of environmental, hormonal, and social factors on these changes.
In addition to Lui, Chen, and Bayanagari, other NYU Langone Health and NYU researchers involved in the study were Sohae Chung, Ph.D., and Yao Wang, Ph.D.
Read the original article on: Medical Xpress
Read more: AI-Created Gene Editing Tools Successfully Alter Human DNA
