Predicting Male vs. Female from Task-fMRI Brain Connectivity

Area of Science:

• Neuroscience
• Cognitive Science
• Brain Imaging

Background:

• Behavioral and cognitive functions exhibit sex-based differences in the brain.
• Prevalence of psychiatric disorders like ADHD, autism, depression, and schizophrenia varies between males and females.
• Previous research utilized resting-state fMRI to explore gender differences in functional brain networks.

Purpose of the Study:

• To investigate if task-based functional magnetic resonance imaging (task-fMRI) functional connectivity can predict gender.
• To identify specific functional connectivity features that are most predictive of gender.
• To explore gender-specific brain network changes during cognitive tasks.

Main Methods:

• Utilized a large dataset of 475 healthy controls from the Human Connectome Project (HCP) database.
• Extracted Pearson correlation coefficients from mean time-series of anatomical brain regions during an emotion task.
• Employed Partial Least Squares (PLS) regression with feature selection for gender classification.

Main Results:

• Achieved a classification accuracy of 0.88 for predicting male versus female using emotion task fMRI data.
• Identified interhemispheric connectivity as the most crucial feature for gender prediction from task-fMRI.
• Demonstrated the potential of task-fMRI to reveal gender-specific patterns in brain functional connectivity.

Conclusions:

• Task-fMRI functional connectivity is a viable method for predicting gender.
• Interhemispheric brain connectivity plays a significant role in gender differentiation during cognitive tasks.
• Findings may inform understanding of sex-specific risks and responses to stimuli or medications.