Gender Similarities Dominate Mathematical Cognition at the Neural Level: A Japanese fMRI Study Using Advanced Wavelet Analysis and Generative AI

TL;DR

This study used advanced wavelet analysis on fMRI data from Japanese children to show that neural mechanisms of mathematical cognition are largely similar between genders, challenging claims of inherent differences.

Contribution

It introduces a novel wavelet-based dynamic analysis approach to neural data, revealing gender similarities in mathematical cognition at the neural process level.

Findings

89.1% similarity in neural activation patterns between genders

Neural dynamics differences exceeded group differences by 3.2:1

Machine learning classifiers performed at chance in gender prediction

Abstract

Recent large scale behavioral studies suggest early emergence of gender differences in mathematical performance within months of school entry. However, these findings lack direct neural evidence and are constrained by cultural contexts. We conducted functional magnetic resonance imaging (fMRI) during mathematical tasks in Japanese participants (N = 156), employing an advanced wavelet time frequency analysis to examine dynamic brain processes rather than static activation patterns. Wavelet decomposition across four frequency bands (0.01-0.25 Hz) revealed that neural processing mechanisms underlying mathematical cognition are fundamentally similar between genders. Time frequency analysis demonstrated 89.1% similarity in dynamic activation patterns (p = 0.734, d = 0.05), with identical temporal sequences and frequency profiles during mathematical processing. Individual variation in neural dynamics exceeded group differences by 3.2:1 (p $<$ 0.001). Machine learning classifiers achieved only 53.8% accuracy in distinguishing gender based neural patterns essentially at chance level even when analyzing sophisticated temporal spectral features. Cross frequency coupling analysis revealed similar network coordination patterns between genders, indicating shared fundamental cognitive architecture. These findings provide robust process level neural evidence that gender similarities dominate mathematical cognition, particularly in early developmental stages, challenging recent claims of inherent differences and demonstrating that dynamic brain analysis reveals neural mechanisms that static behavioral assessments cannot access.