Individualized functional connectome biomarkers predict clinical symptoms after rTMS treatment in Alzheimer's disease

Pharmacological treatments for Alzheimer's disease (AD) often show limited effectiveness, prompting growing interest in non-drug approaches such as repetitive transcranial magnetic stimulation (rTMS). However, the effects of rTMS can vary widely between individuals with AD, highlighting the need to better understand brain characteristics that may influence treatment response. In this study, we applied a personalized method to divide each participant's brain cortex into functionally meaningful regions based on their brain activity patterns, rather than relying on a standard brain template. Using this individualized brain mapping approach, we examined how rTMS changes functional connectivity (FC) across the brain. We further used support vector regression to estimate whether these individualized FC patterns could predict the severity of clinical symptoms. The results showed that rTMS significantly increased whole-brain individualized FC strength during resting state, with the most prominent effects observed in the default mode and visual networks (Cohen's d > 0.27, corrected p < 0.05). Importantly, the personalized FC features served as predictive biomarkers, demonstrating greater accuracy in forecasting clinical outcomes compared to the conventional group-based approach. These findings suggest that individualized brain connectivity holds significant potential for guiding personalized therapeutic strategies and improving treatment efficacy in AD.