Pain is a subjective and complex experience influenced by various factors, resulting in diverse brain representations across individuals. Previous studies have demonstrated the potential of group-level neuroimaging-based pain biomarkers; however, it r...
Pain is a subjective and complex experience influenced by various factors, resulting in diverse brain representations across individuals. Previous studies have demonstrated the potential of group-level neuroimaging-based pain biomarkers; however, it remains unclear how well they can predict pain without accounting for individual variability. Here, we developed an individualized pain biomarker using densely sampled functional MRI data from a single individual (28 sessions, totaling 56 hours of fMRI scans) and compared it with a population-level model (n = 124). The model exhibited high prediction performance within the individual with well-conserved individual-specific weight patterns. The fronto-parietal, limbic, and dorsal attention networks were identified as important regions for individual pain prediction, while the ventral attention network was crucial for population pain prediction. In addition, we assessed the unique variance explained by the individualized model. Taken together, our study identifies the characteristics of an individualized pain prediction model, underscoring the clinical potential of neuroimaging-based pain biomarkers and advancing personalized pain assessment.