A longitudinal naturalistic fMRI study of trait anxiety and cross-phase consistency in subjective fear signature expression

Persistent fear is a central feature of anxiety, but it remains unclear whether this persistence is more closely related to the consistency of conditioned-threat processing or to the consistency of subjectively experienced fear. Using a longitudinal naturalistic viewing fMRI paradigm, the present study examined whether trait anxiety was associated with cross-phase consistency in the expression of two validated fear-related neural signatures. Forty-two participants completed two fMRI phases separated by approximately three months, viewing the same sequence of 18 naturalistic videos followed by a post-viewing resting-state segment, resulting in 19 corresponding analysis segments for all cross-phase analyses. We applied pre-trained subjective-fear and threat-conditioning signatures to estimate their expression over time and examined how consistently each signature was reexpressed across phases within each participant. We found that higher trait anxiety was significantly associated with greater cross-phase consistency in subjective-fear signature expression, whereas no significant association was observed for threat-conditioning signature expression. This association became more prominent over longer temporal windows and was not detected in the average time series of the amygdala, ventro-medial prefrontal cortex, insula, or hippocampus. These findings suggest that individual differences in trait anxiety are more closely associated with the consistent cross-phase reexpression of distributed neural patterns related to subjective fear than with the stability of conditioned-threat processing or regional neural dynamics. Taken together, this study shows how predictive brain models combined with longitudinal naturalistic viewing provide a new approach for examining individual differences in trait anxiety across repeated experiences of the same naturalistic contexts. HighlightsO_LIHigher trait anxiety predicted greater cross-phase subjective-fear consistency. C_LIO_LINo association emerged for the threat-conditioning neural signature. C_LIO_LIThe subjective-fear effect strengthened over longer temporal windows. C_LIO_LIThe effect was not detected in amygdala, vmPFC, insula, or hippocampus. C_LIO_LILongitudinal naturalistic fMRI tracked fear signatures across three months. C_LI