Life threatening fear after a single exposure evolves in a subset of vulnerable individuals to anxiety, which may persist for their lifetime. Yet neither the whole brain's response to innate acute fear nor how brain activity evolves over time is known. Sustained neuronal activity may be a factor in the development of a persistent fear response. We couple two experimental protocols to provoke acute fear leading to prolonged fear: Predator stress (PS), a naturalistic approach to induce fear in rodents; and Serotonin transporter knockout mouse (SERT-KO) that responds to PS with sustained defensive behavior. Behavior was monitored before, during and at short and long times after PS in wild type (WT) and SERT-KO mice. Both genotypes responded to PS with defensive behavior. SERT-KO retained defensive behavior for 23 days, while WT mice returned to baseline exploratory behavior by 9 days. Thus, differences in neural activity between WT and SERT-KO 9 days after PS identifies neural correlates of persistent defensive behavior, in mice. We used longitudinal manganese-enhanced magnetic resonance imaging (MEMRI) to identify brain-wide neural activity associated with different behaviors. Mn2+ accumulation in active neurons occurs in awake, behaving mice and is retrospectively imaged. Following the same two cohorts of mice, WT and SERT-KO, longitudinally allowed unbiased quantitative comparisons of brain-wide activity by statistical parametric mapping (SPM). During natural behavior in WT, only low levels of activity-induced Mn2+-accumulation were detected, while much more accumulation appeared immediately after PS in both WT and SERT-KO, and evolved at 9 days to a new activity pattern (p < 0.0001, uncorr., T = 5.4). Patterns of accumulation differed between genotypes, with more regions of the brain and larger volumes within regions involved in SERT-KO than WT. A new computational segmentation analysis, using our InVivo Atlas based on a manganese-enhanced MR image of a living mouse, revealed dynamic changes in the volume of significantly enhanced voxels within each segment that differed between genotypes across 45 of 87 segmented regions. At Day 9 after PS, the striatum and ventral pallidum were active in both genotypes but more so in the SERT-KO. SERT-KO also displayed sustained or increased volume of Mn2+ accumulations between Post-Fear and Day 9 in eight segments where activity was decreased or silenced in WT. C-fos staining, an alternative neural activity marker, of brains from the same mice fixed at conclusion of imaging sessions confirmed that MEMRI detected active neurons. Intensity measurements in 12 regions of interest (ROIs) supported the SPM results. Between group comparisons by SPM and of ROI measurements identified specific regions differing between time points and genotypes. We report brain-wide activity in response to a single exposure of acute fear, and, for the first time, its evolution to new activity patterns over time in individuals vulnerable to persistent fear. Our results show multiple regions with dynamic changes in neural activity and that the balance of activity between segments is disordered in the SERT-KO. Thus, longitudinal MEMRI represents a powerful approach to discover how brain-wide activity evolves from the natural state either after an experience or during a disease process.