In the past decades, functional MRI research has investigated task processing in largely static fashion based on evoked responses during blocked and event-related designs. Despite some progress in naturalistic designs, our understanding of threat processing remains largely limited to those obtained with standard paradigms with limited dynamics. In the present work, we applied Switching Linear Dynamical Systems (SLDS), a framework that estimates a generative model of the data with both states and transitions like HMMs, to uncover the dynamics of threat processing during a continuous threat-of-shock paradigm. Importantly, unlike studies in systems neuroscience that frequently assume that systems are decoupled from external inputs, we characterized both endogenous and exogenous contributions to dynamics. First, we demonstrated that the SLDS model learned the regularities of the experimental paradigm, such that states and state transitions estimated from fMRI time series data from 85 regions of interest reflected key properties of the experimental stimuli as well as threat-related processing. Second, we characterized the dynamics of the states and their transitions. The results revealed that threat processing benefits from being viewed in terms of dynamic multivariate patterns whose trajectories are a combination of intrinsic and extrinsic factors, attractor dynamics and stimulus induced perturbations respectively, that jointly determine how the brain temporally evolves during dynamic threat. Finally, we investigated the generalizability of the modeling approach. The successful application of the SLDS model, trained on one paradigm to a separate experiment illustrates the potential of this approach to uncover generalizable neural dynamics of threat that could extend to real-world situations. We believe our work showcases the value of combining computational approaches like SLDS with fMRI to reveal hidden brain dynamics not unveiled with standard experimental designs and analyses, which could inspire wider adoption of these tools in cognitive neuroscience.