The ability to switch between different behaviors is essential to all animals’ survival. Behavior selection is influenced by sensory inputs, internal states, and memory, suggesting that neural mechanisms across the brain coordinate these decisions. While investigating the neural basis of behavioral switching requires whole-brain imaging, current methods face challenges in behaving animals. We employed functional ultrasound imaging (fUS) to record large-scale neural activity in head-fixed mice while simultaneously tracking their behavior. Our aim was to identify brain regions that predict transitions of behavior occurring in the absence of external triggers. To do so, we used the Virtual Burrow Assay, where head-fixed mice can voluntarily egress from an air-floating tube, and a running wheel setup allowing spontaneous initiation of running. We found that mice robustly exhibit distinct, uninstructed behavioral states in these assays, including egress/running, whisking, inactivity, and grooming. Employing brain-wide fUS, we subsequently observed activity patterns associated with these distinct behavioral states and performed whole-brain time-resolved decoding around behavioral transitions. Remarkably, our results revealed that whole-brain activity can predict spontaneous egress and running seconds before its onset, indicating that an uninstructed change in behavior is preceded by a detectable change in brain state. Furthermore, region-wise decoding revealed specific brain areas driving the prediction of behavioral transitions. Using optogenetics we found that inhibiting these regions increases the probability of being in an active state. Through this unbiased approach, our work sheds light on the whole-brain changes preceding transitions of behaviors.