Simulation Before Vision: Interoceptive Origins of Offline Simulation and the Evolution of Mental Imagery

How did mental imagery evolve, and what adaptive problem did it solve? Research tradition has carried an implicit and largely unexamined assumption that mental imagery evolved from sensory systems. Using visual imagery as my primary case, I argue instead that its origins lie not in sensory cortex but in interoceptive systems. A simulation system with no means of evaluating outcomes has no basis for selecting between them; interoception is what converts predicted outcomes into reasons to act. Drawing on comparative neuroanatomy, paleontological evidence, and the cognitive neuroscience of offline simulation, I argue that this interoceptive architecture was established before elaborated visual cortex existed, with motor simulation constituting its earliest form. During the nocturnal bottleneck (a prolonged period during which early mammals had substantially reduced visual systems) the simulation architecture matured under olfactory and interoceptive dominance, independently of high-acuity vision. Vision assumed its dominant role only after this bottleneck, recruited to solve a discriminability problem that coarse interoceptive signals alone could not resolve. This synthesis reconceptualises mental imagery not as an extension of visual perception but as a component of a broader simulation system whose organisational logic was shaped from the outset by the demands of predicting and evaluating bodily states.