Reafferent origins of image-forming vision

We propose that reafference was central to the evolution of image-forming vision. Early neural animals used copies of their own motor commands to predict the optic flow generated by self-motion and suppress the expected component in retinal input. In a closed sensorimotor loop combining photoreception, locomotion, and reafferent feedback, spatial luminance gradients were converted into temporally structured signals. The residual differences yielded motion cues for stabilizing orientation and gaze. Short-term plasticity mechanisms tuned synaptic circuits to prevailing speed and illumination, so that incremental optical refinements paid off without necessitating object recognition. We show the feasibility of this hypothesis by constructing a minimal model: a correlation-type motion detector coupled to a ~20 -photoreceptor compound eye, and show that under dim benthic conditions, the model reliably extracts optic-flow cues. Because the required mechanisms − contrast extraction, temporal integration, and reafference − are widespread across neural metazoans, our account is consistent with the assumption that vison emerged convergently in several Cambrian lineages. On this view, reafferent control over optic flow was the key innovation that made light a usable sense for directed navigation; progressive improvements in circuits, wiring, and optics then transformed residual contrast maps (“proto-images”) into spatial vision. We outline phylogenetically grounded, testable predictions about efferent photoreceptor control, latency matching, and polarity-specific motion channels in early-branching taxa and interpret early visual loops in information-theoretic terms.