Species-specific sensorimotor noise levels explain saccadic suppression strength differences between macaques and zebrafish

Active sensing requires integrating information about self-generated movements and afferent inputs. However, such information is inherently variable and relies on uncertain extrapolations. Saccadic suppression of visual sensitivity is a classic example of sensory motor integration. Here, we used a control theory model based on efficient state estimation to predict that optimal saccadic suppression strength depends on the time-dependent magnitude of unexplained sensory and motor variability. Using two drastically different species, macaques and zebrafish larvae, we experimentally measured saccadic suppression in the superior colliculus and the homologous optic tectum. Species-specific oculomotor noise was quantified experimentally, and sensory noise was estimated. We found that inter-species differences in sensory and motor noise levels, as well as their theoretically predicted impacts on visual sensitivity, are consistent with observed differences in saccadic suppression strength. Our results highlight the importance of computational resource limits of the brain for understanding performance differences that have evolved in homologous brain areas.