In many non-mammalian vertebrate species, retinal rod photoreceptors have been found to couple extensively to one another via gap junctions. Yet the functional role of coupling remains enigmatic. Coupling could enhance night vision by circumventing saturation at the rod output synapse, but it could also reduce absolute sensitivity by rendering the output synapse less effective at separating photon signals from intrinsic rod noise. Recent results suggest that rodent rods are also coupled, but rod circuitry differs significantly between mammals and other vertebrates, so existing data on non-mammalian coupling may not be applicable to understanding mammalian vision; the rod to rod-bipolar pathway is a mammalian specialization, and mammalian rod networks appear distinctive in their limited extent of coupling, and in the type of connexin proteins underlying the junctions.
Here, I show that primate rods are also coupled, suggesting a role for coupling in human vision. I then present more detailed data on rod coupling from guinea pig retina, showing that guinea pig rod-rod junctional conductance is about 350 pS. By developing a rod network model to analyze the combined primate and guinea pig results, I confirm that the junctional conductance in guinea pig and primate is comparable. Finally, I model the effects of coupling on human visual detection, focusing particularly on the interaction between rod coupling and the nonlinear operations of the rod output synapse. Based on primate data alone, the model confirms that coupling can have competing beneficial and detrimental effects on vision depending on the strength of coupling and the parameters of the rod output synapse, as well as the spatial configuration of the stimulus. Incorporating the guinea pig data reveals that the strength of coupling in mammals is appropriate for mitigating detrimental effects on absolute detection while enhancing benefits for detection of small spots.