Retinal circuitry and human vision

The daunting, yet fascinating complexity of the central nervous system, embodied by diverse cell types, circuits, parallel pathways and neural computations is beautifully embodied in the neural retina at the start of the visual process. The retina comprises on the order of 100 cell types; these cell types create functionally diverse circuits that give rise to, by one recent estimate, 30-50 parallel pathways to the brain’s visual processing structures.

Our long-term goal has been to advance understanding of the origins of these pathways in the retina and the underlying neural mechanisms that can create both color coding circuits and motion sensitive direction selective circuits from the same basic neural substrate. The non-human primate provides a perfect model for the human retina, distinguished from other mammals by its specialized foveal structure and trichromatic color code. We developed the first in vitro preparation of the macaque monkey retina that permitted the targeting of primate retinal circuits with single cell intracellular physiology.

Today our research program is focused on two parallel tracks. First, we have identified in the primate the key cell types and circuits that encode the direction of visual motion. To probe all aspects of this circuitry we are utilizing novel visual stimuli, 2-photon calcium imaging of dendritic signals, the voltage clamp to measure light-evoked synaptic currents, electron microscopic reconstruction to identify critical synaptic motifs and finally computational and compartmental modeling of biologically realistic circuits.

Second, to complement our physiological studies in the in vitro macaque retina we have joined an exciting a team of clinicians and histopathologists to apply new methods of ‘volume’ electron microscopy – connectomics – to reconstruct the complete circuitry of the human fovea for the first time. Our approach combines recent advances in electron microscopy, handling big datasets, and artificial intelligence approaches to automate interrogation of complex circuits. An exciting element of this line of investigation is that we are including non-neuronal retinal cell types that are critical to understanding cellular changes that occur in blinding retinal diseases that attack the human fovea.