Parallel Computations in Insect and Mammalian Visual Motion Processing

Abstract

Sensory systems use receptors to extract information from the environment and neural circuits to perform subsequent computations. These computations may be described as algorithms composed of sequential mathematical operations. Comparing these operations across taxa reveals how different neural circuits have evolved to solve the same problem, even when using different mechanisms to implement the underlying math. In this review, we compare how insect and mammalian neural circuits have solved the problem of motion estimation, focusing on the fruit fly Drosophila and the mouse retina. Although the two systems implement computations with grossly different anatomy and molecular mechanisms, the underlying circuits transform light into motion signals with strikingly similar processing steps. These similarities run from photoreceptor gain control and spatiotemporal tuning to ON and OFF pathway structures, motion detection, and computed motion signals. The parallels between the two systems suggest that a limited set of algorithms for estimating motion satisfies both the needs of sighted creatures and the constraints imposed on them by metabolism, anatomy, and the structure and regularities of the visual world.

The Neuronal Basis of an Illusory Motion Percept Is Explained by Decorrelation of Parallel Motion Pathways

“…There exist deep parallels between fly visual circuits and mouse retinal circuits, spanning anatomy [60], neural circuits [4], and the sequence of algorithmic steps in motion detection [5]. These parallels suggest that the phi and reverse-phi complementarity we observed here could also exist in vertebrate motion detection systems.…”

“…In many circuits, including those in mouse retina and in flies, two parallel pathways encode motion in the horizontal plane [4, 5]. These two parallel motion detection pathways respond strongly to either light or dark moving edges.…”

Neural computations combine low- and high-order motion cues similarly, in dragonfly and monkey

“…Consistent with the EMD model, this computation involves a nonlinear interaction of a pair of luminance signals separated in space and time. An analogous nonlinear interaction of local luminance signals is also present in mammalian retina 31,32 , resulting in directionally-selective retinal ganglion cells. In both cases, these local interactions are effectively modeled as a multiplication, and the challenge of implementing a signed multiplication in neural circuitry is met in both cases by a separation of signals into ON and OFF channels 3,30,31,33,34 .…”

“…It is notable that the parallels we found are present at the level of circuitry (i.e., the way that individual cells respond to different kinds of motion signals), and not just at the level of the overall computation performed. The latter might be expected based on normative considerations 5,31 , but the former is rather remarkable, given these species' overwhelming differences in brain size, neuron number, and organizational plan.…”

Spatio-temporal relays control layer identity of direction-selective neuron subtypes inDrosophila

“…30 . For GPC-specific wg{KO;FRT NRT-wg FRT wg + }allele switching rescue experiments 30 , wg{KO;FRT NRT-wg FRT wg + }/GlaBc; UAS-FLP flies were crossed to (1) wg{KO;NRT-wg}/GlaBc 30 , and (2) wg{KO;NRT-wg}/GlaBc; R46E01-Gal4 and maintained at 25°C. For gain-of-function experiments, UAS-cd8-GFP; h 1J3 -Gal4 was crossed to UASarm S10 ; dpp-lacZ Exel.2 (ref.…”