Systematic reduction of the dimensionality of natural scenes allows accurate predictions of retinal ganglion cell spike outputs

Abstract

The mammalian retina engages a broad array of linear and nonlinear circuit mechanisms to convert natural scenes into retinal ganglion cell (RGC) spike outputs. Although many individual integration mechanisms are well understood, we know less about how multiple mechanisms interact to encode the complex spatial features present in natural inputs. Here, we identified key spatial features in natural scenes that shape encoding by primate parasol RGCs. Our approach identified simplifications in the spatial structure of natural scenes that minimally altered RGC spike responses. We observed that reducing natural movies into 16 linearly integrated regions described ∼80% of the structure of parasol RGC spike responses; this performance depended on the number of regions but not their precise spatial locations. We used simplified stimuli to design high-dimensional metamers that recapitulated responses to naturalistic movies. Finally, we modeled the retinal computations that convert flashed natural images into one-dimensional spike counts.

Natural stimuli drive concerted nonlinear responses in populations of retinal ganglion cells

“…Interestingly, ON parasol cells showed more rectified nonlinearities compared to OFF, opposite of what is expected from finding in the macaque retina (Turner and Rieke, 2016). Midget cells also showed substantial rectification, consistent with findings in the peripheral macaque retina (Freedland and Rieke, 2022; Freeman et al, 2015). Cell-type-specific differences in nonlinear components were also prominent in the mouse retina ( Fig.…”

Functional Decomposition of Retinal Ganglion Cell Receptive Fields

“…Specifically OFF parasol cells have been reported to respond to finely structured natural stimuli not captured with the LN model (Turner and Rieke 2016). Similar observations have been made in midget cells that were described better with a nonlinear receptive field (Freedland and Rieke 2022;J. Freeman et al 2015).…”

“…Parasol ganglion cells in the primate retina have been linked to nonlinear Y-cell physiology (Crook et al 2008a,b;Petrusca et al 2007) and their function was successfully described with subunit models Turner and Rieke 2016). Nonlinear responses in midget ganglion cells are less pronounced (Cafaro and Rieke 2013;Petrusca et al 2007), but were also modeled successfully (Freedland and Rieke 2022;J. Freeman et al 2015).…”

Temporal prediction captures retinal spiking responses across animal species

“…We compared the spiking activity of the model units to RGC activity across different animal species in response to natural images and movies. Here, we used publicly available recordings from the retina in macaque, marmoset, mouse and salamander [35][36][37]. We found our model to exhibit similar spike activity to the RGCs across each of these species.…”