scholarly journals Revealing structure components of the retina by deep learning networks

2017 ◽  
Author(s):  
Qi Yan ◽  
Zhaofei Yu ◽  
Feng Chen ◽  
Jian K. Liu

AbstractDeep convolutional neural networks (CNNs) have demonstrated impressive performance on visual object classification tasks. In addition, it is a useful model for predication of neuronal responses recorded in visual system. However, there is still no clear understanding of what CNNs learn in terms of visual neuronal circuits. Visualizing CNN’s features to obtain possible connections to neuronscience under-pinnings is not easy due to highly complex circuits from the retina to higher visual cortex. Here we address this issue by focusing on single retinal ganglion cells with a simple model and electrophysiological recordings from salamanders. By training CNNs with white noise images to predicate neural responses, we found that convolutional filters learned in the end are resembling to biological components of the retinal circuit. Features represented by these filters tile the space of conventional receptive field of retinal ganglion cells. These results suggest that CNN could be used to reveal structure components of neuronal circuits.

2020 ◽  
Author(s):  
Naïg Aurélia Ludmilla Chenais ◽  
Marta Jole Ildelfonsa Airaghi Leccardi ◽  
Diego Ghezzi

AbstractObjectiveRetinal stimulation in blind patients evokes the sensation of discrete points of light called phosphenes, which allows them performing visual guided tasks, such as orientation, navigation, object recognition, object manipulation and reading. However, the clinical benefit of artificial vision in profoundly blind patients is still tenuous, as several engineering and biophysical obstacles keep it away from natural perception. The relative preservation of the inner retinal neurons in hereditary degenerative retinal diseases, such as retinitis pigmentosa, supports artificial vision through the network-mediated stimulation of retinal ganglion cells. However, the response of retinal ganglion cells to repeated electrical stimulation rapidly declines, primarily because of the intrinsic desensitisation of their excitatory network. In patients, upon repetitive stimulation, phosphenes fade out in less than half of a second, which drastically limits the understanding of the percept.ApproachA more naturalistic stimulation strategy, based on spatiotemporal modulation of electric pulses, could overcome the desensitisation of retinal ganglion cells. To investigate this hypothesis, we performed network-mediated epiretinal stimulations paired to electrophysiological recordings in retinas explanted from both male and female retinal degeneration 10 mice.Main resultsThe results showed that the spatial and temporal modulation of the network-mediated epiretinal stimulation prolonged the responsivity of retinal ganglion cells from 400 ms up to 4.2 s.SignificanceA time-varied, non-stationary and interrupted stimulation of the retinal network, mimicking involuntary microsaccades, might reduce the fading of the visual percept and improve the clinical efficacy of retinal implants.


2021 ◽  
Vol 2021 (11) ◽  
pp. pdb.prot106898 ◽  
Author(s):  
Kara G. Pratt

The Xenopus tadpole retinotectal projection is the main component of the amphibian visual system. It comprises the retinal ganglion cells (RGCs) in the eye, which project an axon to synapse onto tectal neurons in the optic tectum. There are many attributes of this relatively simple projection that render it uniquely well-suited for studying the functional development of neural circuits. One major experimental advantage of this circuit is that it can be genetically or pharmacologically altered and then assessed at high resolution via whole-cell electrophysiological recordings using an ex vivo isolated brain preparation. This protocol provides instructions for performing such electrophysiological investigations using the ex-vivo-isolated brain preparation. It allows one to measure many different aspects of synaptic transmission between the RGC axons and individual postsynaptic tectal neurons, including AMPA (α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid) to NMDA (N-methyl-d-aspartate) ratios, strength of individual RGC axons, paired pulse facilitation, and strength of individual synapses.


Author(s):  
Kyril I. Kuznetsov ◽  
Vitaliy Yu. Maslov ◽  
Svetlana A. Fedulova ◽  
Nikolai S. Veselovsky

Sign in / Sign up

Export Citation Format

Share Document