Retinal receptive-field substructure: scaffolding for coding and computation

Zapp, Sören J.; Nitsche, Steffen; Gollisch, Tim

doi:10.1016/j.tins.2022.03.005

Cited by 16 publications

(12 citation statements)

References 113 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…These subunits are thought to correspond to bipolar cells that provide excitatory input to ganglion cells (Demb et al, 2001; Liu et al, 2017). Fitting the parameters of a nonlinear subunit model to spiking data is still an ongoing challenge (Liu et al, 2017; Maheswaranathan et al, 2018; Shah et al, 2020; Zapp et al, 2022). We therefore developed a new approach to fit subunit models to data, which we call the subunit grid model ( Fig.…”

Section: The Nonlinear Receptive Field Captures Responses To Natural ...mentioning

confidence: 99%

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

Karamanlis

Khani

Schreyer

et al. 2023

Preprint

Self Cite

View full text Add to dashboard Cite

The role of the vertebrate retina in early vision is generally described by the efficient coding theory, which predicts that the retina discards spatiotemporal correlations in natural scenes. It is unclear, however, whether the predicted decorrelation in the activity of ganglion cells, the retina's output neurons, holds under gaze shifts, which dominate the natural visual input. We here show that species-specific gaze patterns in natural stimuli can drive strong and correlated spiking responses both within and across distinct types of ganglion cells in marmoset as well as mouse retina. These concerted responses violate efficient coding and signal fixation periods with locally high spatial contrast. Finally, novel model-based analyses of ganglion cell responses to natural stimuli reveal that the observed response correlations follow from nonlinear pooling of ganglion cell inputs. Our results reveal how concerted population activity can surpass efficient coding to detect gaze-related stimulus features.

show abstract

Section: The Nonlinear Receptive Field Captures Responses To Natural ...mentioning

confidence: 99%

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

Karamanlis

Khani

Schreyer

et al. 2023

Preprint

Self Cite

View full text Add to dashboard Cite

show abstract

“…A limitation of the present study is that we restricted our analysis to the use of full-field stimuli without any spatial structure. We made this choice in order to limit the number of free parameters in the model as well as to allow comparison between cell types and models independent of the how well the spatial filtering and potential nonlinearities in spatial stimulus integration [20][21][22][23] are captured. On the downside, the use of full-field stimuli makes our analysis insensitive to the spatio-temporal structure of suppression, in particular by not providing for a dedicated suppressive signal pathway from the receptive-field surround.…”

Section: Discussionmentioning

confidence: 99%

Filter-based models of suppression in retinal ganglion cells: comparison and generalization across species and stimuli

Shahidi

Rozenblit

Khani

et al. 2022

Preprint

Self Cite

View full text Add to dashboard Cite

The dichotomy of excitation and suppression is one of the canonical mechanisms explaining the complexity of the neural activity. Computational models of the interplay of excitation and suppression in single neurons aim at investigating how this interaction affects a neuron's spiking responses and shapes, for example, the encoding of sensory stimuli. Here, we compare the performance of three filter-based stimulus-encoding models in predicting retinal ganglion cell responses recorded from axolotl, mouse, and marmoset retina to different types of temporally varying visual stimuli. Suppression in these models is implemented via subtractive or divisive interactions of stimulus filters or by a response-driven feedback module. For the majority of ganglion cells, the subtractive and divisive models perform similarly and outperform the feedback model as well as a linear-nonlinear (LN) model with no suppression. Comparison between the subtractive and the divisive model depended on cell type, species, and stimulus components, with the divisive model generalizing best across temporal stimulus frequencies and visual contrast and the subtractive model capturing in particular responses for slow temporal stimulus dynamics and for slow axolotl cells. Overall, we conclude that the divisive and subtractive models are well suited for capturing interactions of excitation and suppression in ganglion cells and emphasize different temporal regimes of these interactions.

show abstract

“…Convolution recurrent neural networks (CRNN) could better predict the neural response and reveal the corresponding biological features of the retina. And Receptive filed of ganglion cells is a remarkable biological feature [169][170][171]. Some studies compared the spatial and temporal filters of CRNN with that of CNN and recorded receptive field of retinal ganglion cells [167].…”

Section: Rnnmentioning

confidence: 99%

Artificial intelligence techniques for retinal prostheses: a comprehensive review and future direction

Wang

Fang

Zou³

et al. 2023

J. Neural Eng.

View full text Add to dashboard Cite

Abstract. Objective. Retinal prostheses are promising devices to restore vision for patients with severe AMD or RP disease. The visual processing mechanism embodied in retinal prostheses play an important role in the restoration effect. Its performance depends on our understanding of the retina’s working mechanism and the evolvement of computer vision models. Recently, remarkable progress has been made in the field of processing algorithm for retinal prostheses where the new discovery of the retina’s working principle and state-of-the-arts computer vision models are combined together. Approach. We investigated the related research on artificial intelligence techniques for retinal prostheses. The processing algorithm in these studies could be attributed to three types: computer vision-related methods, biophysical models, and deep learning models. Main results. In this review, we first illustrate the structure and function of the normal and degenerated retina, then demonstrate the vision rehabilitation mechanism of three representative retinal prostheses. It is necessary to summarize the computational frameworks abstracted from the normal retina. In addition, the development and feature of three types of different processing algorithms are summarized. Finally, we analyze the bottleneck in existing algorithms and propose our prospect about the future directions to improve the restoration effect. Significance. This review systematically summarizes existing processing models for predicting the response of the retina to external stimuli. What’s more, the suggestions for future direction may inspire researchers in this field to design better algorithms for retinal prostheses.

show abstract

Retinal receptive-field substructure: scaffolding for coding and computation

Cited by 16 publications

References 113 publications

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

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

Filter-based models of suppression in retinal ganglion cells: comparison and generalization across species and stimuli

Artificial intelligence techniques for retinal prostheses: a comprehensive review and future direction

Contact Info

Product

Resources

About