Nonlinear Decoding of Natural Images from Large-Scale Primate Retinal Ganglion Recordings

AbstractDecoding sensory stimuli from neural activity can provide insight into how the nervous system might interpret the physical environment, and facilitates the development of brain-machine interfaces. Nevertheless, the neural decoding problem remains a significant open challenge. Here, we present an efficient nonlinear decoding approach for inferring natural scene stimuli from the spiking activities of retinal ganglion cells (RGCs). Our approach uses neural networks to improve upon existing decoders in both accuracy and scalability. Trained and validated on real retinal spike data from > 1000 simultaneously recorded macaque RGC units, the decoder demonstrates the necessity of nonlinear computations for accurate decoding of the fine structures of visual stimuli. Specifically, high-pass spatial features of natural images can only be decoded using nonlinear techniques, while low-pass features can be extracted equally well by linear and nonlinear methods. Together, these results advance the state of the art in decoding natural stimuli from large populations of neurons.Author summaryNeural decoding is a fundamental problem in computational and statistical neuroscience. There is an enormous literature on this problem, applied to a wide variety of brain areas and nervous systems. Here we focus on the problem of decoding visual information from the retina. The bulk of previous work here has focused on simple linear decoders, applied to modest numbers of simultaneously recorded cells, to decode artificial stimuli. In contrast, here we develop a scalable nonlinear decoding method to decode natural images from the responses of over a thousand simultaneously recorded units, and show that this decoder significantly improves on the state of the art.

Download Full-text

Neural Networks for Efficient Bayesian Decoding of Natural Images from Retinal Neurons

10.1101/153759 ◽

2017 ◽

Cited By ~ 10

Author(s):

Nikhil Parthasarathy ◽

Eleanor Batty ◽

William Falcon ◽

Thomas Rutten ◽

Mohit Rajpal ◽

...

Keyword(s):

Neural Networks ◽

Ganglion Cells ◽

Natural Images ◽

Retinal Ganglion ◽

Natural Scene Statistics ◽

Neural Signals ◽

Sensory Stimuli ◽

Convolutional Autoencoder ◽

Approximate Bayesian ◽

Linear Decoding

AbstractDecoding sensory stimuli from neural signals can be used to reveal how we sense our physical environment, and is valuable for the design of brain-machine interfaces. However, existing linear techniques for neural decoding may not fully reveal or exploit the fidelity of the neural signal. Here we develop a new approximate Bayesian method for decoding natural images from the spiking activity of populations of retinal ganglion cells (RGCs). We sidestep known computational challenges with Bayesian inference by exploiting artificial neural networks developed for computer vision, enabling fast nonlinear decoding that incorporates natural scene statistics implicitly. We use a decoder architecture that first linearly reconstructs an image from RGC spikes, then applies a convolutional autoencoder to enhance the image. The resulting decoder, trained on natural images and simulated neural responses, significantly outperforms linear decoding, as well as simple point-wise nonlinear decoding. These results provide a tool for the assessment and optimization of retinal prosthesis technologies, and reveal that the retina may provide a more accurate representation of the visual scene than previously appreciated.

Download Full-text

Multiplexed computations in retinal ganglion cells of a single type

10.1101/080135 ◽

2016 ◽

Author(s):

Stephane Deny ◽

Ulisse Ferrari ◽

Emilie Mace ◽

Pierre Yger ◽

Romain Caplette ◽

...

Keyword(s):

Large Scale ◽

Ganglion Cells ◽

Quantitative Model ◽

Visual Scene ◽

Single Type ◽

Retinal Ganglion ◽

Object Code ◽

Stimulus Feature ◽

Homogeneous Population ◽

Common Assumption

AbstractIn the early visual system, cells of the same type perform the same computation in di↵erent places of the visual field. How these cells code together a complex visual scene is unclear. A common assumption is that cells of the same type will extract a single stimulus feature to form a feature map, but this has rarely been observed directly. Using large-scale recordings in the rat retina, we show that a homogeneous population of fast OFF ganglion cells simultaneously encodes two radically different features of a visual scene. Cells close to a moving object code linearly for its position, while distant cells remain largely invariant to the object’s position and, instead, respond non-linearly to changes in the object’s speed. Cells switch between these two computations depending on the stimulus. We developed a quantitative model that accounts for this effect and identified a likely disinhibitory circuit that mediates it. Ganglion cells of a single type thus do not code for one, but two features simultaneously. This richer, flexible neural map might also be present in other sensory systems.

Download Full-text

Speed-Selectivity in Retinal Ganglion Cells is Modulated by the Complexity of the Visual Stimulus

10.1101/350330 ◽

2018 ◽

Author(s):

César R Ravello ◽

Laurent U Perrinet ◽

María-José Escobar ◽

Adrián G Palacios

Keyword(s):

Visual System ◽

Retinal Ganglion Cells ◽

Visual Stimulus ◽

Motion Detection ◽

Ganglion Cells ◽

Large Body ◽

Natural Images ◽

Retinal Ganglion ◽

Complex Stimuli ◽

Different Populations

ABSTRACTMotion detection represents one of the critical tasks of the visual system and has motivated a large body of research. However, is remain unclear precisely why the response of retinal ganglion cells (RGCs) to simple artificial stimuli does not predict their response to complex naturalistic stimuli. To explore this topic, we use Motion Clouds (MC), which are synthetic textures that preserve properties of natural images and are merely parameterized, in particular by modulating the spatiotemporal spectrum complexity of the stimulus by adjusting the frequency bandwidths. By stimulating the retina of the diurnal rodent,Octodon deguswith MC we show that the RGCs respond to increasingly complex stimuli by narrowing their adjustment curves in response to movement. At the level of the population, complex stimuli produce a sparser code while preserving movement information; therefore, the stimuli are encoded more efficiently. Interestingly, these properties were observed throughout different populations of RGCs. Thus, our results reveal that the response at the level of RGCs is modulated by the naturalness of the stimulus - in particular for motion - which suggests that the tuning to the statistics of natural images already emerges at the level of the retina.

Download Full-text

Synaptic circuits for irradiance coding by intrinsically photosensitive retinal ganglion cells

10.1101/442954 ◽

2018 ◽

Cited By ~ 3

Author(s):

Shai Sabbah ◽

Carin Papendorp ◽

Elizabeth Koplas ◽

Marjo Beltoja ◽

Cameron Etebari ◽

...

Keyword(s):

Retinal Ganglion Cells ◽

Ganglion Cells ◽

Bipolar Cells ◽

Retinal Ganglion ◽

Electron Microscopic ◽

Synaptic Circuits ◽

Ribbon Synapses ◽

High Pass ◽

Excitatory Drive ◽

Rod Bipolar Cells

SummaryWe have explored the synaptic networks responsible for the unique capacity of intrinsically photosensitive retinal ganglion cells (ipRGCs) to encode overall light intensity. This luminance signal is crucial for circadian, pupillary and related reflexive responses light. By combined glutamate-sensor imaging and patch recording of postsynaptic RGCs, we show that the capacity for intensity-encoding is widespread among cone bipolar types, including OFF types.Nonetheless, the bipolar cells that drive ipRGCs appear to carry the strongest luminance signal. By serial electron microscopic reconstruction, we show that Type 6 ON cone bipolar cells are the dominant source of such input, with more modest input from Types 7, 8 and 9 and virtually none from Types 5i, 5o, 5t or rod bipolar cells. In conventional RGCs, the excitatory drive from bipolar cells is high-pass temporally filtered more than it is in ipRGCs. Amacrine-to-bipolar cell feedback seems to contribute surprisingly little to this filtering, implicating mostly postsynaptic mechanisms. Most ipRGCs sample from all bipolar terminals costratifying with their dendrites, but M1 cells avoid all OFF bipolar input and accept only ectopic ribbon synapses from ON cone bipolar axonal shafts. These are remarkable monad synapses, equipped with as many as a dozen ribbons and only one postsynaptic process.

Download Full-text

Pan-Retinal Characterisation of Light Responses from Ganglion Cells in the Developing Mouse Retina

10.1101/050393 ◽

2016 ◽

Author(s):

Gerrit Hilgen ◽

Sahar Pirmoradian ◽

Daniela Pamplona ◽

Pierre Kornprobst ◽

Bruno Cessac ◽

...

Keyword(s):

Large Scale ◽

Ganglion Cells ◽

Mouse Retina ◽

Multielectrode Array ◽

Retinal Ganglion ◽

Developmental Profile ◽

Light Responses ◽

Ecological Requirements ◽

Eye Opening ◽

Dorsal Retina

AbstractWe have investigated the ontogeny of light-driven responses in mouse retinal ganglion cells (RGCs). Using a large-scale, high-density multielectrode array, we recorded from hundreds to thousands of RGCs simultaneously at pan-retinal level, including dorsal and ventral locations. Responses to different contrasts not only revealed a complex developmental profile for ON, OFF and ON-OFF RGC types, but also unveiled differences between dorsal and ventral RGCs. At eye-opening, dorsal RGCs of all types were more responsive to light, perhaps indicating an environmental priority to nest viewing for pre-weaning pups. The developmental profile of ON and OFF RGCs exhibited antagonistic behavior, with the strongest ON responses shortly after eye-opening, followed by an increase in the strength of OFF responses later on. Further, we found that with maturation receptive field (RF) center sizes decrease, responses to light get stronger, and centers become more circular while seeing differences in all of them between RGC types. These findings show that retinal functionality is not spatially homogeneous, likely reflecting ecological requirements that favour the early development of dorsal retina, and reflecting different roles in vision in the mature animal.

Download Full-text

The multiple factors determining retinotopic order in the growth of optic fibres into the optic tectum

Philosophical Transactions of the Royal Society of London Series B Biological Sciences ◽

10.1098/rstb.1977.0041 ◽

1977 ◽

Vol 278 (961) ◽

pp. 261-276 ◽

Cited By ~ 50

Keyword(s):

Optic Tectum ◽

Nervous Tissue ◽

Large Scale ◽

Ganglion Cells ◽

Impulse Activity ◽

Nerve Impulse ◽

Retinal Ganglion ◽

Optic Nerve Fibre ◽

Cells Of Origin ◽

Lower Vertebrates

Evidence is presented to support the conclusion that normally functioning optic nerve fibre terminal arborizations are open to continuous modification of their location and that they are capable of large scale gradual movement across the optic tectum in lower vertebrates. The termination of optic fibres at precisely defined tectal locations during normal embryonic development does not appear, in view of this and other evidence, to be due to any restrictions imposed by specializations distinguishing terminal sites themselves. However, there is clear evidence that, on the basis of possibly very simple specializations acquired as part of their embryological origin at particular locations in the retina, growing optic fibres actively and continuously select specific routes to be followed through intervening nervous tissue which eventually lead them to predictable and at least approximately appropriate terminal regions in the tectum. It is proposed that terminals move into and maintain fully retinotopic order as a result of direct interactions between fibres themselves based on features correlated with the retinal proximity of their cells of origin. This may involve further use of specializations due to related embryological origin: correlations in nerve impulse activity among neighbouring retinal ganglion cells may serve to stabilize most favourable terminal combinations. It is argued that fibres are subject to multiple influences which contribute to their orderly growth and that the demands made on the embryological differentiation of nervous tissue can thereby be considerably reduced.

Download Full-text

Activation of ganglion cells and axon bundles using epiretinal electrical stimulation

Journal of Neurophysiology ◽

10.1152/jn.00750.2016 ◽

2017 ◽

Vol 118 (3) ◽

pp. 1457-1471 ◽

Cited By ~ 27

Author(s):

Lauren E. Grosberg ◽

Karthik Ganesan ◽

Georges A. Goetz ◽

Sasidhar S. Madugula ◽

Nandita Bhaskhar ◽

...

Keyword(s):

Electrical Stimulation ◽

Retinal Ganglion Cells ◽

Large Scale ◽

Ganglion Cells ◽

Propagation Speed ◽

Peripheral Retina ◽

Retinal Ganglion ◽

Central Retina ◽

Axon Bundle ◽

Stimulation Current

Epiretinal prostheses for treating blindness activate axon bundles, causing large, arc-shaped visual percepts that limit the quality of artificial vision. Improving the function of epiretinal prostheses therefore requires understanding and avoiding axon bundle activation. This study introduces a method to detect axon bundle activation on the basis of its electrical signature and uses the method to test whether epiretinal stimulation can directly elicit spikes in individual retinal ganglion cells without activating nearby axon bundles. Combined electrical stimulation and recording from isolated primate retina were performed using a custom multielectrode system (512 electrodes, 10-μm diameter, 60-μm pitch). Axon bundle signals were identified by their bidirectional propagation, speed, and increasing amplitude as a function of stimulation current. The threshold for bundle activation varied across electrodes and retinas, and was in the same range as the threshold for activating retinal ganglion cells near their somas. In the peripheral retina, 45% of electrodes that activated individual ganglion cells (17% of all electrodes) did so without activating bundles. This permitted selective activation of 21% of recorded ganglion cells (7% of expected ganglion cells) over the array. In one recording in the central retina, 75% of electrodes that activated individual ganglion cells (16% of all electrodes) did so without activating bundles. The ability to selectively activate a subset of retinal ganglion cells without axon bundles suggests a possible novel architecture for future epiretinal prostheses. NEW & NOTEWORTHY Large-scale multielectrode recording and stimulation were used to test how selectively retinal ganglion cells can be electrically activated without activating axon bundles. A novel method was developed to identify axon activation on the basis of its unique electrical signature and was used to find that a subset of ganglion cells can be activated at single-cell, single-spike resolution without producing bundle activity in peripheral and central retina. These findings have implications for the development of advanced retinal prostheses.

Download Full-text

Decision letter: Reconstruction of natural images from responses of primate retinal ganglion cells

10.7554/elife.58516.sa1 ◽

2020 ◽

Author(s):

Hiroki Asari

Keyword(s):

Retinal Ganglion Cells ◽

Ganglion Cells ◽

Natural Images ◽

Retinal Ganglion

Download Full-text

Selective activation of ganglion cells without axon bundles using epiretinal electrical stimulation

10.1101/075283 ◽

2016 ◽

Cited By ~ 1

Author(s):

Lauren E. Grosberg ◽

Karthik Ganesan ◽

Georges A. Goetz ◽

Sasidhar Madugula ◽

Nandita Bhaskar ◽

...

Keyword(s):

Electrical Stimulation ◽

Retinal Ganglion Cells ◽

Large Scale ◽

Ganglion Cells ◽

Propagation Speed ◽

Peripheral Retina ◽

Retinal Ganglion ◽

Selective Activation ◽

Central Retina ◽

Axon Bundle

AbstractEpiretinal prostheses for treating blindness activate axon bundles, causing large, arc-shaped visual percepts that limit the quality of artificial vision. Improving the function of epiretinal prostheses therefore requires understanding and avoiding axon bundle activation. This paper introduces a method to detect axon bundle activation based on its electrical signature, and uses the method to test whether epiretinal stimulation can directly elicit spikes in individual retinal ganglion cells without activating nearby axon bundles. Combined electrical stimulation and recording from isolated primate retina were performed using a custom multi-electrode system (512 electrodes, 10 µm diameter, 60 µm pitch). Axon bundle signals were identified by their bi-directional propagation, speed, and increasing amplitude as a function of stimulation current. The threshold for bundle activation varied across electrodes and retinas, and was in the same range as the threshold for activating retinal ganglion cells near their somas. In the peripheral retina, 45% of electrodes that activated individual ganglion cells (17% of all electrodes) did so without activating bundles. This permitted selective activation of 21% of recorded ganglion cells (7% of all ganglion cells) over the array. In the central retina, 75% of electrodes that activated individual ganglion cells (16% of all electrodes) did so without activating bundles. The ability to selectively activate a subset of retinal ganglion cells without axon bundles suggests a possible novel architecture for future epiretinal prostheses.New & NoteworthyLarge-scale multi-electrode recording and stimulation were used to test how selectively retinal ganglion cells can be electrically activated without activating axon bundles. A novel method was developed to identify axon activation based on its unique electrical signature, and used to find that a subset of ganglion cells can be activated at single-cell, single-spike resolution without producing bundle activity, in peripheral and central retina. These findings have implications for the development of advanced retinal prostheses.

Download Full-text