Limits...
High Accuracy Decoding of Dynamical Motion from a Large Retinal Population.

Marre O, Botella-Soler V, Simmons KD, Mora T, Tkačik G, Berry MJ - PLoS Comput. Biol. (2015)

Bottom Line: We show that the bar's position can be reconstructed from retinal activity with a precision in the hyperacuity regime using a linear decoder acting on 100+ cells.The classical view would have suggested that the firing rates of the cells form a moving hill of activity tracking the bar's position.As a result, population redundancy was high, and we could find multiple, disjoint subsets of neurons that encoded the trajectory with high precision.

View Article: PubMed Central - PubMed

Affiliation: Department of Molecular Biology and Neuroscience Institute, Princeton University, Princeton, United States of America; Institut de la Vision, INSERM UMRS 968, UPMC UM 80, CNRS UMR 7210, Paris, France.

ABSTRACT
Motion tracking is a challenge the visual system has to solve by reading out the retinal population. It is still unclear how the information from different neurons can be combined together to estimate the position of an object. Here we recorded a large population of ganglion cells in a dense patch of salamander and guinea pig retinas while displaying a bar moving diffusively. We show that the bar's position can be reconstructed from retinal activity with a precision in the hyperacuity regime using a linear decoder acting on 100+ cells. We then took advantage of this unprecedented precision to explore the spatial structure of the retina's population code. The classical view would have suggested that the firing rates of the cells form a moving hill of activity tracking the bar's position. Instead, we found that most ganglion cells in the salamander fired sparsely and idiosyncratically, so that their neural image did not track the bar. Furthermore, ganglion cell activity spanned an area much larger than predicted by their receptive fields, with cells coding for motion far in their surround. As a result, population redundancy was high, and we could find multiple, disjoint subsets of neurons that encoded the trajectory with high precision. This organization allows for diverse collections of ganglion cells to represent high-accuracy motion information in a form easily read out by downstream neural circuits.

No MeSH data available.


Related in: MedlinePlus

Decoding based on the neural image in the salamander retina.A: Neural image in response to the moving bar. Color plot: neural image of the ganglion cell’s population activity at each point in time. White points: most likely position of the bar inferred from the peak in the neural image. Real trajectory in red. B: Population firing rate summed over all the cells as a function of time, for the same time window than A. C: Prediction of the bar’s trajectory using the linear decoding (black); real trajectory (red).
© Copyright Policy
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC4489022&req=5

pcbi.1004304.g004: Decoding based on the neural image in the salamander retina.A: Neural image in response to the moving bar. Color plot: neural image of the ganglion cell’s population activity at each point in time. White points: most likely position of the bar inferred from the peak in the neural image. Real trajectory in red. B: Population firing rate summed over all the cells as a function of time, for the same time window than A. C: Prediction of the bar’s trajectory using the linear decoding (black); real trajectory (red).

Mentions: We first constructed the “neural image” as in previous studies, where its peak peak corresponded well with the position of the moving bar [24, 25]. The neural image is the spatial pattern of firing in the ganglion cell population as a function of time. We calculated it by plotting the firing rate of the cells as a function of their receptive field position (see Methods). Despite the success of the neural image in tracking simpler object motion [22, 25], we found that it did a poor job with our complex trajectory (CC = 0.06, Fig 4A). Similar results were obtained on the 3 salamander retinas (average performance CC = 0.09 ± 0.04). The performance of the neural image decoder was also poor for the guinea pig retina (CC = 0.17, n = 1).


High Accuracy Decoding of Dynamical Motion from a Large Retinal Population.

Marre O, Botella-Soler V, Simmons KD, Mora T, Tkačik G, Berry MJ - PLoS Comput. Biol. (2015)

Decoding based on the neural image in the salamander retina.A: Neural image in response to the moving bar. Color plot: neural image of the ganglion cell’s population activity at each point in time. White points: most likely position of the bar inferred from the peak in the neural image. Real trajectory in red. B: Population firing rate summed over all the cells as a function of time, for the same time window than A. C: Prediction of the bar’s trajectory using the linear decoding (black); real trajectory (red).
© Copyright Policy
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC4489022&req=5

pcbi.1004304.g004: Decoding based on the neural image in the salamander retina.A: Neural image in response to the moving bar. Color plot: neural image of the ganglion cell’s population activity at each point in time. White points: most likely position of the bar inferred from the peak in the neural image. Real trajectory in red. B: Population firing rate summed over all the cells as a function of time, for the same time window than A. C: Prediction of the bar’s trajectory using the linear decoding (black); real trajectory (red).
Mentions: We first constructed the “neural image” as in previous studies, where its peak peak corresponded well with the position of the moving bar [24, 25]. The neural image is the spatial pattern of firing in the ganglion cell population as a function of time. We calculated it by plotting the firing rate of the cells as a function of their receptive field position (see Methods). Despite the success of the neural image in tracking simpler object motion [22, 25], we found that it did a poor job with our complex trajectory (CC = 0.06, Fig 4A). Similar results were obtained on the 3 salamander retinas (average performance CC = 0.09 ± 0.04). The performance of the neural image decoder was also poor for the guinea pig retina (CC = 0.17, n = 1).

Bottom Line: We show that the bar's position can be reconstructed from retinal activity with a precision in the hyperacuity regime using a linear decoder acting on 100+ cells.The classical view would have suggested that the firing rates of the cells form a moving hill of activity tracking the bar's position.As a result, population redundancy was high, and we could find multiple, disjoint subsets of neurons that encoded the trajectory with high precision.

View Article: PubMed Central - PubMed

Affiliation: Department of Molecular Biology and Neuroscience Institute, Princeton University, Princeton, United States of America; Institut de la Vision, INSERM UMRS 968, UPMC UM 80, CNRS UMR 7210, Paris, France.

ABSTRACT
Motion tracking is a challenge the visual system has to solve by reading out the retinal population. It is still unclear how the information from different neurons can be combined together to estimate the position of an object. Here we recorded a large population of ganglion cells in a dense patch of salamander and guinea pig retinas while displaying a bar moving diffusively. We show that the bar's position can be reconstructed from retinal activity with a precision in the hyperacuity regime using a linear decoder acting on 100+ cells. We then took advantage of this unprecedented precision to explore the spatial structure of the retina's population code. The classical view would have suggested that the firing rates of the cells form a moving hill of activity tracking the bar's position. Instead, we found that most ganglion cells in the salamander fired sparsely and idiosyncratically, so that their neural image did not track the bar. Furthermore, ganglion cell activity spanned an area much larger than predicted by their receptive fields, with cells coding for motion far in their surround. As a result, population redundancy was high, and we could find multiple, disjoint subsets of neurons that encoded the trajectory with high precision. This organization allows for diverse collections of ganglion cells to represent high-accuracy motion information in a form easily read out by downstream neural circuits.

No MeSH data available.


Related in: MedlinePlus