Efficient coding of natural scenes improves neural system identification DOI Creative Commons
Yongrong Qiu,

David Klindt,

Klaudia P. Szatko

et al.

PLoS Computational Biology, Journal Year: 2023, Volume and Issue: 19(4), P. e1011037 - e1011037

Published: April 24, 2023

Neural system identification aims at learning the response function of neurons to arbitrary stimuli using experimentally recorded data, but typically does not leverage normative principles such as efficient coding natural environments. Visual systems, however, have evolved efficiently process input from environment. Here, we present a network regularization for models by incorporating, regularizer, hypothesis, which states that neural properties sensory representations are strongly shaped need preserve most stimulus information with limited resources. Using this approach, explored if model can be improved sharing its convolutional filters those an autoencoder encode stimuli. To end, built hybrid predict responses retinal noise This approach did only yield higher performance than "stand-alone" model, it also produced more biologically plausible filters, meaning they closely resembled representation in early visual systems. We found these results applied different artificial and across architectures. Moreover, our normatively regularized performed particularly well predicting direction-of-motion sensitive neurons. The benefit scene statistics became marginal, movies. In summary, indicate encoding environmental inputs improve models, least stimuli, point probing naturalistic

Language: Английский

Ancestral photoreceptor diversity as the basis of visual behaviour DOI
Tom Baden

Nature Ecology & Evolution, Journal Year: 2024, Volume and Issue: 8(3), P. 374 - 386

Published: Jan. 22, 2024

Language: Английский

Citations

21

Neurobehavioral meaning of pupil size DOI Creative Commons
Nikola Grujic, Rafael Polanìa, Denis Burdakov

et al.

Neuron, Journal Year: 2024, Volume and Issue: 112(20), P. 3381 - 3395

Published: June 25, 2024

Pupil size is a widely used metric of brain state. It one the few signals originating from that can be readily monitored with low-cost devices in basic science, clinical, and home settings. is, therefore, important to investigate generate well-defined theories related specific interpretations this metric. What exactly does it tell us about brain? Pupils constrict response light dilate during darkness, but also controls pupil irrespective luminosity. fluctuations resulting ongoing "brain states" are as arousal, what pupil-linked arousal how should interpreted neural, cognitive, computational terms? Here, we discuss some recent findings these issues. We identify open questions propose answer them through combination tasks, neurocomputational models, neurophysiological probing interconnected loops causes consequences size.

Language: Английский

Citations

20

A three-photon head-mounted microscope for imaging all layers of visual cortex in freely moving mice DOI Creative Commons

Alexandr Klioutchnikov,

Damian J. Wallace,

J Sawiński

et al.

Nature Methods, Journal Year: 2022, Volume and Issue: 20(4), P. 610 - 616

Published: Nov. 28, 2022

Abstract Advances in head-mounted microscopes have enabled imaging of neuronal activity using genetic tools freely moving mice but these are restricted to recording minimally lit arenas and upper cortical layers. Here we built a 2-g, three-photon excitation-based microscope, containing z -drive that access all layers while behaved fully environment. The microscope had on-board photon detectors, robust environmental light, the arena lighting was timed end each line-scan, enabling functional from layer 4 6 neurons expressing jGCaMP7f roaming or dark arena. By comparing measured populations show is differentially modulated by conditions during free exploration.

Language: Английский

Citations

57

High-density electrode recordings reveal strong and specific connections between retinal ganglion cells and midbrain neurons DOI Creative Commons
Jérémie Sibille, Carolin Gehr, Jonathan I. Benichov

et al.

Nature Communications, Journal Year: 2022, Volume and Issue: 13(1)

Published: Sept. 5, 2022

Abstract The superior colliculus is a midbrain structure that plays important roles in visually guided behaviors mammals. Neurons the receive inputs from retinal ganglion cells but how these are integrated vivo unknown. Here, we discovered high-density electrodes simultaneously capture activity of axons and their postsynaptic target neurons colliculus, vivo. We show cell mouse provide single precise representation retina as input to colliculus. This isomorphic mapping builds scaffold for retinotopic wiring functionally specific connection strength. Our methods broadly applicable, which demonstrate by recording optic tectum zebra finches. find common rules mice finches visual world encoded connections retinorecipient areas.

Language: Английский

Citations

53

State-dependent pupil dilation rapidly shifts visual feature selectivity DOI
Katrin Franke, Konstantin F. Willeke, Kayla Ponder

et al.

Nature, Journal Year: 2022, Volume and Issue: 610(7930), P. 128 - 134

Published: Sept. 28, 2022

Language: Английский

Citations

47

From water to land: Evolution of photoreceptor circuits for vision in air DOI Creative Commons
Tom Baden

PLoS Biology, Journal Year: 2024, Volume and Issue: 22(1), P. e3002422 - e3002422

Published: Jan. 22, 2024

When vertebrates first conquered the land, they encountered a visual world that was radically distinct from of their aquatic ancestors. Fish exploit strong wavelength-dependent interactions light with water by differentially feeding signals up to 5 spectral photoreceptor types into behavioural programmes. However, above same rules do not apply, and this called for an update circuit strategies. Early tetrapods soon evolved double cone, still poorly understood pair new photoreceptors brought “ancestral terrestrial” complement 7. Subsequent nonmammalian lineages adapted highly parallelised retinal input strategy diverse ecologies. By contrast, mammals shed most ancestral converged on is exceptionally general. In eutherian including in humans, parallelisation emerges gradually as signal traverses layers retina brain.

Language: Английский

Citations

14

Recommendations for measuring and standardizing light for laboratory mammals to improve welfare and reproducibility in animal research DOI Creative Commons
Robert J. Lucas, Annette E. Allen, George C. Brainard

et al.

PLoS Biology, Journal Year: 2024, Volume and Issue: 22(3), P. e3002535 - e3002535

Published: March 12, 2024

Light enables vision and exerts widespread effects on physiology behavior, including regulating circadian rhythms, sleep, hormone synthesis, affective state, cognitive processes. Appropriate lighting in animal facilities may support welfare ensure that animals enter experiments an appropriate physiological behavioral state. Furthermore, proper consideration of light during experimentation is important both when it explicitly employed as independent variable a general feature the environment. This Consensus View discusses metrics to use for quantification nonhuman mammals their application improve quality research. It provides methods measuring these metrics, practical guidance implementation husbandry experimentation, quantitative exposure laboratory mammals. The provided has potential data contribute reduction refinement, helping more ethical use.

Language: Английский

Citations

14

Efficient coding of natural images in the mouse visual cortex DOI Creative Commons
Federico Bolaños, Javier G. Orlandi, Ryo Aoki

et al.

Nature Communications, Journal Year: 2024, Volume and Issue: 15(1)

Published: March 19, 2024

How the activity of neurons gives rise to natural vision remains a matter intense investigation. The mid-level visual areas along ventral stream are selective common class images-textures-but circuit-level understanding this selectivity and its link perception unclear. We addressed these questions in mice, first showing that they can perceptually discriminate between textures statistically simpler spectrally matched stimuli, texture types. Then, at neural level, we found secondary area (LM) exhibited higher degree for compared primary (V1). Furthermore, were represented distinct subspaces whose relative distances correlate with statistical similarity images mice's ability them. Notably, dependencies more pronounced LM, where texture-related smaller than V1, resulting superior stimulus decoding capabilities. Together, our results demonstrate finding linking framework statistics, representations, perceptual sensitivity-a hallmark efficient coding computations.

Language: Английский

Citations

10

Ancestral circuits for vertebrate color vision emerge at the first retinal synapse DOI Creative Commons
Takeshi Yoshimatsu, Philipp Bartel, Cornelius Schröder

et al.

Science Advances, Journal Year: 2021, Volume and Issue: 7(42)

Published: Oct. 13, 2021

In vivo recordings of cone photoreceptor outputs in a tetrachromate reveal efficient ancestral strategy for color processing.

Language: Английский

Citations

44

Panoramic visual statistics shape retina-wide organization of receptive fields DOI Creative Commons
Divyansh Gupta, Wiktor Młynarski, Anton Sumser

et al.

Nature Neuroscience, Journal Year: 2023, Volume and Issue: 26(4), P. 606 - 614

Published: March 23, 2023

Abstract Statistics of natural scenes are not uniform—their structure varies dramatically from ground to sky. It remains unknown whether these nonuniformities reflected in the large-scale organization early visual system and what benefits such adaptations would confer. Here, by relying on efficient coding hypothesis, we predict that changes receptive fields across space increase efficiency sensory coding. Using mouse ( Mus musculus ) as a model species, show retinal ganglion cells change their shape along dorsoventral axis, with marked surround asymmetry at horizon, agreement our predictions. Our work demonstrates that, according principles coding, panoramic is exploited retina cell types.

Language: Английский

Citations

23