Vision: Depth perception in climbing mice

Depth perception helps animals interact with a three-dimensional world. A new study presents novel paradigm for studying depth in naturally climbing mice and links their behavior to binocular disparity signals primary visual cortical neurons. We perceive the world around us as three-dimensional, even though projections of objects on retinas our eyes are two-dimensional. How does brain infer from two-dimensional retinal images? One major cue is provided by slightly different views two eyes. The eyes’ horizontal separation creates small spatial offsets — known disparities images, enabling distance (Figure 1A). Binocular has been investigated detail species forward-facing eyes, such humans non-human primates. In primates, cortex contains neurons that receive inputs respond preferentially specific disparities1Parker A.J. cerebral cortex.Nat. Rev. Neurosci. 2007; 8: 379-391Crossref PubMed Scopus (275) Google Scholar. encode also found mouse cortex2Scholl B. Burge J. Priebe N.J. integration selectivity cortex.J. Neurophysiol. 2013; 109: 3013-3024Crossref (57) However, compared have much more laterally placed smaller zone 1B). This raises question whether mouse, which become popular vision, relies cues perception. this issue Current Biology, Boone, Samonds et al.3Boone H.C. J.M. Crouse E.C. Barr C. McGee A.W. Natural discrimination explained activity.Curr. Biol. 2021; 31: 2191-2198Abstract Full Text PDF (2) Scholar provide insights into linking encoding cortex. developed approach was inspired classic cliff task4Gibson E.J. Walk R.D. “visual cliff”.Sci. Am. 1960; 202: 64-71Crossref (366) Scholar, glass plate above platforms at depths. Animals instinctively move ‘safest’ platform closest avoid deep ‘visual cliffs’, demonstrating depth. mice, standard task typically engages lower part field little overlap between larger upper 1C), extending front animal’s head. therefore reasoned orienting head towards ground would engage field. authors exploited fact natural climbers introduced vertical pole cliff. Mice were top climbed down, pointing down 1D). They reliably descended shallow (2.5 cm below plate) when other exceeded 10 cm. Are using cues, particular disparity, during descent task? As requires input both repeated same experiments one eye sutured closed. contrast performing only monocular did not descend shallower platforms, difference large 60 open take longer pole, suggesting closure impair performance. Other ‘motion parallax’, speeds close far retina an animal moving, could principle still be used these mice. These indicate use identify ‘safest’, most platform. Next, link response properties (V1). V1 area important retinas, arriving via thalamus. To test information task, two-photon imaging measure neural responses separate group head-fixed First, associated depths estimated behavioral data freely moving Based disparities, stimuli generated presented polarization-preserving screen polarization filters techniques similar those cinema. allowed estimate ‘neural discrimination’ performance, indicating how well discriminated based responses. Strikingly, performance challenge naturalistic tasks it always clear making decision. enabled determine choosing measuring angle (indicating four turning descent; Figure turned reaching bottom about 5 Neural relatively robust against variations viewing distance. Closer distances, however, caused wider range better discrimination. provides possible explanation why made minds up late. confirm idea, forced make discriminations further away, increasing but keeping absolute constant. Consistent data, worse distances. Finally, examined changes alignment behaving rapid, saccadic movements gaze shifts5Meyer A.F. O’Keefe Poort Two distinct types eye-head coupling mice.Curr. 2020; 30: 2116-2130Abstract (30) Scholar,6Michaiel A.M. Abe E.T. Niell C.M. Dynamics control prey capture mice.eLife. 9e57458Crossref (21) Saccade sizes can differ across differences (the latter measurements). analyzed previous study5Meyer tracked simulated effects perceived comparable middle 50% observed saccade sizes. Such robustness very accurate conditions. emerging line research exploits behaviors, defensive7Yilmaz M. Meister Rapid innate defensive looming stimuli.Curr. 23: 2011-2015Abstract (250) hunting behaviors8Hoy J.L. Yavorska I. Wehr Vision drives laboratory 2016; 26: 3046-3052Abstract (89) processing Major advantages behaviors they do require extensive training results easily generalized ‘real world’. There are, challenges monitoring motor outputs modify processing, movements. Additionally, certain experimental difficult perform animals. linked collected precisely controlled input. Discrimination cells matched movement. supports idea contribute prediction tested future work. Recent technological advances expand way we manipulate activity. For example, precise tracking location9Mathis A. Mamidanna P. Cury K.M. T. Murthy V.N. Mathis M.W. Bethge DeepLabCut: markerless pose estimation user-defined body parts learning.Nat. 2018; 21: 1281-1289Crossref (883) including positions10Wallace D.J. Greenberg D.S. Sawinski Rulla S. Notaro G. Kerr J.N.D. Rats maintain overhead expense constant fusion.Nature. 498: 65-69Crossref (172) Scholar,11Meyer Sahani Linden J.F. head-mounted camera system untegrates detailed multichannel electrophysiology mice.Neuron. 100: 46-60.e7Abstract (51) enable reconstruction input10Wallace Using tools, interesting investigate what extent orientation counter-acted previously reported rodents5Meyer Scholar,10Wallace 11Meyer 12Oommen B.S. Stahl J.S. Eye static tilts its relationship spontaneous pitch mouse.Brain Res. 2008; 1193: 57-66Crossref (47) Perhaps greatest difficulty stimulus. Recently head-coupled large-field projection systems13Del Grosso N.A. Graboski J.J. Chen W. Blanco-Hernández E. Sirota Virtual reality freely-moving rodents.bioRxiv. 2017; https://doi.org/10.1101/161232Crossref (0) Scholar,14Stowers J.R. Hofbauer Bastien R. Griessner Higgins Farooqui Fischer R.M. Nowikovsky K. Haubensak Couzin I.D. al.Virtual animals.Nat. Methods. 14: 995-1002Crossref (120) promising dissociate impact motion parallax, pole. tools conjunction large-scale recording techniques15Jun Steinmetz Siegle J.H. Denman Bauza Barbarits Lee A.K. Anastassiou C.A. Andrei Aydın Ç. al.Fully integrated silicon probes high-density activity.Nature. 551: 232-236Crossref (754) Scholar,16Zong Wu Li Hu Y. Rong H. Xu Lu al.Fast high-resolution miniature microscopy mice.Nat. 713-719Crossref (229) activity disparity-sensitive ability genetically defined local cell populations inter-areal projections17Luo L. Callaway E.M. Svoboda Genetic dissection circuits: decade progress.Neuron. 98: 256-281Abstract (185) permits dissecting roles circuits visually-guided contribution higher areas perception18La Chioma Bonhoeffer Hübener Area-specific mapping cortex.Curr. 2019; 29: 2954-2960Abstract (31) Scholar,19Thomas O.M. Cumming B.G. Parker specialization relative V2.Nat. 2002; 5: 472-478Crossref (167) “old method watching wondering behavior”20Tinbergen N. Ethology stress diseases.Science. 1974; 185: 20-27Crossref (75) favored ethologists helping gain fundamental basis New easier timely rich, help understand mechanisms