Table Of Content

Visual receptive field properties of cells in the optic tectum of the archer fish Mor Ben-Tov, Ivgeny Kopilevich, Opher Donchin, Ohad Ben-Shahar, Chen Giladi and Ronen Segev J Neurophysiol 110:748-759, 2013. First published 8 May 2013; doi: 10.1152/jn.00094.2013 You might find this additional info useful... This article cites 48 articles, 17 of which you can access for free at: http://jn.physiology.org/content/110/3/748.full#ref-list-1 Updated information and services including high resolution figures, can be found at: http://jn.physiology.org/content/110/3/748.full Additional material and information about Journal of Neurophysiology can be found at: D http://www.the-aps.org/publications/jn o w n lo a d This information is current as of August 1, 2013. e d fro m h ttp ://jn .p h y s io lo g y .o rg a/ t M a rin e B io lo g ic a l L a b o n A u g u s t 1 , 2 0 1 3 Journal of Neurophysiology publishes original articles on the function of the nervous system. It is published 24 times a year (twice monthly) by the American Physiological Society, 9650 Rockville Pike, Bethesda MD 20814-3991. Copyright © 2013 the American Physiological Society. ESSN: 1522-1598. Visit our website at http://www.the-aps.org/. JNeurophysiol110:748–759,2013. FirstpublishedMay8,2013;doi:10.1152/jn.00094.2013. Visual receptive field properties of cells in the optic tectum of the archer fish Mor Ben-Tov,1,2 Ivgeny Kopilevich,2,3 Opher Donchin,2,3,4 Ohad Ben-Shahar,2,5 Chen Giladi,6 and Ronen Segev1,2 1DepartmentofLifeSciences,Ben-GurionUniversityoftheNegev,Be’er-Sheva,Israel;2ZlotowskiCenterforNeuroscience, Ben-GurionUniversityoftheNegev,Be’er-Sheva,Israel;3DepartmentofBiomedicalEngineering,Ben-GurionUniversityof theNegev,Be’er-Sheva,Israel;4DepartmentofNeuroscience,ErasmusUniversityMedicalCenter,Rotterdam,TheNetherlands; 5DepartmentofComputerScience,Ben-GurionUniversityoftheNegev,Be’er-Sheva,Israel;and6DepartmentofPhysics, Ben-GurionUniversityoftheNegev,Be’er-Sheva,Israel Submitted5February2013;acceptedinfinalform7May2013 D o Ben-TovM,KopilevichI,DonchinO,Ben-ShaharO,GiladiC, Inadditiontotheadvantagesdescribedabove,thearcherfish w SegevR.Visualreceptivefieldpropertiesofcellsintheoptictectum isaninterestinganimaltostudyasafishwithasemiterrestrial nlo ofthearcherfish.JNeurophysiol110:748–759,2013.Firstpublished visual environment. Put differently, it is an animal that, like ad May 8, 2013; doi:10.1152/jn.00094.2013.—The archer fish is well e most mammals, processes a visual landscape above water d khnanogwinngfoornivtsegeexttarteimoneavbiosuvaelwbaethearv.iTohriisnfisshhoiostiangprwomatiesrinjgetsmaotdeplreiny level,whileasafish,itdoessowithabrainwhosestructureis fro m the study of visual system function because it can be trained to verydifferentfromthatofmammals.Thusstudiesofthevisual h respondtoartificialtargetsandthustoprovidevaluablepsychophys- processinginthearcherfishcanleadtobroaderperspectiveon ttp iocvaelrdathtae.Apaltshtoutwghomduecchadbeesh,alviitotlrealidsatkanhoawvneianbdoeuetdtbheeenfucnocllteiocnteadl twhiethfusnuccthioennavlirpornompeerntite.s of visual systems that have to cope ://jn.p organization of the main visual area supporting this visual behavior, The central unit of the fish visual information processing is hy namely,thefishoptictectum.Inthisarticlewefocusonafundamen- s the optic tectum. The information from each eye travels di- io tal aspect of this functional organization and provide a detailed rectlytothetectallobeontheoppositesideofthebrain,where lo analysis of receptive field properties of cells in the archer fish optic g tectum.Usingextracellularmeasurementstorecordactivitiesofsingle information processing work flow is believed to occur (Ki- y.o cells,wefirstmeasuretheirretinotectalmapping.Wethendetermine noshita and Ito 2006; Kinoshita et al. 2002; Vanegas and Ito rg their receptive field properties such as size, selectivity for stimulus 1983).Informationprocessinginthemammalianvisualcortex a/ directionandorientation,tuningforspatialfrequency,andtuningfor appears to progress hierarchically, where the visual stream t M a temporal frequency. Finally, on the basis of all these measurements, ascends from one cortical region to another with progressive rin we demonstrate that optic tectum cells can be classified into three representationandprocessingofvisualinformation.Theoptic e categories: orientation-tuned cells, direction-tuned cells, and direc- tectum of fish, however, appears to be much more uniform Bio tion-agnostic cells. Our results provide an essential basis for future acrossthetectumsurfaceintermsoffunctionalunits(Nevinet lo investigations of information processing in the archer fish visual g al. 2010); i.e., the cells in the optic tectum are diverse in their ic system. a responsetodifferentfeaturesofthevisualstimuli,andtheyare l L archer fish; optic tectum; early visual processing; retinotectal map- not arranged in columnar organization. Furthermore, next to a b ping;orientationanddirectionselectivity visualsignalsfromtheretina,inputfromothersensorymodal- o n ities such as auditory, lateral line, and somatosensory are also A u mapped topographically over the tectum (Bodznick 1990; g INRECENTYEARS,THEARCHERFISHhasprovedtobeaveryproduc- u Catania et al. 2010; Knudsen 1982). Anatomically, it was s tivemodelinthestudyofmanyaspectsofvisualbehavior,from found in other teleosts that the most important distinction is t 1 retinalencoding(Tsvillingetal.2012;Vassermanetal.2010),eye betweenthesuperficiallayers,whicharemainlyvisual,andthe , 2 movements(Ben-Simonetal.2009),animalcompetition(Schus- deep layers, which are multimodal and motor (Bastian 1982; 013 ter2011),andvisualacuity(Ben-Simonetal.2012a;Templeetal. Bodznick 1990). Focusing on vision, we have therefore se- 2010, 2013) to tracking of moving targets (Ben-Simon et al. lected the superficial layers of the archer fish optic tectum as 2012b; Schlegel and Schuster 2008; Schuster et al. 2006; Tim- our target of research. mermans2000,2001)andvisualsearch(Mokeichevetal.2010). Studies have described the retinotectal maps in fish by re- The productivity of the model can be explained, in part, by the cording from the superficial tectum using single electrodes (Ja- ability of the archer fish to overtly report its visual decisions by cobson and Gaze 1964; Schwassmann and Kruger 1965). These employing a squirt of water from its mouth aimed at selected studies showed that there is correspondence between dorsal and targets. The archer fish can be trained to shoot at a variety of rostral points within the visual field of the eye and points in the artificialtargetsfromsmallfoodpelletstoprintedtargetsonpaper contralateraltectum.Inaddition,researchwasdevotedtofindthe andeventargetspresentedoncomputermonitors.Thuscomplex sizeofthereceptivefields.Itwasfoundthatthesmallestsizeof experimentsofrecognitionanddecisionmakingcanbeconducted the receptive field was 7.7° and 1.8° for zebra fish and goldfish, withcarefulexperimentalsettingstoexplorevariousaspectsofvisual respectively (Damjanovic´ et al. 2009b; Sajovic and Levinthal informationprocessinginapsychophysicallyrigorousmanner. 1982). Knowledge about spatial mapping between the primary visual area of the fish and the visual field and the sizes of the Addressforreprintrequestsandothercorrespondence:R.Segev,Dept.of receptivefieldscanprovidetheopportunityforfurtherinvestiga- Life Sciences, Ben-Gurion Univ. of the Negev, Be’er-Sheva 84105, Israel (e-mail:[email protected]). tionofspatialdiscriminationandvisualacuity. 748 0022-3077/13Copyright©2013theAmericanPhysiologicalSociety www.jn.org VISUAL RECEPTIVE FIELDS IN THE ARCHER FISH OPTIC TECTUM 749 Seekingtostudytheoverallfunctionalorganizationoftheoptic and were in accordance with government regulations of the State of tectumofarcherfish,thespecificpurposeofourpresentworkis Israel. Twenty-eight archer fish (Toxotes chatareus; Fig. 1A), 6–13 twofold: 1) to reveal the retinotectal mapping of the archer fish cminlength,10–15gbodyweight,wereusedinthisstudy.Thefish werenormallykeptinawatertank50(cid:1)70(cid:1)30cminsize((cid:2)100 optic tectum, and 2) to describe the functional properties of the liters)filledwithbrackishwater(2–2.5gofredseasaltmixfor1liter receptivefieldofoptictectumcells.Ourresultswillthusprovide ofwater)at26–28°C.Theroomwasilluminatedwithartificiallight afoundationforfuturestudiesofvisualprocessinginthearcher witha16:8-hday-nightcycle. fish and of the mapping between its visual behavior and neural visualprocessing. Surgery Fish were anesthetized with MS-222 (100 mg per liter of tank MATERIAL AND METHODS water;A-5040,Sigma).ThepHlevelwasadjustedtocompensatefor Animals MS-222 acidity. After the fish lost its buoyancy balance and flipped onto its back, it was restrained in a special device and its gills were All experiments with fish were approved by the Ben-Gurion Uni- watered continuously with tank water containing MS-222. The wa- versity of the Negev Institutional Animal Care and Use Committee tering of the gills was essential due to a possible respiratory failure D o w n A B Stimulus Recording lo a d e d fro m h A/D ttp ://jn .p h y s io lo g y Water Fig. 1. The archer fish and the experimental .o with air setup.A:anarcherfishshootsataninsectabove rg pump waterlevel.B:aschematicviewoftheexperi- a/ mental setup, which consists of a monitor to t M presentthestimuli,arecordingsetupof1ex- a C D tracellularelectrode,andalifesupportsystem rin forthefish.Thefish’sgillswerecontinuously e On-Off Diffuse Flash Static Bar wateredthroughatubeinsertedintoitsmouth. Bio Thewaterreservoirisequippedwithairpump lo thathelpsinsufflateairinthewatertanksothat g oxygenisavailabletothefish.Astreambreaker ica deviceisusedtopreventanelectricalcurrent l L from passing through the water molecules. a b C–F:the4typesofvisualstimuliusedinthe o experiment.C:on-offdiffuseflash,arepeating n A loopconsistingof1sofwhitescreenfollowed u by1sofblackscreen. D:staticbar,astationary g u barin8differentorientations(orientationsvar- s iedby22.5°).E:movingbar,abarmovingin8 t 1 differentdirections(directionsvariedby45°). , 2 Thedirectionofmotionwasorthogonaltothe 01 bar’sorientation.F:sinusoidalgratings,drifting 3 in different spatial and temporal frequencies. Spatialfrequencyvariedbetween0.03and0.7 E F cycles/deg(6logarithmicsteps),andtemporal Moving Bar Sinusoidal Grating frequency varied between 1.5 and 9 cycles/s (6linearsteps). JNeurophysiol•doi:10.1152/jn.00094.2013•www.jn.org 750 VISUAL RECEPTIVE FIELDS IN THE ARCHER FISH OPTIC TECTUM causedbyexposuretoMS-222.Anincisionwasmadeovertheoptic ofthereceptivefield.Inasimilarway,thebarwasmovedindifferent tectum,theskinandfattytissuewereremoved,andlidocaine(L-7757, directionstodeterminethebordersofthereceptivefield.Thismethod Sigma)wasappliedtotheboundariesoftheincision.Atthispointwe enables us to mark the receptive field boundary fast enough that injectedthefishwith5–15(cid:1)lofthenondepolarizingmusclerelaxant mapping of many cells from each animal was possible. The error in gallaminetriethiodide(17g/l;G8134,Sigma)toitsspine,towardits determiningtheexactlocationoftheedgesis(cid:2)1.5°. tail,topreventmusclemovementduringtheexperiment.Specifically, weconfirmedthattheeyemovementsofthefishwereeliminated,and VisualStimuliandDataAnalysis onlythendidwecontinuewiththerestoftheprocedure.Adentaldrill (Fine Science Tools micro drill, no. 097883, with a 2.7-mm tip Four types of stimuli were presented on the screen, and they are diameter; stainless steel trephines, no. 18004-27) was then used to demonstratedinFig.1,C–F.Allstimuliweregeneratedusingeither open the skull and meninges over the optic tectum. A silver wire thePsychophysicstoolboxinMatlabordirectlyinC(cid:4)(cid:4). (76.2-(cid:1)mdiameter,tipcoatedwithsilverchloride)wasplacedinthe ON-OFFstimulus.TheON-OFFstimulusconsistedof1sofwhite cerebrospinal fluid near the optic tectum to use as a reference elec- screenfollowedby1sofblackscreen(lightintensityvaluesof2.66 trode. The tectum was then covered with a thin layer of Vaseline to and0.41(cid:1)W/cm2,respectively)in20–30repeatingloops(Fig.1C). preventitsdehydration. Foreachcondition,wecalculatedtheresponseasthemeannumberof spikes in a time window of 50 ms around the peak response time. D InVivoElectrophysiology Afterobtainingtheseresponses,wecomputedtheresponseratio(R) ow asfollows: n lo smTalhleerwfisahterantadnkth(elenrgesthtra2i5nicnmg,wdeivditche6wcmer,ehepilgahcte6dctmog)efithlleerdwinitha R(cid:2)RON(cid:3)ROFF, ade b(nraocMkisSh-2w2a2teart(trheids ssteaagesa;lst,ee2–F2i.g5.g1/Bl)).uTphteofi0s.5h’csmgilalbsowveereeyceolnetvine-l RON(cid:4)ROFF d fro where R is the response to the onset of the flash and R is the m uouslywateredthroughatubeinsertedintoitsmouthtocompensate responseOtNotheoffsetoftheflash. OFF h for possible respiratory degradation. The fish was placed so that its Stationary bar. The stimulus consisted of a bar (white over black ttp roifghthteeyfieshw)aisn0t.h3eccmenftreormofththeegltaasnsk.wTahllis(pgalraaslslewltaollthweassahgiigtthaelrptlhanane background) in eight different orientations (orientations step 22.5°; ://jn seeFig.1D).Thelengthandwidthofthebarwereadjustedtomatch .p theotherwalls(12-cmheight)andthusallowedthefishawidevisual h thereceptivefieldsize,asestimatedbytheinitialstimulationproce- y field (about 110° in both the vertical and horizontal axes). Using a s single electrode (tungsten, glass coated, 250-(cid:1)m diameter, 2-M(cid:3) dscurreee(ns,eteoaabvoovied).aEnaychadbaaprtaptrieosnenetfafteicotns,wians7f–o1ll2owreepdebaytin0g.5losoopfsb.lFacokr iolo impedance, 60-mm length; no. 366-060620-11, Alpha-Omega, Naz- g eachorientation,wecalculatedtheresponseastheaveragefiringrate y areth,Israel)mountedonacalibratedmanipulator(Narishige,Tokyo, .o Japan), recordings were made from the superficial layers (up to 500 in a time window of 200 ms around the peak response time. To rg (cid:1)m deep) of the optic tectum. The signal obtained was magnified determinewhetherthecellisorientationtunedornot,wecalculated a/ ((cid:1)104)andfiltered(band-passboxfilter,300Hz–10kHzrange)byan theorientationindexasfollows: t M kacmhHazpn,lniafienlesdr:2(1D))tAthhMeessi5igg0nn,aaWllwwPeaIn)stastnahmdroptuhlgeehdnaatnrnadannsremacloiotgrtedndeodtthctorhoaufiglcthoermtw(pHouutepmraraBatlul2ge0l, OrientationIndex(cid:2)RRpp(cid:4)(cid:3)RRoo, arine B Q(AuMes9t,SGcriaesnstiIfincs)t,rurmemenotvsi)nagnd50anHozs,ciallnodscothpeen(TtoDSan21a0u,dTioekmtroonniitxo)r. whereRpistheresponserateatthepreferredorientationandRoisthe iolog response rate at the orthogonal orientation. According to previous ic Inthiswaytheneuralresponsecouldbebothheardandseeninreal studies(Baronetal.2007;Wangetal.2010),weusedthevalueof0.5 a ti3in0mg–es4ed0susrmiioninng.wthaese(cid:2)xp5erhim, aenndt.Twheewavereeraagbeledutroathioonldosfiangtylepiucnailtsreucoprdto- tooriednistatitniognu-iashgnboesttwiceecnelltsw.oCceelllsl cwlaitshseso:rioenriteantitoantioinn-dtuenxed(cid:5)c0e.5llswaenrde l Lab o consideredorientation-tunedcells.Wethenfittedparametriccurvesto n responsesasafunctionoftheorientationofthebar,todeterminethe A u SpikeSorting width of the orientation tuning. The fit was based on the probability g u density function of the von Mises distribution (Fisher 1995) as s Spikesortingwasdoneofflineusinganin-houseMatlabprogram. describedinEq.1: t 1 First, the signal was filtered with a band-pass filter with additional , 2 noisereductionat50-Hzfrequency.Potentialspikesweredetectedin p.e(cid:6)cos(cid:1)x-(cid:1)(cid:2) 01 M(x)(cid:2) , (1) 3 events where the filtered signal crossed a threshold of 4.5 times the 2(cid:7)I ((cid:6)) 0 standard deviation of the signal. The peak in every segment was determined,and1.5msofsignalwaskeptbeforeandafterthepeak. where I (x) is the modified Bessel function of order 0 and the fitted 0 Usingagraphicaluserinterface,wemanuallyremovedeventsthatdid parametersarep,peakresponse;(cid:1),centerorientationatthepeak;and nothavetheshapeofaspike.Wethenclusteredthespikesintoone (cid:6),concentrationfactor. ormoregroupsbasedonspikeamplitudeandwidth. Weexploredtheparameter(cid:6)tofindtheorientationtuningwidth. TheadvantageinfittingavonMisesdistributiontothiskindofdata EstimationoftheLocationoftheReceptiveField isthatitenablesusnotonlytofindthepreferredorientation,butalso tofacilitatetheextractionofdatasuchasthetuningwidth. We stimulated the visual system using a hand-held flashlight to Moving bar. A circle 120% of the receptive field size was drawn determine whether each single cell could produce a high-quality around the estimated receptive field location, and a bar, with length neural response. In well-isolated cells with a high-quality neural andwidthadjustedtomatchthereceptivefieldsize,wasmovedfrom response,thestimulatingcomputerscreenwaspositioned20cmfrom onesideofthecircletotheoppositeside(1cycleper1.5s,followed theanimal’seye.Toestimatethelocationofthecell’sreceptivefield, by 0.5 s of black screen, to avoid any adaptation effects, in 7–12 a white bar was moved interactively by the experimenter across the repeating loops) with direction of motion orthogonal to the bar’s screen in different orientations and moving directions to detect the orientation(8differentdirections;seeFig.1E).Foreachdirection,we limitsofthereceptivefield.Thebarwasmovedacrossthescreenuntil calculated the response as the mean number of spikes in a time astrongreactionwasheard;thispointwasmarkedasoneoftheedges window of 200 ms around the peak response time. To determine JNeurophysiol•doi:10.1152/jn.00094.2013•www.jn.org VISUAL RECEPTIVE FIELDS IN THE ARCHER FISH OPTIC TECTUM 751 whetherthecellisdirectiontunedornot,wecalculatedthedirection EstimationofConfidenceIntervalsonCellTypeRatiointhe indexasfollows: Population DirectionIndex(cid:2)Rp(cid:3)Ro, For each cell type ratio, we determined the 95% confidence R (cid:4)R intervalsusingtheClopper-Pearsonmethod.Inaddition,weestimated p o the confidence interval of the event that there is an additional cell whereR istheresponserateatthepreferreddirectionandR isthe groupwedidnotobserveinthedataset. p o response rate at the opposite direction. Cells with direction index (cid:5)0.5 were considered direction-tuned cells. For cells with direction index(cid:5)0.5,wecalculatedtheorientationindex(asdescribedabove). RESULTS Cells with orientation index (cid:5)0.5 were considered orientation-tuned We measured the receptive field location and functional cells. We then fitted parametric curves to responses as a function of the direction of the bar to the cells with direction index (cid:5)0.5, to properties of cells in the optic tectum of the archer fish using extracellular electrodes. The fish was immobilized during ex- determinethewidthofthedirectiontuning.Thefitwasbasedonavon MisesfunctionsasdescribedinEq.1. periments, and respiration was reduced by perfusing the gills. For comparison, and for ensuring our results are robust, we also Overall, we measured from a total of 166 cells in 28 different D fitted the response of the cell to the different directions with a animals in different stimulus conditions. o w second-orderharmonicfunction(HF)assuggestedbyMaximovetal. n (2005): lo Retinotectal Map of Archer Fish Optic Tectum a d HF(x)(cid:2)a (cid:4)a cos(x(cid:3)(cid:1))(cid:4)a cos(2x(cid:3)2(cid:1)), (2) e 0 1 1 2 2 Torevealtheretinotectalmappingofreceptivefieldsofcells d where a are the amplitude and (cid:1) are the phases of the different in the optic tectum, we mapped the location of 35 cells in 5 fro i i m harmonics. The amplitudes of the harmonics (a0, a1, and a2) were different animals by using an interactively moving bar in h usedtoclassifythecellsintothreedifferentgroups:direction-tuned, different orientations. For each location on the optic tectum ttp oclraiesnsitfiateidona-studnireedc,tiaonnd-tduinreecdtiiofnth-aegirnfiosrstitchcaerlmlso.nSipcehcaifidcdailsltyi,nccetlrleslawtievree surface (see MATERIALS AND METHODS), we found several cells ://jn atahme(cid:6)iprlis0teu.c5doaen,,dia.hen.ad,rmaa1sod(cid:6)niirceahc2taiadonndd-iasagti1nnoc(cid:6)sttri0ec.l5aoatni0ve,seaiasfmtohpreliiertneutldaaettii,ovin.ee-a.t,umanp2el(cid:6)dituoadn1eesasnoidff fteiraoocnmhaldnoidcfafedtirioemnnetonansnitoihmnesa.olspTtahincedtaemvceteuraamsguesruelrdofcatachteeiowrneecareenpdtthivdeeinmficeaenllcsdiuollonatceoad-f .physio 2 0 torevealthemapbetweenthespecificplaceintheoptictectum lo both their first and second harmonics were small (Maximov et al. g 2005). and the location in the visual field (Fig. 2, A and B). y.o Sinusoidal gratings. The stimulus consisted of drifting sinusoidal We found that the dorsal visual field projects to the lateral rg gratings (maximum and minimum light intensity values of 2.66 and part of the optic tectum and the ventral visual field projects to a/ 0.41(cid:1)W/cm2,respectively)indifferentspatialandtemporalfrequen- themedialpartoftheoptictectum.Thevisualfieldinthenasal t M cfoiellsow(Feidg.by1F0)..5Esaocfhbplarceksesnctraeteionn,toofavaogidraatinnygaadpappetaatrieodnefoffrec1t.s5,isn, dteimrepctoiroanl dpirroejcetcitosntoprothjeectrsosttoratlhepacratuadnadl pvaisrutaolffitehledtienctuthme arine 7–12 repeating loops. The direction of the grating was chosen to be B thepreferreddirectionobtainedfromthemovingbarstimulus.Spatial (Fig. 2, A and B). io Inaddition,wefoundthatthereceptivefieldscanbedivided lo frequency varied between 0.03 and 0.7 cycles/deg (6 logarithmic g steps), and temporal frequency varied between 1.5 and 9 cycles/s (6 into two groups according to their location, size, and shape. ica linearsteps),resultinginatotalof36gratingconditions.Firingrates Thefirstgroupconsistsofcellswithreceptivefieldsinfrontof l L foreachgratingwereobtainedbyaveragingthenumberofspikesover thefish’seyeinthedorsal-ventralplane,i.e.,atthesameheight ab the1.5-sstimulusdurationacrossalltrialsandsubtractingthefirst50 as the fish eye. The second group consists of cells with o n ms, to avoid the onset response. Collecting these responses for all receptive fields located more dorsally (Fig. 2B). A stimulus conditions, we then determined the preferred spatial and Exploringtheshapeofthereceptivefields,wefoundthatthe ug temporal frequency combination for each cell. Finally, we evaluated majority of receptive fields that were located in front of the us the spatial frequency tuning at the preferred temporal frequency and fish’s eye in the dorsal-ventral plane were elongated in their t 1 the temporal frequency tuning at the preferred spatial frequency by verticaldimension(Fig.2C,red,yellow,andgreencircles).In , 2 0 fittingthemwithaGaussianfunction. cells with receptive fields located more dorsally, the receptive 13 fields were slightly more elongated in their horizontal dimen- RetinotectalMappingoftheOpticTectum sion than in their vertical dimension (Fig. 2C, blue and cyan circles). A closer examination of the width and height of the To determine the gross mapping between the visual field to the optictectumsurface,wefirsttookaphotooftheoptictectumandaligned receptive fields shows that the receptive fields of cells with ittoagridconsistingofsmallregionsalongtheanterior-posteriorand receptive fields located in front of the fish’s eye in the dorsal- the medial-lateral axes. This procedure divided the optic tectum ventralplanearesignificantlynarrowerandlongerthanrecep- surfacetosmallregionsofinterest(Fig.2A).Onceawell-isolatedcell tive fields of cells with receptive fields located more dorsally, with a high-quality neural response was found, we marked the loca- which were wider and shorter. tionofthepenetrationtotheoptictectumonthephotowepreviously Wedeterminedthesizeofthereceptivefieldasthegeomet- took.Theaccuracyofthedeterminingtheexactpenetrationpointwas ricalmeanofitstwoedges(Fig.2D).Wefoundthattherewas (cid:2)0.4 mm. At the end, we grouped single units to those regions by a significant difference in the size distribution of the two their tectal location. For each unit, we estimated the receptive field groups (Kolmogorov-Smirnov test, P (cid:5) 0.001). Specifically, dimension and location as described in above. We then assigned an small receptive fields (range: 3.12°–9.48°, mean: 5.32°) were averagereceptivefieldlocationanddimensionforeachgridregionby averagingthosepropertiesofallcellsthatcorrespondedtothatregion. locatedabovethetemporal-nasalline.Largereceptivefieldsof This procedure was done in different animals and allowed us to cells(range:5.45°–17.69°,mean:10.14°)werelocatedinfront combinedatafromdifferentexperiments. of the fish eye. JNeurophysiol•doi:10.1152/jn.00094.2013•www.jn.org 752 VISUAL RECEPTIVE FIELDS IN THE ARCHER FISH OPTIC TECTUM A B DORSAL L LATERAL A R N O A P S Fig.2.Spaceandsizeorganizationofreceptive R M A fieldsinthearcherfishoptictectum.A:aphotoof AL OS TE L the archer fish brain with the left optic tectum D T delineated(top)andtheelectrodepositionsonthe U R left optic tectum (bottom). B: topographical ar- A A D C L o rangementofreceptivefieldsintherightvisual w n fieldofthearcherfish.Theblackcirclerepresents 10° lo thefisheye.Eachcolorcorrespondstotheelec- a MEDIAL d trodepositionshowninA.Solidrectanglesrep- VENTRAL ed rdeismeenntstihoen,aavnedracgoentroeucreprteicvteanfigeleldsrleopcraetisoenntatnhde fro m areaintowhichallthereceptivefieldsbelonging C D tattooisotnathhfeeoufsnealrcemetcicoeternpogtodriivfedeiptfsofiahlselie.ltdiiCgosh:nitwz.sehEi.doatwchhnocfionaloArre.ccDeop:rrtdievisseptrofiinbedulds- h (deg) s 76 http://jn.ph Widt Cell 5 ysio d of 4 log ve Fiel umber 32 ay.org/ pti N t M Rece 10 arine 0 5 10 15 20 B Receptive Field Height (deg) Receptive Field Size (deg) io lo g ic Inaddition,weexaminedtheareainwhichallthereceptive We found that most cells respond to both ON and OFF a fieldsbelongingtothesamegridfall(Fig.2B,contourrectan- stimulus and that there is a continuum in the response of cells l La b gles). We found that there is no significant difference in the to this kind of stimulus, since the ON/OFF response ratio of o widths of the circumscribed receptive field areas (permutation different cells spans continuously the interval (cid:7)1 to 1 (Fig. n A test, P (cid:6) 0.05). However, there is a significant difference 3D). In addition, there is a slight preference for the OFF u g between the two groups in the variability in the dorsal-ventral stimulus, since the peak of the response histogram is negative us axis (permutation test, P (cid:5) 0.01). The height of the circum- ((cid:7)0.14 (cid:8) 0.05, mean (cid:8) SE). t 1 scribedareaofthereceptivefieldsclosertothefisheyeissmaller , 2 0 than the height of the circumscribed area of the receptive fields 1 Orientation Selectivity in Archer Fish Optic Tectum 3 locatedmoredorsally. The orientation selectivity of cells was tested by flashing a ArcherFishOpticTectumCellsRespondtoanON/OFFStimulus bar on the receptive field center. This was done first by Wetestedtheresponseofthecellstotheonsetandoffsetof detecting the receptive field boundaries (see MATERIALS AND alightbypresentingafull-fieldflashstimulus.Werecordeda METHODS) and then by flashing a bar with a matched size in total of 83 cells from 15 animals using this type of stimulus. eight different orientations. We recorded 38 cells from 8 Severalcellswithdifferentresponsepropertiesaredepicted animalsusingthistypeofstimulus.Wequantifiedtheresponse in Fig. 3, A–C. Generally, cells responded to the onset and of different cells to orientation by calculating the orientation offset of light in various degrees. To quantify the response index and determined the width of the orientation tuning by profiles,foreachcellwedeterminedtheONresponseleveland fitting a von Mises function only to the cells with high theOFFresponseandcalculatedtheON/OFFresponseratioR, orientation index (for additional details, see MATERIALS AND which quantifies the ON/OFF polarity of the cell. Cells that METHODS). exhibitastrongONresponse(Fig.3A)willhaveresponseratio Wefoundthattheresponseofcellscanbeclassifiedintotwo close to 1, whereas cells that exhibit a strong OFF response types. The first is a strong response to specific orientation, (Fig. 3B) will have response ratios close to (cid:7)1. The response where the response to the preferred orientation could be up to ratio of an ON-OFF cell (Fig. 3C) will be close to 0. 20 times higher than the response to the null orientation, and JNeurophysiol•doi:10.1152/jn.00094.2013•www.jn.org VISUAL RECEPTIVE FIELDS IN THE ARCHER FISH OPTIC TECTUM 753 Response to a Diffuse Flash A B ON Cell OFF Cell # # at at e e p p e e R R Fig.3.CellresponsestoON-OFFstimulus. A:exampleofanONcell,acellthatresponds totheonsetofaflash.B:exampleofanOFF cell, a cell that responds to the offset of a D o 0 1000 2000 0 1000 2000 flash.C:exampleofanON-OFFcell,acell w Time (ms) Time (ms) thatrespondstoboththeonsetandoffsetof nlo aflash.D:adistributionoftheresponseratio a d C D of ON and OFF responses shows that the e ON-OFF Cell cbeeltlwsefeanlloOnNacaonndtinOuuFmFocflaresssepso.nsTehpeattmeronsst d fro 12 abundant pattern is a symmetrical ON-OFF m response. h 10 ttp # Cells 8 ://jn.p peat r of 6 hysio Re mbe 4 logy u .o N rg 2 a/ t M 0 1000 2000 0 a Time (ms) -1 -0.R5 e s p o n 0s e R a t i 0o.5 1 rine B therefore the orientation index is (cid:6)0.5 (Fig. 4, A and B). In cellsinthearcherfishoptictectumhaverelativelysharporienta- iolo addition,wefoundcellswithveryweakdependenceonorien- tiontuningprofiles. gic tation,wherethedifferencebetweenthemaximalandminimal a l L orientation spike count is marginal, for which the orientation Direction and Orientation Selectivity of Tectal Neurons in a index is (cid:5)0.5 (Fig. 4, C and D). We define cells with such Response to Moving Bars b o weak dependence on orientation as orientation-agnostic neu- n A rons. Thedirectionselectivityofcellsinresponsetomovingbarswas u g Using this analysis, we found that 58% of the cells were testedbyfirstlocatingthereceptivefieldcenterandthenmovingabar u s orientation tuned (Fig. 4E). We tested the sensitivity of these acrossthereceptivefield(seeMATERIALSANDMETHODS).Werecorded t 1 numbers to the threshold between cell types and found that 36cellsfrom8animalsusingthisstimulus.Toclassifycellsinto , 2 0 10% change in the threshold results with only 1 cell moving different response groups, we first calculated the spike count in 1 3 between groups. Moreover, the preferred orientation had a response to specific direction (see MATERIALS AND METHODS) and strong bias for the vertical orientation with 59% of the orien- thencalculatedthedirectionindexandtheorientationindex.We tation-tuned cells having a preference for the vertical orienta- then fitted a von Mises function to the response to measure the tion or its neighboring orientations ((cid:8)22.5°), 13.6% of the directiontuningwidth.Forcomparison,wealsofittedaharmonic orientation-tuned cells having a preference for the horizontal function to the response to determine whether our results are orientationoritsneighboringorientations((cid:8)22.5°),and27.2% robust(seeMATERIALSANDMETHODSformoredetails). oftheorientation-tunedcellshavingapreferencefortheother diagonal orientations or its neighboring orientations ((cid:8)22.5°) Wefoundthreetypesofcellsaccordingtotheirresponsetoa moving bar: 1) an orientation-tuned response, where the cell (Fig. 4F). respondswithhighfiringratetomovementsalongaspecificaxis WefittedvonMisesfunctiontotheorientation-tunedcellsand explored the parameter (cid:6)to find their orientation tuning width. in space (Fig. 5A); 2) a direction-tuned response, where the cell Wefoundthat45%ofthecellshadasharptuningwidthbetween respondswithhighratetomovementinaspecificdirection(Fig. 15°and30°,41%ofthecellshadamediumtuningwidthbetween 5B); and 3) an agnostic response, where the cell responds with 30°and45°,and14%ofthecellshadatuningwidthof45°to60°. marginaldifferencesbetweendirections(Fig.5C). Onaverage,meanorientationtuningwidthwas33.26°(SE2.36°; Onthebasisofthequantitativeanalysisdescribedabove,we Fig.4G);i.e.,themostprevalentwidthwas(cid:5)34°,indicatingthat found that 11% of the cells were direction tuned, 39% were JNeurophysiol•doi:10.1152/jn.00094.2013•www.jn.org 754 VISUAL RECEPTIVE FIELDS IN THE ARCHER FISH OPTIC TECTUM Response to a Flashing Bar A B Orientation-Tuned Cell 60 nt u n o o C ati e 40 ent pik Ori e S ar ag 20 B er v A 0 Fig.4.Orientationselectivityprofilesassessed 0 1000 2000 usingthestaticbarstimulus.A:rasterplotsof Time (ms) Bar Orientation D the responses to the 8 oriented bars for an ow orientation-tunedcell.Blacklinesrepresentthe C D n startandendtimeofthebardisplay.B:orien- Orientation Agnostic Cell lo a tation tuning profile for the orientation-tuned 20 de cell from A. The dashed black line repre- nt d sentsthefittedvonMissesdistributionfunction. n ou fro C:rasterplotsoftheresponsestothe8oriented o C 15 m barsforanorientation-agnosticcell.D:orien- ati e h tc9ae5tli%lofnrcotoumnnfiiCndg.enEpcr:eodfiiinlsettrefirbovurattlih)oenfooorrifoesnreitelaentcitoatintvi-oiatnyg-nt(uownsitetihdc Orient e Spik 10 ttp://jn andorientation-agnosticcells.F:distributionof ar ag 5 .ph preferredorientation(with95%confidencein- B er ys terval).Themajorityofcellspreferredthever- Av io ticalorientation.G:distributionoftuningwidth 0 log fortheorientation-tunedcells. 0 1000 2000 y Time (ms) Bar Orientation .org a/ E F G t M a ells 70 n=38 ells 70 n=22 ells 40 n=22 rine C 60 C 60 C B entage of 54320000 entage of 54320000 entage of 321000 iological L rc 10 rc 10 rc ab Pe 0 Agnostic OSreielenctatitvioen Pe P0referred Orientation Pe 0 10-2T0 u 2n0i-3n0g 3W0-4id0 t h40 -(5d0 e g50)-60 on A u g u s orientation tuned, and 50% were classified as direction agnos- Using the second method, fitting of a harmonic function, to t 1 tic(Fig.5D).Wetestedthesensitivityofthesenumberstothe determine the selectivity of the cell in response to a moving , 2 thresholdbetweencelltypesandfoundthat10%changeinthe bar, we found similar results with a few exceptions. All the 01 3 threshold results with 5 cells moving between groups. cells that were classified as direction-tuned cells with the first Of the orientation-tuned cells, 86% preferred the vertical ori- method were also classified as direction-tuned cells with the entation (as shown in Fig. 5A) and 14% preferred the diagonal second method. The second method classified an additional orientation. We calculated the confidence interval for the possi- cellasadirection-tunedcell.Asfortheorientationtuning,the second method missed two cells that were classified as orien- bilityoftheexistenceofanadditionalcelltypethatwehavenot tation-tunedwiththefirstmethodandaddedanadditionalcell observedinourdataset(seeMATERIALSANDMETHODS),orientation- asorientation-tuned.Overall,itseemsthatuseofthedirection selectivecellswithapreferenceforthehorizontalorientation.We and orientation indexes and the von Mises distribution gives found that with 95% of certainty, the proportion of this possible robust results, and moreover, it adds the ability to determine missinggroupisbetween0and0.1. the tuning width, compared with the second method. Wefoundthathalfofthedirection-tunedcellshadthegreatest response to a moving bar in the temporal-nasal direction, as Spatial and Temporal Frequency Tuning determined by exploring the fitted parameter (cid:1) in Eq. 1 and shown in Fig. 5B. The other direction-tuned cells preferred the The spatial and temporal tuning of cells was evaluated by diagonaldirection,onefromnasal-ventraltocaudal-dorsaldirec- analyzingtheresponseofcellstosinusoidalgratingdriftingalong tion and the other from nasal-dorsal to caudal-ventral direction. thecell’spreferreddirection.Oncethecellpreferreddirectionwas Thewidthsofthedirectiontuningvariedbetween60°and90°. detected, we stimulated it with all combinations of six spatial JNeurophysiol•doi:10.1152/jn.00094.2013•www.jn.org VISUAL RECEPTIVE FIELDS IN THE ARCHER FISH OPTIC TECTUM 755 Response to a Moving Bar A B Orientation-Tuned Cell Direction-Tuned Cell Dorsal Nasal Temporal D o w Fig.5.Directionselectivityprofilesassessed n usingthemovingbarstimulus.A:orientation- lo 1 s tunedcellprofile.B:direction-tunedcellpro- ad Ventral file.C:direction-agnosticcellprofile.D:dis- ed tirnitbeurvtiaoln) ofofrsealgencotisvtiicty, o(wriietnhta9t5io%n-ctuonnefidd,eanncde from C D direction-tunedcells. h ttp Direction Agnostic Cell ://jn n=36 .p h y s io lo g y .o rg a/ t M a rin e B Agnostic OSreielenctatitvioen SDeirleeccttiiovne iolo g ic a l L frequencies(rangingfrom0.03to0.7cycles/deg)andsixtemporal that we have not observed in our data (see MATERIALS AND ab frequencies (ranging from 1.5 to 9 Hz). For this stimulus, we METHODS).Wefoundthatwith95%ofcertainty,theproportion on recordedatotalof30cellsfrom6animals. of a possible missing group is between 0 and 0.12. A u For each cell, we found the combination of spatial and In a similar way, to analyze the spatial tuning of a cell, we g u temporal frequencies that elicited the maximal response. The tookitsbesttemporalfrequencyandfittedaGaussianfunction s cells had a higher response to a unique combination of spatial totheresponseatallcorrespondingspatialfrequencies(i.e.,to t 1 and temporal frequencies (Fig. 6, A and B). 1columnoftheresponsematrix).Wefoundtwogeneraltypes , 2 0 We analyzed the cell response to obtain further frequency oftuningprofiles:lowpassandbandpass.Figure6Fshowsa 13 tuningproperties.Toanalyzethetemporalfrequencytuningof low-pass tuning profile. The maximal response of this cell to a cell, we took the best spatial frequency and analyzed the the sinusoidal grating is at the lowest spatial frequency tested cellular response for all the different temporal frequencies at (0.03 cycles/deg), and at higher spatial frequencies the re- that spatial frequency (i.e., 1 row of the response matrix). sponse gradually decreases. Figure 6G shows a band-pass Fitting a Gaussian function to these responses, we obtained tuning profile. The response of this cell to the sinusoidal three types of temporal frequency tuning profiles: a low-pass gratingatthelowestspatialfrequencytested(0.03cycles/deg) tuningprofile,inwhichcellselicitthemaximalresponsetothe andatthehighestspatialfrequencytest(0.7cycles/deg)islow, lowertemporalfrequenciestested(Fig.6C);ahigh-passtuning and the maximal response is for a spatial frequency in the profile, in which cells elicit the maximal response to the middle of the range (0.09–0.18 cycles/deg). We found that highest temporal frequencies tested (Fig. 6D); and an all-pass 57%ofthecellshadalow-passtuningprofileand43%ofthe tuning profile, in which there was almost no difference in the cells had a band-pass tuning profile (Fig. 6H). We calculated response to the different temporal frequencies (Fig. 6E). We theconfidenceintervalforthepossibilityoftheexistenceofan foundthat23%ofthecellshadalow-passtuningprofile,54% additional tuning profile group that we have not observed in hadahigh-passtuningprofile,and23%hadanall-passtuning ourdata(seeMATERIALSANDMETHODS).Wefoundthatwith95% profile(Fig.6H).Wecalculatedtheconfidenceintervalforthe of certainty, the proportion of a possible missing group is possibilityoftheexistenceofanadditionaltuningprofilegroup between 0 and 0.12. JNeurophysiol•doi:10.1152/jn.00094.2013•www.jn.org 756 VISUAL RECEPTIVE FIELDS IN THE ARCHER FISH OPTIC TECTUM Response to a Sinusoidal Grating A B 1 1 g) 0.7 g) 0.7 e 0.9 e 0.9 d N d N e/ 0.8 o e/ 0.8 o cycl 0.35 0.7 rma cycl 0.35 0.7 rma ncy ( 0.18 0.6 lized ncy ( 0.18 0.6 lized e 0.5 S e 0.5 S equ 0.09 0.4 pik equ 0.09 0.4 pik atial Fr 0.05 00..23 e Coun atial Fr 0.05 00..23 e Coun Down p t p t lo S 0.03 0.1 S 0.03 0.1 a d 0 0 ed Te1.m5por3al Fre4.q5uen6cy (c7y.c5le/s9) Te1.m5por3al Fre4.q5uen6cy (c7y.c5le/s)9 from C D E http ://jn .p nt nt nt hy u u u s o o o io e C e C e C log k k k y pi pi pi .o S S S rg alized alized alized at Ma/ orm orm orm rine N N N B io lo Temporal Frequency (cycle/s) Temporal Frequency (cycle/s) Temporal Frequency (cycle/s) g ic a F G H l La 80 n=30 b o n ount ount Cells 7600 Augu d Spike C d Spike C entage of 543000 st 1, 2013 e e c z z r ali ali Pe 20 m m 10 r r o o N N 0 Low High All Low Band Pass Pass Pass Pass Pass Spatial Frequency (cycle/deg) Spatial Frequency (cycle/deg) Temporal Tuning Spatial Tuning Fig.6.Responsestospatialandtemporalfrequenciesandtuningprofilesassessedusingsinusoidalgratings.A:exampleofacellforwhichthemaximalresponse rate is at the lowest spatial and temporal frequencies tested. B: example of a cell for which the maximal response rate is at one of the central spa- tialandthehighesttemporalfrequenciestested.C:low-passtemporalfrequencytuningprofileshowingthehighestresponserateatthelowesttemporalfre- quencytested.D:high-passtemporalfrequencytuningprofileshowingthehighestresponserateatthehighesttemporalfrequencytested.E:all-passtemporal frequencytuningprofile.F:low-passspatialfrequencytuningprofileshowingthehighestresponserateatthelowestspatialfrequencytested.G:band-passspatial frequencytuningprofile.H:distributionoftuningprofiles(with95%confidenceinterval).Lightgray,temporalfrequencytuning;darkgray,spatialfrequency tuning.Wecalculatedtheconfidenceintervalforthepossibilityoftheexistenceofanadditionaltuningprofilegroupthatwehavenotobservedinourdata(see MATERIALSANDMETHODS).Forboththespatialfrequencyandtemporalfrequencytuning,wefoundthatwith95%ofcertainty,theproportionofapossiblemissing groupisbetween0and0.12. JNeurophysiol•doi:10.1152/jn.00094.2013•www.jn.org

Description:
Downloaded from system. archer fish; optic tectum; early visual processing; retinotectal map- ping; orientation .. parameters are p, peak response; , center orientation at the peak; and .. to the electrode position shown in A. D: distribu- .. archer fish live in water that is a three-dimensional w
ebook img

Visual receptive field properties of cells in the optic tectum of the archer fish PDF

13 Pages·2013·1.21 MB·English
by  
Save to my drive
Quick download
Download
Upgrade Premium
Most books are stored in the elastic cloud where traffic is expensive. For this reason, we have a limit on daily download.

Preview Visual receptive field properties of cells in the optic tectum of the archer fish

Visual receptive field properties of cells in the optic tectum of the archer fish Mor Ben-Tov, Ivgeny Kopilevich, Opher Donchin, Ohad Ben-Shahar, Chen Giladi and Ronen Segev J Neurophysiol 110:748-759, 2013. First published 8 May 2013; doi: 10.1152/jn.00094.2013 You might find this additional info useful... This article cites 48 articles, 17 of which you can access for free at: http://jn.physiology.org/content/110/3/748.full#ref-list-1 Updated information and services including high resolution figures, can be found at: http://jn.physiology.org/content/110/3/748.full Additional material and information about Journal of Neurophysiology can be found at: D http://www.the-aps.org/publications/jn o w n lo a d This information is current as of August 1, 2013. e d fro m h ttp ://jn .p h y s io lo g y .o rg a/ t M a rin e B io lo g ic a l L a b o n A u g u s t 1 , 2 0 1 3 Journal of Neurophysiology publishes original articles on the function of the nervous system. It is published 24 times a year (twice monthly) by the American Physiological Society, 9650 Rockville Pike, Bethesda MD 20814-3991. Copyright © 2013 the American Physiological Society. ESSN: 1522-1598. Visit our website at http://www.the-aps.org/. JNeurophysiol110:748–759,2013. FirstpublishedMay8,2013;doi:10.1152/jn.00094.2013. Visual receptive field properties of cells in the optic tectum of the archer fish Mor Ben-Tov,1,2 Ivgeny Kopilevich,2,3 Opher Donchin,2,3,4 Ohad Ben-Shahar,2,5 Chen Giladi,6 and Ronen Segev1,2 1DepartmentofLifeSciences,Ben-GurionUniversityoftheNegev,Be’er-Sheva,Israel;2ZlotowskiCenterforNeuroscience, Ben-GurionUniversityoftheNegev,Be’er-Sheva,Israel;3DepartmentofBiomedicalEngineering,Ben-GurionUniversityof theNegev,Be’er-Sheva,Israel;4DepartmentofNeuroscience,ErasmusUniversityMedicalCenter,Rotterdam,TheNetherlands; 5DepartmentofComputerScience,Ben-GurionUniversityoftheNegev,Be’er-Sheva,Israel;and6DepartmentofPhysics, Ben-GurionUniversityoftheNegev,Be’er-Sheva,Israel Submitted5February2013;acceptedinfinalform7May2013 D o Ben-TovM,KopilevichI,DonchinO,Ben-ShaharO,GiladiC, Inadditiontotheadvantagesdescribedabove,thearcherfish w SegevR.Visualreceptivefieldpropertiesofcellsintheoptictectum isaninterestinganimaltostudyasafishwithasemiterrestrial nlo ofthearcherfish.JNeurophysiol110:748–759,2013.Firstpublished visual environment. Put differently, it is an animal that, like ad May 8, 2013; doi:10.1152/jn.00094.2013.—The archer fish is well e most mammals, processes a visual landscape above water d khnanogwinngfoornivtsegeexttarteimoneavbiosuvaelwbaethearv.iTohriisnfisshhoiostiangprwomatiesrinjgetsmaotdeplreiny level,whileasafish,itdoessowithabrainwhosestructureis fro m the study of visual system function because it can be trained to verydifferentfromthatofmammals.Thusstudiesofthevisual h respondtoartificialtargetsandthustoprovidevaluablepsychophys- processinginthearcherfishcanleadtobroaderperspectiveon ttp iocvaelrdathtae.Apaltshtoutwghomduecchadbeesh,alviitotlrealidsatkanhoawvneianbdoeuetdtbheeenfucnocllteiocnteadl twhiethfusnuccthioennavlirpornompeerntite.s of visual systems that have to cope ://jn.p organization of the main visual area supporting this visual behavior, The central unit of the fish visual information processing is hy namely,thefishoptictectum.Inthisarticlewefocusonafundamen- s the optic tectum. The information from each eye travels di- io tal aspect of this functional organization and provide a detailed rectlytothetectallobeontheoppositesideofthebrain,where lo analysis of receptive field properties of cells in the archer fish optic g tectum.Usingextracellularmeasurementstorecordactivitiesofsingle information processing work flow is believed to occur (Ki- y.o cells,wefirstmeasuretheirretinotectalmapping.Wethendetermine noshita and Ito 2006; Kinoshita et al. 2002; Vanegas and Ito rg their receptive field properties such as size, selectivity for stimulus 1983).Informationprocessinginthemammalianvisualcortex a/ directionandorientation,tuningforspatialfrequency,andtuningfor appears to progress hierarchically, where the visual stream t M a temporal frequency. Finally, on the basis of all these measurements, ascends from one cortical region to another with progressive rin we demonstrate that optic tectum cells can be classified into three representationandprocessingofvisualinformation.Theoptic e categories: orientation-tuned cells, direction-tuned cells, and direc- tectum of fish, however, appears to be much more uniform Bio tion-agnostic cells. Our results provide an essential basis for future acrossthetectumsurfaceintermsoffunctionalunits(Nevinet lo investigations of information processing in the archer fish visual g al. 2010); i.e., the cells in the optic tectum are diverse in their ic system. a responsetodifferentfeaturesofthevisualstimuli,andtheyare l L archer fish; optic tectum; early visual processing; retinotectal map- not arranged in columnar organization. Furthermore, next to a b ping;orientationanddirectionselectivity visualsignalsfromtheretina,inputfromothersensorymodal- o n ities such as auditory, lateral line, and somatosensory are also A u mapped topographically over the tectum (Bodznick 1990; g INRECENTYEARS,THEARCHERFISHhasprovedtobeaveryproduc- u Catania et al. 2010; Knudsen 1982). Anatomically, it was s tivemodelinthestudyofmanyaspectsofvisualbehavior,from found in other teleosts that the most important distinction is t 1 retinalencoding(Tsvillingetal.2012;Vassermanetal.2010),eye betweenthesuperficiallayers,whicharemainlyvisual,andthe , 2 movements(Ben-Simonetal.2009),animalcompetition(Schus- deep layers, which are multimodal and motor (Bastian 1982; 013 ter2011),andvisualacuity(Ben-Simonetal.2012a;Templeetal. Bodznick 1990). Focusing on vision, we have therefore se- 2010, 2013) to tracking of moving targets (Ben-Simon et al. lected the superficial layers of the archer fish optic tectum as 2012b; Schlegel and Schuster 2008; Schuster et al. 2006; Tim- our target of research. mermans2000,2001)andvisualsearch(Mokeichevetal.2010). Studies have described the retinotectal maps in fish by re- The productivity of the model can be explained, in part, by the cording from the superficial tectum using single electrodes (Ja- ability of the archer fish to overtly report its visual decisions by cobson and Gaze 1964; Schwassmann and Kruger 1965). These employing a squirt of water from its mouth aimed at selected studies showed that there is correspondence between dorsal and targets. The archer fish can be trained to shoot at a variety of rostral points within the visual field of the eye and points in the artificialtargetsfromsmallfoodpelletstoprintedtargetsonpaper contralateraltectum.Inaddition,researchwasdevotedtofindthe andeventargetspresentedoncomputermonitors.Thuscomplex sizeofthereceptivefields.Itwasfoundthatthesmallestsizeof experimentsofrecognitionanddecisionmakingcanbeconducted the receptive field was 7.7° and 1.8° for zebra fish and goldfish, withcarefulexperimentalsettingstoexplorevariousaspectsofvisual respectively (Damjanovic´ et al. 2009b; Sajovic and Levinthal informationprocessinginapsychophysicallyrigorousmanner. 1982). Knowledge about spatial mapping between the primary visual area of the fish and the visual field and the sizes of the Addressforreprintrequestsandothercorrespondence:R.Segev,Dept.of receptivefieldscanprovidetheopportunityforfurtherinvestiga- Life Sciences, Ben-Gurion Univ. of the Negev, Be’er-Sheva 84105, Israel (e-mail:[email protected]). tionofspatialdiscriminationandvisualacuity. 748 0022-3077/13Copyright©2013theAmericanPhysiologicalSociety www.jn.org VISUAL RECEPTIVE FIELDS IN THE ARCHER FISH OPTIC TECTUM 749 Seekingtostudytheoverallfunctionalorganizationoftheoptic and were in accordance with government regulations of the State of tectumofarcherfish,thespecificpurposeofourpresentworkis Israel. Twenty-eight archer fish (Toxotes chatareus; Fig. 1A), 6–13 twofold: 1) to reveal the retinotectal mapping of the archer fish cminlength,10–15gbodyweight,wereusedinthisstudy.Thefish werenormallykeptinawatertank50(cid:1)70(cid:1)30cminsize((cid:2)100 optic tectum, and 2) to describe the functional properties of the liters)filledwithbrackishwater(2–2.5gofredseasaltmixfor1liter receptivefieldofoptictectumcells.Ourresultswillthusprovide ofwater)at26–28°C.Theroomwasilluminatedwithartificiallight afoundationforfuturestudiesofvisualprocessinginthearcher witha16:8-hday-nightcycle. fish and of the mapping between its visual behavior and neural visualprocessing. Surgery Fish were anesthetized with MS-222 (100 mg per liter of tank MATERIAL AND METHODS water;A-5040,Sigma).ThepHlevelwasadjustedtocompensatefor Animals MS-222 acidity. After the fish lost its buoyancy balance and flipped onto its back, it was restrained in a special device and its gills were All experiments with fish were approved by the Ben-Gurion Uni- watered continuously with tank water containing MS-222. The wa- versity of the Negev Institutional Animal Care and Use Committee tering of the gills was essential due to a possible respiratory failure D o w n A B Stimulus Recording lo a d e d fro m h A/D ttp ://jn .p h y s io lo g y Water Fig. 1. The archer fish and the experimental .o with air setup.A:anarcherfishshootsataninsectabove rg pump waterlevel.B:aschematicviewoftheexperi- a/ mental setup, which consists of a monitor to t M presentthestimuli,arecordingsetupof1ex- a C D tracellularelectrode,andalifesupportsystem rin forthefish.Thefish’sgillswerecontinuously e On-Off Diffuse Flash Static Bar wateredthroughatubeinsertedintoitsmouth. Bio Thewaterreservoirisequippedwithairpump lo thathelpsinsufflateairinthewatertanksothat g oxygenisavailabletothefish.Astreambreaker ica deviceisusedtopreventanelectricalcurrent l L from passing through the water molecules. a b C–F:the4typesofvisualstimuliusedinthe o experiment.C:on-offdiffuseflash,arepeating n A loopconsistingof1sofwhitescreenfollowed u by1sofblackscreen. D:staticbar,astationary g u barin8differentorientations(orientationsvar- s iedby22.5°).E:movingbar,abarmovingin8 t 1 differentdirections(directionsvariedby45°). , 2 Thedirectionofmotionwasorthogonaltothe 01 bar’sorientation.F:sinusoidalgratings,drifting 3 in different spatial and temporal frequencies. Spatialfrequencyvariedbetween0.03and0.7 E F cycles/deg(6logarithmicsteps),andtemporal Moving Bar Sinusoidal Grating frequency varied between 1.5 and 9 cycles/s (6linearsteps). JNeurophysiol•doi:10.1152/jn.00094.2013•www.jn.org 750 VISUAL RECEPTIVE FIELDS IN THE ARCHER FISH OPTIC TECTUM causedbyexposuretoMS-222.Anincisionwasmadeovertheoptic ofthereceptivefield.Inasimilarway,thebarwasmovedindifferent tectum,theskinandfattytissuewereremoved,andlidocaine(L-7757, directionstodeterminethebordersofthereceptivefield.Thismethod Sigma)wasappliedtotheboundariesoftheincision.Atthispointwe enables us to mark the receptive field boundary fast enough that injectedthefishwith5–15(cid:1)lofthenondepolarizingmusclerelaxant mapping of many cells from each animal was possible. The error in gallaminetriethiodide(17g/l;G8134,Sigma)toitsspine,towardits determiningtheexactlocationoftheedgesis(cid:2)1.5°. tail,topreventmusclemovementduringtheexperiment.Specifically, weconfirmedthattheeyemovementsofthefishwereeliminated,and VisualStimuliandDataAnalysis onlythendidwecontinuewiththerestoftheprocedure.Adentaldrill (Fine Science Tools micro drill, no. 097883, with a 2.7-mm tip Four types of stimuli were presented on the screen, and they are diameter; stainless steel trephines, no. 18004-27) was then used to demonstratedinFig.1,C–F.Allstimuliweregeneratedusingeither open the skull and meninges over the optic tectum. A silver wire thePsychophysicstoolboxinMatlabordirectlyinC(cid:4)(cid:4). (76.2-(cid:1)mdiameter,tipcoatedwithsilverchloride)wasplacedinthe ON-OFFstimulus.TheON-OFFstimulusconsistedof1sofwhite cerebrospinal fluid near the optic tectum to use as a reference elec- screenfollowedby1sofblackscreen(lightintensityvaluesof2.66 trode. The tectum was then covered with a thin layer of Vaseline to and0.41(cid:1)W/cm2,respectively)in20–30repeatingloops(Fig.1C). preventitsdehydration. Foreachcondition,wecalculatedtheresponseasthemeannumberof spikes in a time window of 50 ms around the peak response time. D InVivoElectrophysiology Afterobtainingtheseresponses,wecomputedtheresponseratio(R) ow asfollows: n lo smTalhleerwfisahterantadnkth(elenrgesthtra2i5nicnmg,wdeivditche6wcmer,ehepilgahcte6dctmog)efithlleerdwinitha R(cid:2)RON(cid:3)ROFF, ade b(nraocMkisSh-2w2a2teart(trheids ssteaagesa;lst,ee2–F2i.g5.g1/Bl)).uTphteofi0s.5h’csmgilalbsowveereeyceolnetvine-l RON(cid:4)ROFF d fro where R is the response to the onset of the flash and R is the m uouslywateredthroughatubeinsertedintoitsmouthtocompensate responseOtNotheoffsetoftheflash. OFF h for possible respiratory degradation. The fish was placed so that its Stationary bar. The stimulus consisted of a bar (white over black ttp roifghthteeyfieshw)aisn0t.h3eccmenftreormofththeegltaasnsk.wTahllis(pgalraaslslewltaollthweassahgiigtthaelrptlhanane background) in eight different orientations (orientations step 22.5°; ://jn seeFig.1D).Thelengthandwidthofthebarwereadjustedtomatch .p theotherwalls(12-cmheight)andthusallowedthefishawidevisual h thereceptivefieldsize,asestimatedbytheinitialstimulationproce- y field (about 110° in both the vertical and horizontal axes). Using a s single electrode (tungsten, glass coated, 250-(cid:1)m diameter, 2-M(cid:3) dscurreee(ns,eteoaabvoovied).aEnaychadbaaprtaptrieosnenetfafteicotns,wians7f–o1ll2owreepdebaytin0g.5losoopfsb.lFacokr iolo impedance, 60-mm length; no. 366-060620-11, Alpha-Omega, Naz- g eachorientation,wecalculatedtheresponseastheaveragefiringrate y areth,Israel)mountedonacalibratedmanipulator(Narishige,Tokyo, .o Japan), recordings were made from the superficial layers (up to 500 in a time window of 200 ms around the peak response time. To rg (cid:1)m deep) of the optic tectum. The signal obtained was magnified determinewhetherthecellisorientationtunedornot,wecalculated a/ ((cid:1)104)andfiltered(band-passboxfilter,300Hz–10kHzrange)byan theorientationindexasfollows: t M kacmhHazpn,lniafienlesdr:2(1D))tAthhMeessi5igg0nn,aaWllwwPeaIn)stastnahmdroptuhlgeehdnaatnrnadannsremacloiotgrtedndeodtthctorhoaufiglcthoermtw(pHouutepmraraBatlul2ge0l, OrientationIndex(cid:2)RRpp(cid:4)(cid:3)RRoo, arine B Q(AuMes9t,SGcriaesnstiIfincs)t,rurmemenotvsi)nagnd50anHozs,ciallnodscothpeen(TtoDSan21a0u,dTioekmtroonniitxo)r. whereRpistheresponserateatthepreferredorientationandRoisthe iolog response rate at the orthogonal orientation. According to previous ic Inthiswaytheneuralresponsecouldbebothheardandseeninreal studies(Baronetal.2007;Wangetal.2010),weusedthevalueof0.5 a ti3in0mg–es4ed0susrmiioninng.wthaese(cid:2)xp5erhim, aenndt.Twheewavereeraagbeledutroathioonldosfiangtylepiucnailtsreucoprdto- tooriednistatitniognu-iashgnboesttwiceecnelltsw.oCceelllsl cwlaitshseso:rioenriteantitoantioinn-dtuenxed(cid:5)c0e.5llswaenrde l Lab o consideredorientation-tunedcells.Wethenfittedparametriccurvesto n responsesasafunctionoftheorientationofthebar,todeterminethe A u SpikeSorting width of the orientation tuning. The fit was based on the probability g u density function of the von Mises distribution (Fisher 1995) as s Spikesortingwasdoneofflineusinganin-houseMatlabprogram. describedinEq.1: t 1 First, the signal was filtered with a band-pass filter with additional , 2 noisereductionat50-Hzfrequency.Potentialspikesweredetectedin p.e(cid:6)cos(cid:1)x-(cid:1)(cid:2) 01 M(x)(cid:2) , (1) 3 events where the filtered signal crossed a threshold of 4.5 times the 2(cid:7)I ((cid:6)) 0 standard deviation of the signal. The peak in every segment was determined,and1.5msofsignalwaskeptbeforeandafterthepeak. where I (x) is the modified Bessel function of order 0 and the fitted 0 Usingagraphicaluserinterface,wemanuallyremovedeventsthatdid parametersarep,peakresponse;(cid:1),centerorientationatthepeak;and nothavetheshapeofaspike.Wethenclusteredthespikesintoone (cid:6),concentrationfactor. ormoregroupsbasedonspikeamplitudeandwidth. Weexploredtheparameter(cid:6)tofindtheorientationtuningwidth. TheadvantageinfittingavonMisesdistributiontothiskindofdata EstimationoftheLocationoftheReceptiveField isthatitenablesusnotonlytofindthepreferredorientation,butalso tofacilitatetheextractionofdatasuchasthetuningwidth. We stimulated the visual system using a hand-held flashlight to Moving bar. A circle 120% of the receptive field size was drawn determine whether each single cell could produce a high-quality around the estimated receptive field location, and a bar, with length neural response. In well-isolated cells with a high-quality neural andwidthadjustedtomatchthereceptivefieldsize,wasmovedfrom response,thestimulatingcomputerscreenwaspositioned20cmfrom onesideofthecircletotheoppositeside(1cycleper1.5s,followed theanimal’seye.Toestimatethelocationofthecell’sreceptivefield, by 0.5 s of black screen, to avoid any adaptation effects, in 7–12 a white bar was moved interactively by the experimenter across the repeating loops) with direction of motion orthogonal to the bar’s screen in different orientations and moving directions to detect the orientation(8differentdirections;seeFig.1E).Foreachdirection,we limitsofthereceptivefield.Thebarwasmovedacrossthescreenuntil calculated the response as the mean number of spikes in a time astrongreactionwasheard;thispointwasmarkedasoneoftheedges window of 200 ms around the peak response time. To determine JNeurophysiol•doi:10.1152/jn.00094.2013•www.jn.org VISUAL RECEPTIVE FIELDS IN THE ARCHER FISH OPTIC TECTUM 751 whetherthecellisdirectiontunedornot,wecalculatedthedirection EstimationofConfidenceIntervalsonCellTypeRatiointhe indexasfollows: Population DirectionIndex(cid:2)Rp(cid:3)Ro, For each cell type ratio, we determined the 95% confidence R (cid:4)R intervalsusingtheClopper-Pearsonmethod.Inaddition,weestimated p o the confidence interval of the event that there is an additional cell whereR istheresponserateatthepreferreddirectionandR isthe groupwedidnotobserveinthedataset. p o response rate at the opposite direction. Cells with direction index (cid:5)0.5 were considered direction-tuned cells. For cells with direction index(cid:5)0.5,wecalculatedtheorientationindex(asdescribedabove). RESULTS Cells with orientation index (cid:5)0.5 were considered orientation-tuned We measured the receptive field location and functional cells. We then fitted parametric curves to responses as a function of the direction of the bar to the cells with direction index (cid:5)0.5, to properties of cells in the optic tectum of the archer fish using extracellular electrodes. The fish was immobilized during ex- determinethewidthofthedirectiontuning.Thefitwasbasedonavon MisesfunctionsasdescribedinEq.1. periments, and respiration was reduced by perfusing the gills. For comparison, and for ensuring our results are robust, we also Overall, we measured from a total of 166 cells in 28 different D fitted the response of the cell to the different directions with a animals in different stimulus conditions. o w second-orderharmonicfunction(HF)assuggestedbyMaximovetal. n (2005): lo Retinotectal Map of Archer Fish Optic Tectum a d HF(x)(cid:2)a (cid:4)a cos(x(cid:3)(cid:1))(cid:4)a cos(2x(cid:3)2(cid:1)), (2) e 0 1 1 2 2 Torevealtheretinotectalmappingofreceptivefieldsofcells d where a are the amplitude and (cid:1) are the phases of the different in the optic tectum, we mapped the location of 35 cells in 5 fro i i m harmonics. The amplitudes of the harmonics (a0, a1, and a2) were different animals by using an interactively moving bar in h usedtoclassifythecellsintothreedifferentgroups:direction-tuned, different orientations. For each location on the optic tectum ttp oclraiesnsitfiateidona-studnireedc,tiaonnd-tduinreecdtiiofnth-aegirnfiosrstitchcaerlmlso.nSipcehcaifidcdailsltyi,nccetlrleslawtievree surface (see MATERIALS AND METHODS), we found several cells ://jn atahme(cid:6)iprlis0teu.c5doaen,,dia.hen.ad,rmaa1sod(cid:6)niirceahc2taiadonndd-iasagti1nnoc(cid:6)sttri0ec.l5aoatni0ve,seaiasfmtohpreliiertneutldaaettii,ovin.ee-a.t,umanp2el(cid:6)dituoadn1eesasnoidff fteiraoocnmhaldnoidcfafedtirioemnnetonansnitoihmnesa.olspTtahincedtaemvceteuraamsguesruelrdofcatachteeiowrneecareenpdtthivdeeinmficeaenllcsdiuollonatceoad-f .physio 2 0 torevealthemapbetweenthespecificplaceintheoptictectum lo both their first and second harmonics were small (Maximov et al. g 2005). and the location in the visual field (Fig. 2, A and B). y.o Sinusoidal gratings. The stimulus consisted of drifting sinusoidal We found that the dorsal visual field projects to the lateral rg gratings (maximum and minimum light intensity values of 2.66 and part of the optic tectum and the ventral visual field projects to a/ 0.41(cid:1)W/cm2,respectively)indifferentspatialandtemporalfrequen- themedialpartoftheoptictectum.Thevisualfieldinthenasal t M cfoiellsow(Feidg.by1F0)..5Esaocfhbplarceksesnctraeteionn,toofavaogidraatinnygaadpappetaatrieodnefoffrec1t.s5,isn, dteimrepctoiroanl dpirroejcetcitosntoprothjeectrsosttoratlhepacratuadnadl pvaisrutaolffitehledtienctuthme arine 7–12 repeating loops. The direction of the grating was chosen to be B thepreferreddirectionobtainedfromthemovingbarstimulus.Spatial (Fig. 2, A and B). io Inaddition,wefoundthatthereceptivefieldscanbedivided lo frequency varied between 0.03 and 0.7 cycles/deg (6 logarithmic g steps), and temporal frequency varied between 1.5 and 9 cycles/s (6 into two groups according to their location, size, and shape. ica linearsteps),resultinginatotalof36gratingconditions.Firingrates Thefirstgroupconsistsofcellswithreceptivefieldsinfrontof l L foreachgratingwereobtainedbyaveragingthenumberofspikesover thefish’seyeinthedorsal-ventralplane,i.e.,atthesameheight ab the1.5-sstimulusdurationacrossalltrialsandsubtractingthefirst50 as the fish eye. The second group consists of cells with o n ms, to avoid the onset response. Collecting these responses for all receptive fields located more dorsally (Fig. 2B). A stimulus conditions, we then determined the preferred spatial and Exploringtheshapeofthereceptivefields,wefoundthatthe ug temporal frequency combination for each cell. Finally, we evaluated majority of receptive fields that were located in front of the us the spatial frequency tuning at the preferred temporal frequency and fish’s eye in the dorsal-ventral plane were elongated in their t 1 the temporal frequency tuning at the preferred spatial frequency by verticaldimension(Fig.2C,red,yellow,andgreencircles).In , 2 0 fittingthemwithaGaussianfunction. cells with receptive fields located more dorsally, the receptive 13 fields were slightly more elongated in their horizontal dimen- RetinotectalMappingoftheOpticTectum sion than in their vertical dimension (Fig. 2C, blue and cyan circles). A closer examination of the width and height of the To determine the gross mapping between the visual field to the optictectumsurface,wefirsttookaphotooftheoptictectumandaligned receptive fields shows that the receptive fields of cells with ittoagridconsistingofsmallregionsalongtheanterior-posteriorand receptive fields located in front of the fish’s eye in the dorsal- the medial-lateral axes. This procedure divided the optic tectum ventralplanearesignificantlynarrowerandlongerthanrecep- surfacetosmallregionsofinterest(Fig.2A).Onceawell-isolatedcell tive fields of cells with receptive fields located more dorsally, with a high-quality neural response was found, we marked the loca- which were wider and shorter. tionofthepenetrationtotheoptictectumonthephotowepreviously Wedeterminedthesizeofthereceptivefieldasthegeomet- took.Theaccuracyofthedeterminingtheexactpenetrationpointwas ricalmeanofitstwoedges(Fig.2D).Wefoundthattherewas (cid:2)0.4 mm. At the end, we grouped single units to those regions by a significant difference in the size distribution of the two their tectal location. For each unit, we estimated the receptive field groups (Kolmogorov-Smirnov test, P (cid:5) 0.001). Specifically, dimension and location as described in above. We then assigned an small receptive fields (range: 3.12°–9.48°, mean: 5.32°) were averagereceptivefieldlocationanddimensionforeachgridregionby averagingthosepropertiesofallcellsthatcorrespondedtothatregion. locatedabovethetemporal-nasalline.Largereceptivefieldsof This procedure was done in different animals and allowed us to cells(range:5.45°–17.69°,mean:10.14°)werelocatedinfront combinedatafromdifferentexperiments. of the fish eye. JNeurophysiol•doi:10.1152/jn.00094.2013•www.jn.org 752 VISUAL RECEPTIVE FIELDS IN THE ARCHER FISH OPTIC TECTUM A B DORSAL L LATERAL A R N O A P S Fig.2.Spaceandsizeorganizationofreceptive R M A fieldsinthearcherfishoptictectum.A:aphotoof AL OS TE L the archer fish brain with the left optic tectum D T delineated(top)andtheelectrodepositionsonthe U R left optic tectum (bottom). B: topographical ar- A A D C L o rangementofreceptivefieldsintherightvisual w n fieldofthearcherfish.Theblackcirclerepresents 10° lo thefisheye.Eachcolorcorrespondstotheelec- a MEDIAL d trodepositionshowninA.Solidrectanglesrep- VENTRAL ed rdeismeenntstihoen,aavnedracgoentroeucreprteicvteanfigeleldsrleopcraetisoenntatnhde fro m areaintowhichallthereceptivefieldsbelonging C D tattooisotnathhfeeoufsnealrcemetcicoeternpogtodriivfedeiptfsofiahlselie.ltdiiCgosh:nitwz.sehEi.doatwchhnocfionaloArre.ccDeop:rrtdievisseptrofiinbedulds- h (deg) s 76 http://jn.ph Widt Cell 5 ysio d of 4 log ve Fiel umber 32 ay.org/ pti N t M Rece 10 arine 0 5 10 15 20 B Receptive Field Height (deg) Receptive Field Size (deg) io lo g ic Inaddition,weexaminedtheareainwhichallthereceptive We found that most cells respond to both ON and OFF a fieldsbelongingtothesamegridfall(Fig.2B,contourrectan- stimulus and that there is a continuum in the response of cells l La b gles). We found that there is no significant difference in the to this kind of stimulus, since the ON/OFF response ratio of o widths of the circumscribed receptive field areas (permutation different cells spans continuously the interval (cid:7)1 to 1 (Fig. n A test, P (cid:6) 0.05). However, there is a significant difference 3D). In addition, there is a slight preference for the OFF u g between the two groups in the variability in the dorsal-ventral stimulus, since the peak of the response histogram is negative us axis (permutation test, P (cid:5) 0.01). The height of the circum- ((cid:7)0.14 (cid:8) 0.05, mean (cid:8) SE). t 1 scribedareaofthereceptivefieldsclosertothefisheyeissmaller , 2 0 than the height of the circumscribed area of the receptive fields 1 Orientation Selectivity in Archer Fish Optic Tectum 3 locatedmoredorsally. The orientation selectivity of cells was tested by flashing a ArcherFishOpticTectumCellsRespondtoanON/OFFStimulus bar on the receptive field center. This was done first by Wetestedtheresponseofthecellstotheonsetandoffsetof detecting the receptive field boundaries (see MATERIALS AND alightbypresentingafull-fieldflashstimulus.Werecordeda METHODS) and then by flashing a bar with a matched size in total of 83 cells from 15 animals using this type of stimulus. eight different orientations. We recorded 38 cells from 8 Severalcellswithdifferentresponsepropertiesaredepicted animalsusingthistypeofstimulus.Wequantifiedtheresponse in Fig. 3, A–C. Generally, cells responded to the onset and of different cells to orientation by calculating the orientation offset of light in various degrees. To quantify the response index and determined the width of the orientation tuning by profiles,foreachcellwedeterminedtheONresponseleveland fitting a von Mises function only to the cells with high theOFFresponseandcalculatedtheON/OFFresponseratioR, orientation index (for additional details, see MATERIALS AND which quantifies the ON/OFF polarity of the cell. Cells that METHODS). exhibitastrongONresponse(Fig.3A)willhaveresponseratio Wefoundthattheresponseofcellscanbeclassifiedintotwo close to 1, whereas cells that exhibit a strong OFF response types. The first is a strong response to specific orientation, (Fig. 3B) will have response ratios close to (cid:7)1. The response where the response to the preferred orientation could be up to ratio of an ON-OFF cell (Fig. 3C) will be close to 0. 20 times higher than the response to the null orientation, and JNeurophysiol•doi:10.1152/jn.00094.2013•www.jn.org VISUAL RECEPTIVE FIELDS IN THE ARCHER FISH OPTIC TECTUM 753 Response to a Diffuse Flash A B ON Cell OFF Cell # # at at e e p p e e R R Fig.3.CellresponsestoON-OFFstimulus. A:exampleofanONcell,acellthatresponds totheonsetofaflash.B:exampleofanOFF cell, a cell that responds to the offset of a D o 0 1000 2000 0 1000 2000 flash.C:exampleofanON-OFFcell,acell w Time (ms) Time (ms) thatrespondstoboththeonsetandoffsetof nlo aflash.D:adistributionoftheresponseratio a d C D of ON and OFF responses shows that the e ON-OFF Cell cbeeltlwsefeanlloOnNacaonndtinOuuFmFocflaresssepso.nsTehpeattmeronsst d fro 12 abundant pattern is a symmetrical ON-OFF m response. h 10 ttp # Cells 8 ://jn.p peat r of 6 hysio Re mbe 4 logy u .o N rg 2 a/ t M 0 1000 2000 0 a Time (ms) -1 -0.R5 e s p o n 0s e R a t i 0o.5 1 rine B therefore the orientation index is (cid:6)0.5 (Fig. 4, A and B). In cellsinthearcherfishoptictectumhaverelativelysharporienta- iolo addition,wefoundcellswithveryweakdependenceonorien- tiontuningprofiles. gic tation,wherethedifferencebetweenthemaximalandminimal a l L orientation spike count is marginal, for which the orientation Direction and Orientation Selectivity of Tectal Neurons in a index is (cid:5)0.5 (Fig. 4, C and D). We define cells with such Response to Moving Bars b o weak dependence on orientation as orientation-agnostic neu- n A rons. Thedirectionselectivityofcellsinresponsetomovingbarswas u g Using this analysis, we found that 58% of the cells were testedbyfirstlocatingthereceptivefieldcenterandthenmovingabar u s orientation tuned (Fig. 4E). We tested the sensitivity of these acrossthereceptivefield(seeMATERIALSANDMETHODS).Werecorded t 1 numbers to the threshold between cell types and found that 36cellsfrom8animalsusingthisstimulus.Toclassifycellsinto , 2 0 10% change in the threshold results with only 1 cell moving different response groups, we first calculated the spike count in 1 3 between groups. Moreover, the preferred orientation had a response to specific direction (see MATERIALS AND METHODS) and strong bias for the vertical orientation with 59% of the orien- thencalculatedthedirectionindexandtheorientationindex.We tation-tuned cells having a preference for the vertical orienta- then fitted a von Mises function to the response to measure the tion or its neighboring orientations ((cid:8)22.5°), 13.6% of the directiontuningwidth.Forcomparison,wealsofittedaharmonic orientation-tuned cells having a preference for the horizontal function to the response to determine whether our results are orientationoritsneighboringorientations((cid:8)22.5°),and27.2% robust(seeMATERIALSANDMETHODSformoredetails). oftheorientation-tunedcellshavingapreferencefortheother diagonal orientations or its neighboring orientations ((cid:8)22.5°) Wefoundthreetypesofcellsaccordingtotheirresponsetoa moving bar: 1) an orientation-tuned response, where the cell (Fig. 4F). respondswithhighfiringratetomovementsalongaspecificaxis WefittedvonMisesfunctiontotheorientation-tunedcellsand explored the parameter (cid:6)to find their orientation tuning width. in space (Fig. 5A); 2) a direction-tuned response, where the cell Wefoundthat45%ofthecellshadasharptuningwidthbetween respondswithhighratetomovementinaspecificdirection(Fig. 15°and30°,41%ofthecellshadamediumtuningwidthbetween 5B); and 3) an agnostic response, where the cell responds with 30°and45°,and14%ofthecellshadatuningwidthof45°to60°. marginaldifferencesbetweendirections(Fig.5C). Onaverage,meanorientationtuningwidthwas33.26°(SE2.36°; Onthebasisofthequantitativeanalysisdescribedabove,we Fig.4G);i.e.,themostprevalentwidthwas(cid:5)34°,indicatingthat found that 11% of the cells were direction tuned, 39% were JNeurophysiol•doi:10.1152/jn.00094.2013•www.jn.org 754 VISUAL RECEPTIVE FIELDS IN THE ARCHER FISH OPTIC TECTUM Response to a Flashing Bar A B Orientation-Tuned Cell 60 nt u n o o C ati e 40 ent pik Ori e S ar ag 20 B er v A 0 Fig.4.Orientationselectivityprofilesassessed 0 1000 2000 usingthestaticbarstimulus.A:rasterplotsof Time (ms) Bar Orientation D the responses to the 8 oriented bars for an ow orientation-tunedcell.Blacklinesrepresentthe C D n startandendtimeofthebardisplay.B:orien- Orientation Agnostic Cell lo a tation tuning profile for the orientation-tuned 20 de cell from A. The dashed black line repre- nt d sentsthefittedvonMissesdistributionfunction. n ou fro C:rasterplotsoftheresponsestothe8oriented o C 15 m barsforanorientation-agnosticcell.D:orien- ati e h tc9ae5tli%lofnrcotoumnnfiiCndg.enEpcr:eodfiiinlsettrefirbovurattlih)oenfooorrifoesnreitelaentcitoatintvi-oiatnyg-nt(uownsitetihdc Orient e Spik 10 ttp://jn andorientation-agnosticcells.F:distributionof ar ag 5 .ph preferredorientation(with95%confidencein- B er ys terval).Themajorityofcellspreferredthever- Av io ticalorientation.G:distributionoftuningwidth 0 log fortheorientation-tunedcells. 0 1000 2000 y Time (ms) Bar Orientation .org a/ E F G t M a ells 70 n=38 ells 70 n=22 ells 40 n=22 rine C 60 C 60 C B entage of 54320000 entage of 54320000 entage of 321000 iological L rc 10 rc 10 rc ab Pe 0 Agnostic OSreielenctatitvioen Pe P0referred Orientation Pe 0 10-2T0 u 2n0i-3n0g 3W0-4id0 t h40 -(5d0 e g50)-60 on A u g u s orientation tuned, and 50% were classified as direction agnos- Using the second method, fitting of a harmonic function, to t 1 tic(Fig.5D).Wetestedthesensitivityofthesenumberstothe determine the selectivity of the cell in response to a moving , 2 thresholdbetweencelltypesandfoundthat10%changeinthe bar, we found similar results with a few exceptions. All the 01 3 threshold results with 5 cells moving between groups. cells that were classified as direction-tuned cells with the first Of the orientation-tuned cells, 86% preferred the vertical ori- method were also classified as direction-tuned cells with the entation (as shown in Fig. 5A) and 14% preferred the diagonal second method. The second method classified an additional orientation. We calculated the confidence interval for the possi- cellasadirection-tunedcell.Asfortheorientationtuning,the second method missed two cells that were classified as orien- bilityoftheexistenceofanadditionalcelltypethatwehavenot tation-tunedwiththefirstmethodandaddedanadditionalcell observedinourdataset(seeMATERIALSANDMETHODS),orientation- asorientation-tuned.Overall,itseemsthatuseofthedirection selectivecellswithapreferenceforthehorizontalorientation.We and orientation indexes and the von Mises distribution gives found that with 95% of certainty, the proportion of this possible robust results, and moreover, it adds the ability to determine missinggroupisbetween0and0.1. the tuning width, compared with the second method. Wefoundthathalfofthedirection-tunedcellshadthegreatest response to a moving bar in the temporal-nasal direction, as Spatial and Temporal Frequency Tuning determined by exploring the fitted parameter (cid:1) in Eq. 1 and shown in Fig. 5B. The other direction-tuned cells preferred the The spatial and temporal tuning of cells was evaluated by diagonaldirection,onefromnasal-ventraltocaudal-dorsaldirec- analyzingtheresponseofcellstosinusoidalgratingdriftingalong tion and the other from nasal-dorsal to caudal-ventral direction. thecell’spreferreddirection.Oncethecellpreferreddirectionwas Thewidthsofthedirectiontuningvariedbetween60°and90°. detected, we stimulated it with all combinations of six spatial JNeurophysiol•doi:10.1152/jn.00094.2013•www.jn.org VISUAL RECEPTIVE FIELDS IN THE ARCHER FISH OPTIC TECTUM 755 Response to a Moving Bar A B Orientation-Tuned Cell Direction-Tuned Cell Dorsal Nasal Temporal D o w Fig.5.Directionselectivityprofilesassessed n usingthemovingbarstimulus.A:orientation- lo 1 s tunedcellprofile.B:direction-tunedcellpro- ad Ventral file.C:direction-agnosticcellprofile.D:dis- ed tirnitbeurvtiaoln) ofofrsealgencotisvtiicty, o(wriietnhta9t5io%n-ctuonnefidd,eanncde from C D direction-tunedcells. h ttp Direction Agnostic Cell ://jn n=36 .p h y s io lo g y .o rg a/ t M a rin e B Agnostic OSreielenctatitvioen SDeirleeccttiiovne iolo g ic a l L frequencies(rangingfrom0.03to0.7cycles/deg)andsixtemporal that we have not observed in our data (see MATERIALS AND ab frequencies (ranging from 1.5 to 9 Hz). For this stimulus, we METHODS).Wefoundthatwith95%ofcertainty,theproportion on recordedatotalof30cellsfrom6animals. of a possible missing group is between 0 and 0.12. A u For each cell, we found the combination of spatial and In a similar way, to analyze the spatial tuning of a cell, we g u temporal frequencies that elicited the maximal response. The tookitsbesttemporalfrequencyandfittedaGaussianfunction s cells had a higher response to a unique combination of spatial totheresponseatallcorrespondingspatialfrequencies(i.e.,to t 1 and temporal frequencies (Fig. 6, A and B). 1columnoftheresponsematrix).Wefoundtwogeneraltypes , 2 0 We analyzed the cell response to obtain further frequency oftuningprofiles:lowpassandbandpass.Figure6Fshowsa 13 tuningproperties.Toanalyzethetemporalfrequencytuningof low-pass tuning profile. The maximal response of this cell to a cell, we took the best spatial frequency and analyzed the the sinusoidal grating is at the lowest spatial frequency tested cellular response for all the different temporal frequencies at (0.03 cycles/deg), and at higher spatial frequencies the re- that spatial frequency (i.e., 1 row of the response matrix). sponse gradually decreases. Figure 6G shows a band-pass Fitting a Gaussian function to these responses, we obtained tuning profile. The response of this cell to the sinusoidal three types of temporal frequency tuning profiles: a low-pass gratingatthelowestspatialfrequencytested(0.03cycles/deg) tuningprofile,inwhichcellselicitthemaximalresponsetothe andatthehighestspatialfrequencytest(0.7cycles/deg)islow, lowertemporalfrequenciestested(Fig.6C);ahigh-passtuning and the maximal response is for a spatial frequency in the profile, in which cells elicit the maximal response to the middle of the range (0.09–0.18 cycles/deg). We found that highest temporal frequencies tested (Fig. 6D); and an all-pass 57%ofthecellshadalow-passtuningprofileand43%ofthe tuning profile, in which there was almost no difference in the cells had a band-pass tuning profile (Fig. 6H). We calculated response to the different temporal frequencies (Fig. 6E). We theconfidenceintervalforthepossibilityoftheexistenceofan foundthat23%ofthecellshadalow-passtuningprofile,54% additional tuning profile group that we have not observed in hadahigh-passtuningprofile,and23%hadanall-passtuning ourdata(seeMATERIALSANDMETHODS).Wefoundthatwith95% profile(Fig.6H).Wecalculatedtheconfidenceintervalforthe of certainty, the proportion of a possible missing group is possibilityoftheexistenceofanadditionaltuningprofilegroup between 0 and 0.12. JNeurophysiol•doi:10.1152/jn.00094.2013•www.jn.org 756 VISUAL RECEPTIVE FIELDS IN THE ARCHER FISH OPTIC TECTUM Response to a Sinusoidal Grating A B 1 1 g) 0.7 g) 0.7 e 0.9 e 0.9 d N d N e/ 0.8 o e/ 0.8 o cycl 0.35 0.7 rma cycl 0.35 0.7 rma ncy ( 0.18 0.6 lized ncy ( 0.18 0.6 lized e 0.5 S e 0.5 S equ 0.09 0.4 pik equ 0.09 0.4 pik atial Fr 0.05 00..23 e Coun atial Fr 0.05 00..23 e Coun Down p t p t lo S 0.03 0.1 S 0.03 0.1 a d 0 0 ed Te1.m5por3al Fre4.q5uen6cy (c7y.c5le/s9) Te1.m5por3al Fre4.q5uen6cy (c7y.c5le/s)9 from C D E http ://jn .p nt nt nt hy u u u s o o o io e C e C e C log k k k y pi pi pi .o S S S rg alized alized alized at Ma/ orm orm orm rine N N N B io lo Temporal Frequency (cycle/s) Temporal Frequency (cycle/s) Temporal Frequency (cycle/s) g ic a F G H l La 80 n=30 b o n ount ount Cells 7600 Augu d Spike C d Spike C entage of 543000 st 1, 2013 e e c z z r ali ali Pe 20 m m 10 r r o o N N 0 Low High All Low Band Pass Pass Pass Pass Pass Spatial Frequency (cycle/deg) Spatial Frequency (cycle/deg) Temporal Tuning Spatial Tuning Fig.6.Responsestospatialandtemporalfrequenciesandtuningprofilesassessedusingsinusoidalgratings.A:exampleofacellforwhichthemaximalresponse rate is at the lowest spatial and temporal frequencies tested. B: example of a cell for which the maximal response rate is at one of the central spa- tialandthehighesttemporalfrequenciestested.C:low-passtemporalfrequencytuningprofileshowingthehighestresponserateatthelowesttemporalfre- quencytested.D:high-passtemporalfrequencytuningprofileshowingthehighestresponserateatthehighesttemporalfrequencytested.E:all-passtemporal frequencytuningprofile.F:low-passspatialfrequencytuningprofileshowingthehighestresponserateatthelowestspatialfrequencytested.G:band-passspatial frequencytuningprofile.H:distributionoftuningprofiles(with95%confidenceinterval).Lightgray,temporalfrequencytuning;darkgray,spatialfrequency tuning.Wecalculatedtheconfidenceintervalforthepossibilityoftheexistenceofanadditionaltuningprofilegroupthatwehavenotobservedinourdata(see MATERIALSANDMETHODS).Forboththespatialfrequencyandtemporalfrequencytuning,wefoundthatwith95%ofcertainty,theproportionofapossiblemissing groupisbetween0and0.12. JNeurophysiol•doi:10.1152/jn.00094.2013•www.jn.org

Description:
Downloaded from system. archer fish; optic tectum; early visual processing; retinotectal map- ping; orientation .. parameters are p, peak response; , center orientation at the peak; and .. to the electrode position shown in A. D: distribu- .. archer fish live in water that is a three-dimensional w
See more

The list of books you might like

Upgrade Premium
Most books are stored in the elastic cloud where traffic is expensive. For this reason, we have a limit on daily download.
DMCA & Copyright: Dear all, most of the website is community built, users are uploading hundred of books everyday, which makes really hard for us to identify copyrighted material, please contact us if you want any material removed.
Copyright @ PDFdrive.to, 2022 | We love our users