topological and modality-specific representation of … · drosophila has six major types of...

NEUROSCIENCE

Topological and modality-specificrepresentation of somatosensoryinformation in the fly brainAsako Tsubouchi,1* Tomoko Yano,1,2* Takeshi K. Yokoyama,1 Chloé Murtin,1,2

Hideo Otsuna,3 Kei Ito1,2,3,4†

Insects and mammals share similarities of neural organization underlying the perception ofodors, taste, vision, sound, and gravity.We observed that insect somatosensation alsocorresponds to that of mammals. In Drosophila, the projections of all the somatosensoryneuron types to the insect’s equivalent of the spinal cord segregated into modality-specificlayers comparable to those in mammals. Some sensory neurons innervate the ventral braindirectly to form modality-specific and topological somatosensory maps. Ascendinginterneurons with dendrites in matching layers of the nerve cord send axons that converge torespective brain regions. Pathways arising from leg somatosensory neurons encode distinctqualities of leg movement information and play different roles in ground detection.Establishment of the ground pattern and genetic tools for neuronal manipulation shouldprovide the basis for elucidating the mechanisms underlying somatosensation.

Responding to external stimuli is a crucialfunction, even for the simplest unicellulareukaryotes, and these organisms share spe-cific molecules that are used for sensorydetection in higher animals (1). In meta-

zoans, the organization of sense organs and as-sociated sensory centers in the brain often showsconserved organizational principles for manysensory modalities, even between the species be-longing to distant evolutionary clades such asinsects and mammals (2–5 ). Is there such cor-respondence in the somatosensory system?Mam-malian and insect somatosensory systems initiallyappear rather different. In mammals, somato-sensory neurons derive from the neural crest, andthey reside deep in the body in the dorsal rootganglia. They send their axons to specific lami-nae in the dorsal spinal cord, from where infor-mation is sent to different parts of the thalamus,depending on somatosensory modality (6, 7). Ininsects, somatosensory neurons derive from theepidermis and reside directly beneath the exo-skeletal body surface (Fig. 1A) (8). Terminals ofdifferent types of insect somatosensory neu-rons tend to form a layered organization in theventral nerve cord (VNC), the ganglia that areequivalent to the vertebrate’s dorsal spinal cord(9, 10). However, despite extensive studies on thestructure and functional properties of primaryneurons and associated local secondary neu-rons in the VNC (11–14), knowledge about theirpatterning and connections to the higher brainhave remained limited (15, 16). Here we provide

the first systematic mapping of insect somato-sensory neural circuitry using Drosophila mela-nogaster as a model.

Modality-specific layered organizationof the somatosensory axon terminalsin the VNC

To understand the global architecture of the so-matosensory pathways in the Drosophila centralnervous system (CNS), we first analyzed the or-ganization of somatosensory axon terminals inthe VNC. There are a number of GAL4 expres-sion driver strains to visualize somatosensoryneurons (8, 14). We screened more than 10,000promoter-fusion and enhancer-trap GAL4/LexAdriver lines (17–21) and identified an array of 22strains. Each strain labels specific types of soma-tosensory neurons in the legs, wings, halteres,and abdominal body surface (figs. S1 and S2). Weselected a set of strains so that they together labelessentially all of the somatosensory cells. Usingthese specific driver lines, we then mapped thedistribution of axon terminals in the respectiveneuropils of the VNC (Fig. 1 and fig. S3).Drosophila has six major types of somato-

sensory neurons (Fig. 1A). Gustatory sensilla (gs)neurons detect food stimuli through the pores ofexternal taste hairs (22, 23). External sensilla (es)neurons detect movement of the bristles causedby contact, wind, etc. (24). Stretch receptor (sr)neurons contact neighboring leg joints to detectstretch and tension between them (25). The chor-dotonal organs (co) have thin elongated struc-tures that span across different parts of theexoskeleton to detect tension or vibration (26, 27).Campaniform sensilla (cs) feature dome-shapedthin cuticles to detect deformation of the exo-skeleton (28, 29). Finally, multidendritic (md)neurons extend complex dendrites spanning un-der the epidermis to detect diverse stimuli suchas touch, pain, heat, and coldness (30).

Among these different types of somatosensoryneurons, the legs feature all but the last neurontype (i.e., gs, es, sr, co, and cs) (Fig. 1, B and C, andfig. S1) (8, 24). Most of their axons terminate inthe bulbous leg neuropils of the VNC (Fig. 1G).Previous studies suggested a layered organiza-tion of some neuron types (e.g., es lying ventrallyand cs and co dorsally) (10, 13, 24); we found thataxon terminals of all five types of leg somato-sensory neurons have clearly segregated layeredprojections (Fig. 1, H and K, and movie S1). Theleg gs neurons are distributed in the distalmostleg segments (fig. S1, A and B). They express 27gustatory receptor genes and were categorizedinto seven types according to their projection pat-terns (23). One of the sugar-detecting neurontypes (VT, expressing Gr5a) arborizes only in thethoracic VNC, which we here call Lgs1NT (where“L” stands for the leg, “1N” for primary neuron,and “T” for thorax). All other gs neurons projectboth to the VNC and the brain, which we col-lectively call Lgs1NB (“B” for brain). In the VNC,axons of both Lgs1NT and Lgs1NB terminate inthemost ventral part of the leg neuropil (Fig. 1H1and fig. S3, A and B).The leg es neurons (Les1N) are studded along

the entire leg surface (fig. S1C) (24). Their axonterminals in the VNC lie slightly more dorsallythan that of the leg gs neurons, in the shape ofa flat disc that horizontally spreads across theventral leg neuropil (Fig. 1H2 and fig. S3C). Thedisc features concentric fine-scale somatotopy.We surgically cut off the legs at different jointsand observed the projection patterns 2 weeks laterafter axons arising from the excised parts aredegenerated. Axons of the es neurons in the dis-tal leg segments terminate in the central parts ofthe disc, whereas those deriving from the moreproximal segments terminate inmore peripheralrings (fig. S3H).Leg sr neurons (Lsr1N) are scattered in several

leg joints (fig. S1D) (25); in the VNC they arborizeat the level just above the disc of the es neuronterminals (Fig. 1H3 and fig. S3D). Most of the legco neurons express the inactive gene (Lco1N1)(fig. S1E) (31), whereas another line labels a dif-ferent small subset (Lco1N2) (fig. S1F). Lcs1N labelsa subset of leg cs neurons (fig. S1G). The axons ofco and cs neurons terminate in the most dorsallayer, just in themiddle of the leg neuropil (Fig. 1,H4 andH5, and fig. S3, E to G). Their projectionsare segregated within a layer: The axon termi-nals of the co neurons form several thick bundles(fig. S3, F2 and G2), and cs neurons terminatein the space between these bundles (fig. S3E2).Cross-sectional views of the registered imagesshow that the presynaptic sites of different neurontypes hardly overlap (fig. S3I). Thus, the leg neu-ropils have amodality-specific layeredorganizationin the order of gs, es, sr, co, and cs axon terminalsfrom ventral to dorsal (Fig. 1, H6 and K, andmovie S1).Somatosensory neurons of the thoracic body

surface (es) and wings (gs, es, and cs) send theiraxons to the wing neuropil (32–34) (Fig. 1, Dand G). The thoracic es neurons scattered onthe body surface (Tes1N) (32) project broadly

RESEARCH

Tsubouchi et al., Science 358, 615–623 (2017) 3 November 2017 1 of 8

1Institute of Molecular and Cellular Biosciences, The Universityof Tokyo, Yayoi, Bunkyo-ku, 113-0032 Tokyo, Japan.2Department of Computational Biology, Graduate School ofFrontier Sciences, The University of Tokyo, Kashiwanoha,Kashiwa, 277-0882 Chiba, Japan. 3Howard Hughes Medical Institute,Janelia Research Campus, Ashburn, VA 20147, USA. 4Institute ofZoology, University of Cologne, 50674 Cologne, Germany.*These authors contributed equally to this work. †Correspondingauthor: Email: [email protected]

on January 9, 2021

http://science.sciencemag.org/

Dow

nloaded from


to the most ventral layer of the wing neuropil(Fig. 1I1 and fig. S3J). The wing gs neurons(Wgs1N) and wing es neurons (Wes1N) (33, 35)are located along the anteriorwingmargin (fig. S2,A and B). They terminate in the wing neuropilin the form of thin bundles: Axons of Wgs1N(Fig. 1I2 and fig. S3K) terminate just above theterminals of Tes1N (Fig. 1I1 and fig. S3J), whereasWes1N terminate slightly more dorsally (Fig. 1I3and fig. S3L). Finally, wing cs neurons (Wcs1N)(34) are located on the dorsal and ventral sur-faces of the wing base, as well as on the L3 wingvein (fig. S2, C and E). Their axon terminals arelocated further dorsally in the middle and dor-

solateral parts of the wing neuropil (Fig. 1I4and fig. S3M). Thus, the wing neuropil has fourmodality-specific layers in the order of thoracices and wing gs, es, and cs (Fig. 1, I5 and K, andmovie S2).Somatosensory neurons in the haltere ter-

minate in the haltere neuropil (Fig. 1, E andG). Unlike wings, halteres have no gs neuronsand only a few es neurons. Haltere cs neurons(Hcs1N) (34) are located on the dorsal and ven-tral sides of its basal part (fig. S2, D and F);their axons terminate in a comparable dorsal level(fig. S3N) as that of the wing cs neurons in thewing neuropil (fig. S3M).

The abdominal body surface features threetypes of mechanosensory neurons (md, es, andco) (Fig. 1F). The abdominalmdneurons (Amd1N)extend dendrites under the dorsal and lateral bodywalls (fig. S2G). Abdominal es neurons (Aes1N)(36) are located on the dorsal and ventralmostabdomen (fig. S2H), and a few co neurons (Aco1N)(37) lie along the ventralmidline (fig. S2I). Amongthem, the abdominal co neurons do not arborizein the abdominal ganglia; instead, they projectdirectly to the prothoracic ganglion to convergewith the terminals of the foreleg co neurons. Themd and es neurons terminate in the abdominalganglia to form segregated layers; axons of the


Fig. 1. Layeredsomatosensoryaxon terminals inthe ventral nerve cord.(A) Schematics ofmechanosensoryorgans. Cross-sectionalview. The planar viewis also shown for therightmost panel. Redand blue indicateneurons and supportcells, respectively.Scale bars, 10 mm. (B toF) Schematics of theentire adult fly (B),somatosensory neuronsin the leg (C), wing (D),haltere (E), and abdom-inal body wall of thefourth and fifthsegments (F). See figs.S1 and S2 for detail.(G) Entire neuropilof the ventral nervecord (VNC). Three-dimensional (3D)reconstruction image ofthe nc82 antibodylabeling. Dotted rectan-gles indicate the areasvisualized in thefollowing panels. (H1to H6) 3D reconstruc-tion of the layered axonterminals in the legneuropil. Fluorescentsignals were segmentedusing Amira software.The anterior obliqueview is shown. Lgs1N[(H1), magenta],Les1N [(H2), green],Lsr1N [(H3), pink], Lco1N[(H4), yellow], Lcs1N[(H5), blue], and mergedimage (H6). See fig. S3,A to G, and movie S1 for detail. (I1 to I5) 3D reconstruction of the layeredaxon terminals in the wing neuropil. Tes1N [(I1), green],Wgs1N [(I2), purple],Wes1N [(I3), yellow],Wcs1N [(I4), blue], and merged image (I5). See fig. S3,J to M, and movie S2 for detail. (J1 to J3) 3D reconstruction of the layeredaxon terminals in the abdominal ganglia. Amd1N [(J1), magenta], Aes1N

[(J2), green], and merged image (J3). See fig. S3O for detail. (H to J)Directions with arrows: D (dorsal), V (ventral), L (lateral), M (medial), andA (anterior). Scale bars, 50 mm. (K) Schematics of the arrangement ofpresynaptic terminals for each type of primary somatosensory neurons in thefrontal cross section.

D

E

C Fsr

sr

md

md

dorsalmidline

ventralmidline

AbdomenLeg

Haltere

Wing

H2 H3 H5H1 H4 H6Lco1N Lcs1NLsr1NLes1NLgs1N

Leg neuropil: 3D reconstruction (right side, anterior oblique view)

Wing neuropil: 3D reconstruction (right side, anterior oblique view)

I1 I3I2 I4Wgs1N Wcs1NWes1NTes1N I5

AV

L

K

Amd1N

abdominal ganglia

V

D

Wcs1N

Wgs1N

wing neuropil

Tes1N

Les1N

leg neuropil

Lgs1N

Wes1NWgs1NgsWgs1NWgs1NWgs1NWgs1N

Lco1NLcs1N

Lsr1N

I

J

G VNC (ventral view)

H

abdominalganglia

wing neuropil

leg neuropil

haltere neuropil

proximal

proximal distaldistal

distal

cs cs

cs cs

cs

cs

cs

cs

cs

gs

co co

co

B

Leg

WingHaltere

Abdomen

AV

L

Abdominal ganglia: 3D reconstruction (both sides, frontal cross section)

J1 J3J2 Aes1NAmd1N

Lco1NLcs1N

Lsr1NLes1NLgs1N

Wgs1N

Wcs1NWes1N

Tes1N

Aes1NAmd1N

V

D

Adult fly

esgs

gs

es

es

es

es

es

es

es

es

eses

es

Aes1N

proximal

leg gustatory sensilla leg external sensilla leg stretch receptors leg chordotonal organ

thoracic external sensilla wing gustatory sensilla wing external sensilla

abdominal multidendrite abdominal external sensilla merge

wing campaniform sensilla merge

leg campaniform sensilla merge

L M

Gustatory sensilla (gs) External sensilla (es) Stretch receptor (sr) Chordotonal organ (co) Campaniform sensilla (cs) Multidendritic neuron (md)

taste bristle contact stretch, tension tension, vibration deformation touch, pain, heat, cold

A

coxa

femur

tibia

tarsus

trochanter

RESEARCH | RESEARCH ARTICLEon January 9, 2021


Dow

nloaded from


md neurons project to the most ventral layer,whereas es neurons arborize in a more dorsallayer (Fig. 1J and fig. S3O). Thus, the abdominalVNC neuropils again have segregated layers thatconsist of md and es axon terminals, respectively(Fig. 1, J and K).Our analysis revealed that although the rep-

ertoires of somatosensory neuron types that ter-minate in the leg, wing, haltere, and abdominalVNC neuropils are different, all neuropils havemodality-specific layered organization of soma-tosensory axon terminals.

Modality-specific and somatotopicorganization of somatosensory axonterminals in the brain

We next investigated what parts of the brain re-ceive somatosensory information, bearing in mindtwo likely arrangements: direct pathways by theaxons of the peripheral somatosensory neuronsand indirect pathways by the secondary inter-neurons ascending from the VNC.We first anal-yzed the former (Fig. 2A).Most of the wing and haltere cs neurons form

two axon branches that terminate in the VNCand project to the brain, respectively. The latterform a single common bundle, and upon enter-ing the brain they spread into four sub-bundlesthat fan out in the posterior gnathal ganglia[GNG (Fig. 2D and fig. S4A); see definitions bythe Insect Brain Name Consortium (38)]. Thelateral two sub-bundles extend dorsolaterallyto arborize in the region of the GNG just posteriorto zones C and E of the antennalmechanosensoryand motor center, which receives informationabout antennal deflection (5, 39), whereas themedial two sub-bundles terminate more proxi-mally in the posterior medio-ventral GNG.Because the axons arising from wing and hal-

tere cs neurons project via a common bundle fromthe VNC, their projection targets are not imme-diately apparent. To analyze topological organi-zation of their terminals, we surgically excisedthe wing or haltere of one side and observed theprojection patterns after axonal degeneration.Only the lateral two sub-bundles remained whena haltere was cut off (fig. S4E); they are thereforespecific to the wing cs neurons (named Wl andWmforwing-lateral andwing-medial sub-bundles,respectively). Conversely, the two medial sub-bundles remained after wing excision (fig. S4F),showing their origin from the haltere cs neurons(namedHl andHm sub-bundles). Thus, wing andhaltere cs neurons form a distinct somatotopicsensorymap in the formof parallel sub-bundles inthe posterior GNG.Whereas a type of sweetness-responsive gs

neurons (Lgs1NT) terminate within the VNC, allother gs neurons detecting sweetness, bitter-ness, contact pheromones, etc. (collectively calledLgs1NB) project their axons directly to the mid-dle part of the GNG (23) (Fig. 2E and fig. S4B).They terminate in the brain region just posteriorto the terminals of the bitterness-responsive gus-tatory sensory neurons of the proboscis (20).Most of the leg co neurons terminate within

the VNC, but some project toward the brain.We

distinguish those subgroups as T and B for thethorax and brain, respectively. The axons of bothLco1N1B and Lco1N2B subpopulations form asingle bundle that runs anterior-dorsally in thebrain along the lateral side of the GNG (Fig. 2,F and G, and fig. S4, C and D) and wedge (WED)neuropils to terminate in the ventralmost part ofthe anterior ventrolateral protocerebrum (AVLP).Do co neurons arising from different legs makea somatotopic map in this axon bundle? To ad-dress this question, we surgically cut off two ofthe three legs on one side of the animal and ob-served projections after axonal degeneration.The remaining axons terminate at different po-sitions along the trajectory; axons from the fore-leg and midleg project toward the distal end anda slightly proximal part of the bundle, respec-tively (fig. S4, G andH), whereas those from thehindleg terminate much more proximally (fig.S4I). Thus, the leg co neurons also make a clearsomatotopic map in the form of longitudinalprojection (fig. S4J).

Modality-specific convergence ofprimary neurons and secondaryinterneurons in the brain

Because many somatosensory neurons terminatewithin the VNC, information should also be con-veyed to the brain via secondary ascending in-terneurons. If we can identify interneuronswhosedendritic arbors arise in a specific somatosen-sory layer of the VNC neuropils, they would belikely candidates to convey modality-specific in-formation to the brain.We screened GAL4 driverlines and identified six types of such interneu-rons (Fig. 2, B and H to M; fig. S5, G to L; andmovie S3).One type of wing cs interneurons (Wcs2N1,

where “2N” stands for putative secondary inter-neuron) and one type of haltere cs interneurons(Hcs2N1) have postsynaptic sites in matchinglayers of the primary cs neuron axon terminalsof the wing and haltere in the respective neuro-pils (fig. S5, A andB). In the brain, they terminatein the distal and proximal regions of the pos-terior GNG (Fig. 2, H and I, and fig. S5, G andH),where they converge with that of the primaryneurons Wcs1N1 and Hcs1N1, respectively (Fig.2N and fig. S5M).A comparable trajectory was also observed in

the leg co neurons. The postsynaptic sites of thetwo types of putative leg co interneurons (Lco2N1and Lco2N2) overlap with the terminal branchesof the primary neurons Lco1N1 in the leg neuro-pil (fig. S5, C and D). In the brain, Lco2N1 axonsproject via the same trajectory as that of the pri-mary co neurons (Fig. 2J and fig. S5I), whereasLco2N2 axons run more medially through themiddle of the GNG (Fig. 2K and fig. S5J). Never-theless, they eventually converge in the sameregion of the ventralmost AVLP neuropil, whichoverlaps with the distalmost projection of theprimary co neurons Lco1N1B and Lco1N2B (Fig.2O and fig. S5N).Because none of the es neurons project directly

to the brain, their information should be trans-mitted via interneurons. Do they supply the same

brain region as the mechanosensory neurons ofother modalities? We found two types of inter-neurons (Les2N1 and Les2N2) that have post-synaptic sites in the ventralmost leg neuropillayers that receive terminals of Les1N1 as well asLgs1N1 (fig. S5, E and F). Les2N1 and Les 2N2 havethree [dorsal (D), middle (M), and ventral (V)]and two (D and V) axon projections in the brain,respectively (Fig. 2B), some of which are reminis-cent of the previously reported taste-associatedinterneurons TPN1 and TPN 2 (40) (fig. S5, E andF). The ventral projections (Les2N1VandLes2N2V)terminate in the middle part of the GNG thatis adjacent to the terminals of the leg gs neu-rons. Other projections pass through the GNG toproject further dorsally: Les2N1D and Les2N1Mterminate in the dorsal and middle parts of theposterior ventrolateral protocerebrum (PVLP),respectively, whereas Les2N2D terminates in theventral part of the superior lateral protocerebrum(SLP) (Fig. 2, B, L, M, and P, and fig. S5, K, L, andO). The targets of those indirect pathways arisingfrom the leg es and gs neurons do not overlapwith, but lie slightly above and medial to, thedirect and indirect targets of the leg co neurons.These results suggest a rather simple archi-

tectural concept of the putative somatosensorycenters in the fly CNS (Fig. 2Q). First, axonsarising from different types of primary somato-sensory neurons terminate in different layers ofthe VNC and different subregions of the brain.Second, direct and indirect pathways of the samesensory modality tend to converge into overlap-ping subregions of the brain. These character-istics infer that specific subregions of the brainreceive distinct types of somatosensory informa-tion, forming somatotopic and modality-specificsensory representation.

Difference of mechanosensory informationconveyed to the brain by different typesof primary and secondary neurons

So far we found that different somatosensoryneurons of the same modality send overlappinginformation to certain subregions of the brain.To reveal differences of their functional proper-ties, we next examined the activity of the primaryand secondary neurons using cell-specific cal-cium imaging. We placed the heads of the fliesbeneath a thin metal plate, so that the tetheredflies could move their wings and legs freely on atrackball, and observed activity of the axon ter-minals of the neurons expressing GCaMP6s (41)using a two-photon microscope (Fig. 3A andmovie S4).Because cs neurons respond to the deforma-

tion of the cuticular epidermis (29), wemeasuredthe activity of Wcs1N and Hcs1N in the brainwhen we mechanically moved the wing or hal-tere with a tungsten rod. Only the lateral (Wl andWm) or medial (Hl and Hm) axon sub-bundlesshowed activity when we specifically stimulatedthewing or haltere, respectively (Fig. 3, B toD, andfig. S6, A and B). Thus, the somatosensory repre-sentation we anatomically revealed earlier (fig.S4, E and F) is confirmed physiologically as theselective activity patterns of axon sub-bundles.




Dow

nloaded from



Fig. 2. Projection of primary and secondary neurons in the brain. (A andB) Schematics of the projection patterns of the identified mechanosensoryneurons (A) and ascending interneurons (B), dorsal view. Cell bodies andarborizations are schematizedwith open and painted circles, respectively. Brainregions that contain terminal arborizations are painted in pale colors.(C) Anterior view of the entire neuropil of the brain. 3D reconstructionimages of the nc82 antibody labeling. Dotted rectangles indicate the areasvisualized in the following panels. (D to G) Neuronal fibers (magenta) andpresynaptic terminals (green-white) of the primary somatosensory neu-rons, anterior view. See fig. S4, A to D, for lateral view. Axons derivingfrom Wcs1N (Wl and Wm in yellow arrows) and Hcs1N (Hl and Hm in whitearrows) (D), Lgs1NB (E), Lco1N1B (F), and Lco1N2B (G) are shown. Comparedwith the axons of Lco1N1B, those of Lco1N2B in the AVLP neuropil are longer

and have branched terminals at their distal ends (blue arrow). (H toM)Neuronalfibers (magenta) and presynaptic terminals (green-white) of the secondaryascending interneurons in the brain, anterior view. See fig. S5, G to O, forlateral view.Wcs2N1 (H), Hcs2N1 (I), Lco2N1 (J), Lco2N2 (K), Les2N1 (L), andLes2N2 (M). (N to P) Overlay of the axons of primary and secondaryneurons. (N) Wcs1N (green), Hcs1N (blue),Wcs2N1 (magenta) and Hcs2N1(yellow). (O) Lco1N1B (yellow), Lco1N2B (blue), Lco2N1 (green), and Lco2N2(magenta). (P) Lgs1NB (blue), Les2N1 (yellow), and Les2N2 (magenta).Regions of the brain (GNG, SLP, PVLP, AVLP, and WED) are indicated withwhite doted lines. (See movie S3 for detail.) Scale bars in (C) to (P), 50 mm.(Q) Schematic somatosensory pathways. Note that although neuronsappear to terminate in diverse brain regions, their terminals mostly occupya small volume of the brain that span across neighboring regions.



Dow

nloaded from


Consistent with the reported response propertyof the leg cs (14), activity persisted during theperiod of deflection (fig. S6, A and B). Neuronsalso showed strong responses upon large andrapid deflections during flight behavior (Fig. 3,B to D, and fig. S6C).To analyze the physiological somatosensory

representation of the leg co neurons, we surgi-cally removed two of the three legs on one sideof the body and recorded the axons of the re-maining leg co neurons (Fig. 3, E to I, and fig.

S7A). Flies with only forelegs andmidlegs showedspecific responses during walking behavior at thedistal and slightly proximal portion of the axonbundle in the AVLPandWEDregions, respectively(Fig. 3, E, F, and I), whereas those with onlyhindlegs showed responses not in the AVLP andWED regions but only in the proximal part of theaxon bundle in the posterior GNG (Fig. 3, G, H,and I). This finding nicely correlates with theanatomy of axon projections from the co neuronsof each leg that we examined (fig. S4, G to J).

Primary neurons and secondary interneuronsof the leg co converge in the ventral AVLP. Dodifferent neurons convey different aspects ofsomatosensory signals to the brain?We comparedvideo-recorded leg movement of the flies withthe responses of leg co neurons. We classified theflies’ leg movement into seven classes dependingon whether the legs move or not and how theycontact with other objects. When the flies are onthe trackball, they either walk, move their legs togroom the body, or stop. When we lower the


Fig. 3. Calcium imaging of specific primarysensory neurons and secondary interneurons.(A) Schematics of imaging setup (left), photograph ofa fly with a tungsten rod stimulation (yellow arrow) tothe wing (middle), and schematics showing areas ofrecording in the ventral brain (right). Neuronal activitywas measured at the terminal parts of axons. Seemovie S4 and figs. S6 to S8 for examples of actualimaging. (B and C) Responses of wing cs neurons (B1to B3) and haltere cs neurons (C1 to C3). Dottedcircles indicate the areas that show activity differ-ences. + and – symbols indicate the relativeintensity of the interested signals. Scale bars, 10 mm.(D) Responses of wing and haltere cs neuronsduring artificial deflection of wing or haltere andduring tethered flight (n = 3 to 4 animals, 3 to8 data sets in total). Box plot of the averageintensity for 2 s in the region of interest [shownin (B) and (C)], measured from 3 s before and afterthe onset of wing or haltere deflection and flight(where F is the average intensity before the onsetof deflection or flight). **P < 0.01. The datarepresent responses of Wm and Hm sub-bundles;Wl and Hl sub-bundles also showed similarresponses. (E to H) Responses of co neuronsfrom single legs. Other legs were excised shortlybefore assay. Activity in the area of AVLP andWED neuropils (for each leg) and GNG (for hindleg)was measured. Dotted circles indicate the areasthat show activity differences. + and – symbolsindicate the relative intensity of responses. Scalebars, 10 mm. (I) Responses of leg co neuronsderiving from a single leg (n = 3 animals, 9 to19 data sets). Intensity for 2 s just before the fliesstart another behavior was averaged (where Fis the value for the longest “stop” period duringan imaging session). **P < 0.01; dash indicatesnot significant. (J toM) Responses of primaryand secondary neurons during different classesof leg movement behavior (n = 3 to 9 animals/3 to83 data sets). Observed behaviors: walk, groom(groom legs and body parts), stop on thetrackball, touch (touch other objects brieflywhile moving legs in the air), flail (move legsin the air without contact to other objects), andstay (stay silent in the air). Intensity for a 2-s periodjust before the flies start another behavior wasaveraged (where F is the value for the longest“stop” period during an imaging session).**P < 0.01; *P < 0.05; dash indicates notsignificant. P values were determinedby Wilcoxon-Mann–Whitney U-tests forpairwise comparisons and Kruskal-Wallistests for multiple comparisons.



Dow

nloaded from


trackball and keep the flies in the air, they eithergroom their legs or body, move their legs withbrief contact to other objects such as the metalplate of the imaging setup (touch), move theirlegs freely in the air (flail), or stay without mov-ing (stay) (movie S4).Both types of primary co neurons respondmore

strongly when the flies move their legs thanwhen they do not (walk and/or groom versus stopon the trackball; groom, touch, and/or flail versusstay in the air) (Fig. 3, J and K, and fig. S7, B andC). They showweakest activitywhen the legs havenomovement and ground contact (stay in the air).Response levels in other behavioral classes weremore varied in Lco1N1B than in Lco1N2B.Secondary interneuronswhose dendrites over-

lap with the axon terminals of the leg somato-sensory neurons responded to leg movement,suggesting the sensory input. Whereas the pri-mary co neurons showed stronger responses whenthe legs move, a coassociated secondary inter-neuron type Lco2N1 showed stronger responseswhen the legs contact objects continuously. Legmovement with ground contact (walk, groom,and/or stop) or persistent contact with other ob-jects (groom in the air) showed higher activitylevel than during the leg movement with briefor no contact (touch, flail, and/or stay) (Fig. 3L,and fig. S8, A and B). Likewise, the responsewhenthe flies stop on the ball (stop) was stronger thanwhen they stay in the air (stay). Secondary inter-neurons Les2N1D, which are associated with theleg es and gs, showed the inverse pattern: Re-sponses are weaker when the legs contact objectscontinuously (walk, groom, stop on the ball,and groom in the air) than during legmovementwith brief contact (touch) (Fig. 3M and fig. S8, Cand D).

Importance of contact-responsivesensory neurons and interneurons forthe walking response against wind

Our physiological assay revealed varying patternsof leg neuron responses depending on cell types,leg movement, and ground contact. To furtherunderstand the roles of those neurons in behav-ior control, we disturbed their functions to ana-lyze behavioral effects. Flies show a rapid arrestof walking when they encounter strong airflow(Fig. 4A), called wind-induced suppression of loco-motion (WISL), and a specific subset of Johnston’sorgan neurons are responsible for wind detec-tion (39). However, the same speed of airflowstimulus does not cause arrest when the flies arein the air. Thus, to induce WISL the flies shouldintegrate wind stimuli with signals from otherbody parts to determine whether they are on theground or in the air.To identify the neurons involved in ground

detection, we induced expression of an inwardrectifier potassium channel, Kir2.1 (42), to sup-press the functions of specific neurons in the legs(movie S5). Control flies that carry only the Kir2.1transgene or only the GAL4 construct showednormal WISL (Fig. 4B and blue lines in Fig. 4, Cto J). The strain we used for labeling Lco1N1also activates expression in the wind-detecting


Fig. 4. Behavioral analysis of the responses against wind. (A) Schematics of the WISL assay.A group of ~25 flies was placed in a thin transparent chamber at one time, illuminated from belowand video recorded for 240 s from above. Airflow (1.35 m/s) was provided from one side of thechamber for 120 s. Movement of the flies was tracked to calculate walking velocity betweenneighboring video frames (1/30 s). Twelve trials (total: 300 flies) were repeated, and movementof 68 ± 17 flies (mean ± SD) was tracked successfully for each video frame. See movie S5 forexamples of actual video recording. (B to J) WISL analysis. Thin vertical lines represent mean ± SDof velocity for each video frame. (B) Genetic control (carrying UAS-Kir2.1 but no GAL4 to activateit) shows robust WISL. Black, mean; gray, SD. (C to J) Lines represent velocity of genetic control flies(GAL4 only, blue, mean; light blue, SD) and flies with silenced neuronal activity (GAL4 + Kir2.1, red,mean; pink, SD). We were not able to measure the specific contribution of Lco1N1, because all of thedriver lines we found for this neuron type induce expression also in the wind-detecting Johnston’s organneurons. (K) WISL index. The ratio between the mean velocity for all frames of wind-on and wind-offstates was calculated for each trial, and mean ± SEM of the 12 trials was compared. **P < 0.01. P valueswere determined by ANOVA (analysis of variance) with post-hoc Tukey HSD tests.



Dow

nloaded from


co neurons of the antennal Johnston’s organ.Accordingly, suppression of these cells severelydisturbed WISL (Fig. 4C). When we suppressedthe leg es neurons, the basal walking activity wasdecreased even without wind (Fig. 4D), indicat-ing that the function of the leg es neurons iscrucial for normalwalking behavior. Nevertheless,the WISL index, which compares the decreaseof average walking velocity betweenwind-off andwind-on states, showed a considerable decreasewhen the leg es neurons are blocked (Fig. 4K),indicating that the flies with no signals fromthese cells cannot respond properly against windstimuli. WISL was also disturbed when we si-lenced leg sr neurons and distinct subsets of legco neurons (Fig. 4, E, F, and K), suggesting theirroles for ground detection. On the contrary, si-lencing leg gs and cs neurons did not disturbWISL (Fig. 4, G, H, and K).Thus, only es, sr, and co neurons are important

for WISL that occurs only on the ground. The esneurons detect contacts between legs and ground,and the sr neurons are known to be importantfor proprioception and load sensing (25). The roleof Lco1N2 for ground detection matches wellwith its physiological property, which showedstronger responses when flies stop on the groundversus in the air (Fig. 3K). Suppression of es- andgs-associated and co-associated interneuronsLes2N1 and Lco2N1, respectively, also disturbedWISL (Fig. 4, I, J, and K). This is reasonable con-sidering that these interneurons showed differ-ent activity levels, depending on ground contact(Fig. 3, L and M).Suppression of any of the three leg primary

neuron types (es, sr, and co) and certain inter-neurons alone was sufficient to disturb WISL,but none of them induced complete suppression.These suggest that integration of the informationfrom all those neuron types would be requiredwhen the flies determinewhether they are on theground or in the air.

Discussion

We performed systematic identification of theprimary and secondary somatosensory neuronsofDrosophila and provided a systematic overviewof the somatosensoryneuronal circuitry of insectsfromperiphery tobrain (Fig. 5A). A set of expressiondriver strains will open up a new research fieldfor detailed functional analysis and selective ma-nipulation of the somatosensory system. Identifi-cation of clearly layered terminals in all the sensoryneuropils of the VNC should provide the basisto identify and analyze further secondary inter-neurons arising from eachmodality-specific layer.

Architecture of the Drosophilasomatosensory system

Only three distinct types of sensory informationare transmitted directly to the brain by primaryneurons (leg gs, co, and wing and haltere cs).Such connection has also been reported in otherinsects (15, 16), suggesting that this might be ageneral feature across insecta. Whereas only asmall portion of leg co neurons project directlyto the brain, most wing and haltere cs neurons

innervate the brain; these cs neurons are knownto detect various aspects ofwing-beat force duringflight to provide feedback control (43). Direct pro-jections to the brainwould be important for theseneurons to enable fast transmission of informa-tion about rapidly changing sensory parametersduring flight.We found that ground detection for WISL,

which would require slower temporal resolutionthan flight control, is mediated by both directand indirect pathways (Fig. 4). Primary neuronsand secondary interneurons of the same sensorymodality tend to converge in specific subregionsof the brain, forming modality-specific somato-sensory representation. In spite of the similar axontrajectory in the brain (Fig. 2O), these neuronsconvey information about leg movement in dif-ferent ways (Fig. 3, J to L).Interneurons associated with the leg co and

es terminate in neighboring but different regionsof the lateral brain (Fig. 2, O and P), yet some ofthem have shared roles inWISL control. Becausetheir signals are transmitted to distinct parts ofthe brain, yet-unidentified higher-order neuronsin the brain should converge those signals to themotor control circuitry.In this respect, it is important to note thatmost

ascending secondary interneurons we identifiedhave presynaptic output sites, not only in the brain

but also in theVNC (fig. S5, A2 to F2 andA4 to F4).Local circuitry in the leg neuropil is importantfor controlling leg movement (11, 14, 44). Thoselocal neurons are likely candidates that receiveoutput from the ascending interneurons, becauseaxon terminals of sensory neurons hardly havepostsynaptic sites. Similar local output has alsobeen found in other sensory modalities; manyolfactory and visual projection interneurons havecollateral output synapses in the antennal lobeand optic lobes (45, 46).There are three pairs of leg neuropils. Among

them, the foreleg neuropil has specialized arbo-rization of the gs neurons that exist only in theforeleg (23). Other than this, we did not find sub-stantial differences of arborization patterns be-tween the fore-, mid-, and hindleg neuropils.

Homology between insect andmammalian somatosensory systems

The present results provide data for a systematiccomparison of the insect somatosensory systemwith its mammalian counterparts. Insects andmammals share similarities of neural organizationunderlying the perception of odors, taste, vision,sound, and gravity (2–5), and our data also revealmarked similarity for the mechanosensory sysem(Fig. 5, A and B). In insects, some primary neuronsproject directly to distinct parts of the ventral and


Fig. 5. Schematic comparison of somatosensory pathways in flies and mammals. (A) Pathwaysin Drosophila identified in this study. (B) Known pathways in mammals. (C) Possible correspondencebetween different types of somatosensory neurons, showing inverse dorsoventral order. Note thatsome fish and amphibians also have putative gustatory neurons on the body, but their projectiontargets are not well known.



Dow

nloaded from


lateral brain, whereas others terminate within theVNC. Likewise, in mammals, some neurons proj-ect directly to the ventral brain at the medullaoblongata, whereas others terminate within thespinal cord (6, 47, 48).Modality-specific pathwaystend to converge in different subregions of themedulla, as well as in the thalamus of the mam-malian brain (6, 49). Similarly, direct and indirectpathways tend to converge in common subregionsof the insect brain, and neurons conveying infor-mation about different somatosensory modal-ities tend to terminate in different subregions.As in mammals, these subregions often lie ad-jacent to each other in certain parts of the brain;for example, the entire terminal arborizations ofthe leg co and es secondary interneurons are con-fined in a 40-mm-wide, 150-mm-tall cylindrical vol-ume in the lateral brain (Fig. 2, O and P, and fig.S5, N and O).Somatosensory signals are sent predominantly

to the ipsilateral brain side in insects and con-tralateral in mammals. Considering that descend-ing neurons tend to project ipsilaterally in insectsbut contralaterally inmammals (50–52), however,somatosensory signals and motor control compu-tation are processed primarily in the same side ofthe brain in both cases.Layers of sensory axon terminals in the insect

VNC and mammalian spinal cord (9, 10, 13, 53)are also organized in a similar order (Fig. 5C).Insect multidendritic neurons and mammalianfree nerve endings share various characteristicsin common: Their dendrites both have free end-ings without forming particular sense organs todetect pain, temperature, and other submodal-ities (54, 55 ). Themd neurons project to themostventral layer of the VNC, whereas free nerve end-ings innervate the most dorsal layer of the spinalcord. Axons from the insect external sensilla andmammalian hair receptors, both of which detecthaptic contact to the tips of the bristles and hairs,terminate in the second-ventral and second-dorsallayers, respectively. Insect chordotonal organ andmammalianmuscle spindle, as well as insect cam-paniform sensilla and mammalian Golgi tendonorgan, also show similarity with respect to theirfunctions in motor control (12). These receptorsystems supply afferents to themost dorsal andmost ventral layers in insects and mammals, re-spectively. A fly’s stretch receptors and mamma-lian Merkel cell neurites—as well as Meissner,Ruffini, and Pacinian corpuscles—terminate inthe third-ventral and third-dorsal layer, respec-tively. Although correspondence between themis less obvious, they similarly detect deforma-tion of the exoskeleton and skin (1, 6, 48). Thus,functionally comparable somatosensory termi-nals are layered in reverse order between the twosystems (Fig. 5C). Considering that the dorso-ventral axis of the mammalian body is develop-mentally upside down compared with the insectone (56, 57 ), the corresponding order of sensoryarrangements is actually conserved exactly be-tween the two systems.Do corresponding somatosensory cell types

express common genes? Modality-dependent mo-lecular specialization is not apparent evenwithin

insects or mammals, because the same genes areoften expressed in multiple cell types and only afew genes share expression in the correspondingcell types across taxa (table S1). This might be arather general feature; receptor molecules as wellas developmental origins of the sensory organsare not identical between insects and mammalsalso in olfactory and auditory systems, yet sen-sory centers in the brain share architectural sim-ilarities (4, 5 ).With our somatosensory analysis, transphy-

letic correspondence of neuronal circuitry hasbeen found in all of the sensory modalities. Cor-responding organization has been suggestedalso for associative centers and motor systems(2–5, 9, 58–60). The fact that essentially all im-portant components of the brain system shareconserved features across the two evolutionaryclades, which have been separated since at leastthe end of the Ediacaran period more than550 million years ago, would suggest that basicdevelopment programs for the orderly and se-crete segregation of those circuitsmay have evolvedbefore deuterostome-protostome or deuterostomia-ecdysozoa divergence.

REFERENCES AND NOTES

1. P. Delmas, J. Hao, L. Rodat-Despoix, Nat. Rev. Neurosci. 12,139–153 (2011).

2. J. S. Joly, G. Recher, A. Brombin, K. Ngo, V. Hartenstein,Curr. Biol. 26, R1001–R1009 (2016).

3. D. A. Clark, J. B. Demb, Curr. Biol. 26, R1062–R1072 (2016).4. J. G. Hildebrand, G. M. Shepherd, Annu. Rev. Neurosci. 20,

595–631 (1997).5. A. Kamikouchi et al., Nature 458, 165–171 (2009).6. V. E. Abraira, D. D. Ginty, Neuron 79, 618–639 (2013).7. A. I. Basbaum, D. M. Bautista, G. Scherrer, D. Julius, Cell 139,

267–284 (2009).8. S. A. Smith, D. Shepherd, J. Comp. Neurol. 364, 311–323 (1996).9. N. J. Strausfeld, Arthropod Brains: Evolution, Functional

Elegance, and Historical Significance (Belknap Press, 2012).10. D. J. Merritt, R. K. Murphey, J. Comp. Neurol. 322, 16–34 (1992).11. E. M. Berg, S. L. Hooper, J. Schmidt, A. Büschges, Curr. Biol.

25, 2012–2017 (2015).12. T. Buschmann, A. Ewald, A. von Twickel, A. Büschges,

Bioinspir. Biomim. 10, 041001 (2015).13. J. C. Tuthill, R. I. Wilson, Curr. Biol. 26, R1022–R1038 (2016).14. J. C. Tuthill, R. I. Wilson, Cell 164, 1046–1059 (2016).15. P. Bräunig, K. Krumpholz, Front. Zool. 10, 54 (2013).16. N. Ando, H. Wang, K. Shirai, K. Kiguchi, R. Kanzaki, J. Insect

Physiol. 57, 1518–1536 (2011).17. L. A. Weiss, A. Dahanukar, J. Y. Kwon, D. Banerjee, J. R. Carlson,

Neuron 69, 258–272 (2011).18. J. A. Ainsley et al., Curr. Biol. 13, 1557–1563 (2003).19. S. Hayashi et al., Genesis 34, 58–61 (2002).20. T. Miyazaki, K. Ito, J. Comp. Neurol. 518, 4147–4181 (2010).21. A. Jenett et al., Cell Rep. 2, 991–1001 (2012).22. F. Ling, A. Dahanukar, L. A. Weiss, J. Y. Kwon, J. R. Carlson,

J. Neurosci. 34, 7148–7164 (2014).23. J. Y. Kwon, A. Dahanukar, L. A. Weiss, J. R. Carlson, J. Biosci.

39, 565–574 (2014).24. R. K. Murphey, D. R. Possidente, P. Vandervorst, A. Ghysen,

J. Neurosci. 9, 3209–3217 (1989).25. B. S. Desai, A. Chadha, B. Cook, Science 343, 1256–1259

(2014).26. L. H. Field, T. Matheson, in Advances in Insect Physiology,

P. D. Evans, Ed. (Academic Press, 1998), vol. 27, pp. 1–228.27. S. R. Shanbhag, K. Singh, R. N. Singh, Int. J. Insect Morphol.

Embryol. 21, 311–322 (1992).28. U. Grünert, W. Gnatzy, Zoomorphology 106, 320–328 (1987).29. S. N. Zill, J. Schmitz, S. Chaudhry, A. Büschges,

J. Neurophysiol. 108, 1453–1472 (2012).30. K. Shimono et al., Neural Dev. 4, 37 (2009).31. Y. Kwon, W. L. Shen, H. S. Shim, C. Montell, J. Neurosci. 30,

10465–10471 (2010).32. A. Ghysen, Dev. Biol. 78, 521–541 (1980).

33. H. Amrein, N. Thorne, Curr. Biol. 15, R673–R684 (2005).34. E. S. Cole, J. Palka, J. Embryol. Exp. Morphol. 71, 41–61 (1982).35. N. C. Inestrosa, C. Sunkel, J. Arriagada, Brain Res. 416,

248–256 (1987).36. D. W. Williams, D. Shepherd, J. Neurobiol. 39, 275–286 (1999).37. A. P. Jarman, Y. Grau, L. Y. Jan, Y. N. Jan, Cell 73, 1307–1321

(1993).38. K. Ito et al., Neuron 81, 755–765 (2014).39. S. Yorozu et al., Nature 458, 201–205 (2009).40. H. Kim, C. Kirkhart, K. Scott, eLife 6, e23386 (2017).41. T. W. Chen et al., Nature 499, 295–300 (2013).42. R. A. Baines, J. P. Uhler, A. Thompson, S. T. Sweeney, M. Bate,

J. Neurosci. 21, 1523–1531 (2001).43. B. H. Dickerson, Z. N. Aldworth, T. L. Daniel, J. Exp. Biol. 217,

2301–2308 (2014).44. M. Burrows, The Neurobiology of an Insect Brain (Oxford Univ.

Press, 1996).45. H. Otsuna, K. Ito, J. Comp. Neurol. 497, 928–958 (2006).46. R. Okada, T. Awasaki, K. Ito, J. Comp. Neurol. 514, 74–91

(2009).47. T. Caspary, K. V. Anderson, Nat. Rev. Neurosci. 4, 289–297

(2003).48. L. Li et al., Cell 147, 1615–1627 (2011).49. K. Sakurai et al., Cell Rep. 5, 87–98 (2013).50. A. I. Basbaum, H. L. Fields, J. Comp. Neurol. 187, 513–531

(1979).51. L. Mu, J. P. Bacon, K. Ito, N. J. Strausfeld, J. Exp. Biol. 217,

2121–2129 (2014).52. M. S. Sivertsen, M. C. Perreault, J. C. Glover, J. Comp. Neurol.

524, 1270–1291 (2016).53. P. D. Wall, J. Physiol. 188, 403–423 (1967).54. K. Y. Kwan, J. M. Glazer, D. P. Corey, F. L. Rice, C. L. Stucky,

J. Neurosci. 29, 4808–4819 (2009).55. S. M. Brierley et al., J. Physiol. 589, 3575–3593 (2011).56. E. Geoffroy Saint-Hilaire, Mém. Mus. Hist. Nat. 9, 89–119 (1822).57. E. M. De Robertis, Y. Sasai, Nature 380, 37–40 (1996).58. N. J. Strausfeld, F. Hirth, Science 340, 157–161 (2013).59. K. G. Pearson, Annu. Rev. Neurosci. 16, 265–297 (1993).60. S. L. Hooper, A. Büschges, Neurobiology of Motor Control:

Fundamental Concepts and New Directions (Wiley-Blackwell,2017).

ACKNOWLEDGMENTS

We are grateful to W. A. Johnson for ppk-GAL4, K. Emoto forppk-LexA, D. van Meyel for tutl-GAL4, C. Montell for inactive-GAL4,J. Carlson and T. Tanimura for Gr5a-GAL4 and Gr89a-GAL4,B. Pfeiffer for LexAop2-mCD8::GFP, J. Urban and G. Technau forMZ-series enhancer trap strains, the members of the Nippon GAL4Consortium and D. Yamamoto for the NP-series strains, G. Rubinand Y. Aso for GMR-GAL4 driver strains, Bloomington Stock Centrefor other transgenic strains, and the Developmental StudiesHybridoma Bank for the nc82 antibody. We thank A. Büschges,A. Kamikouchi, S. Kohatsu, S. Namiki, T. Shimogori, D. Yamada,D. Yamamoto, and M. Yoshihara for discussion and sharingtechnical information; M. Ito and Janelia FlyLight project forsharing neuron image data; T. Kawase for visualizing 3D data;K. Branson and J. Bender for behavior analysis; and S. Shuto,K. Yamashita, H. Hirose, Y. Ishida, and R. Tatsumi for technicalassistance. 3D data of the identified neurons are availablevia the Virtual Fly Brain database (www.virtualflybrain.org/), andthe raw movie data of calcium imaging and behavioral assay arearchived on our laboratory data server (http://jfly.uni-koeln.de/data/science2017/). This work was supported by grants from theCore Research for Evolutional Science and Technology (JapanScience and Technology Agency and Japan Agency for MedicalResearch and Development) program and from the StrategicResearch Program for Brain Sciences to K.I.; a Grant-in-Aidfor Scientific Research from the Ministry of Education, Culture,Sports, Science and Technology of Japan to K.I. and A.T.;and a Japan Society for the Promotion of Science fellowship toT.Y. and T.K.Y.

SUPPLEMENTARY MATERIALS

www.sciencemag.org/content/358/6363/615/suppl/DC1Materials and MethodsFigs. S1 to S8Table S1References (61–121)Movies S1 to S5

13 April 2017; accepted 25 September 201710.1126/science.aan4428




Dow

nloaded from

http://www.virtualflybrain.org/

http://jfly.uni-koeln.de/data/science2017/

http://jfly.uni-koeln.de/data/science2017/

http://www.sciencemag.org/content/358/6363/615/suppl/DC1


brainTopological and modality-specific representation of somatosensory information in the fly

Asako Tsubouchi, Tomoko Yano, Takeshi K. Yokoyama, Chloé Murtin, Hideo Otsuna and Kei Ito

DOI: 10.1126/science.aan4428 (6363), 615-623.358Science

, this issue p. 615Sciencecontrol of upwind behavior that occur only when the flies are on the ground.nerve cord and brain. The authors dissected preferential responses to wing and leg movement and contributions to thefruit flies. The findings revealed topological and modality-specific mechanosensory representations in the insect ventral

systematically mapped the somatosensory circuits inet al.modalities, even between insects and mammals. Tsubouchi The organization of sense organs and sensory brain centers shows conserved principles for many sensory

A somatosensory map in the fly brain

ARTICLE TOOLS http://science.sciencemag.org/content/358/6363/615

MATERIALSSUPPLEMENTARY http://science.sciencemag.org/content/suppl/2017/11/02/358.6363.615.DC1

REFERENCES

http://science.sciencemag.org/content/358/6363/615#BIBLThis article cites 117 articles, 23 of which you can access for free

PERMISSIONS http://www.sciencemag.org/help/reprints-and-permissions

Terms of ServiceUse of this article is subject to the

is a registered trademark of AAAS.ScienceScience, 1200 New York Avenue NW, Washington, DC 20005. The title (print ISSN 0036-8075; online ISSN 1095-9203) is published by the American Association for the Advancement ofScience

Science. No claim to original U.S. Government WorksCopyright © 2017 The Authors, some rights reserved; exclusive licensee American Association for the Advancement of

on January 9, 2021


Dow

nloaded from

http://science.sciencemag.org/content/358/6363/615

http://science.sciencemag.org/content/suppl/2017/11/02/358.6363.615.DC1

http://science.sciencemag.org/content/358/6363/615#BIBL

http://www.sciencemag.org/help/reprints-and-permissions

http://www.sciencemag.org/about/terms-service


topological and modality-specific representation of … · drosophila has six major types of...

Documents