Behavioral modification in choice process of Drosophila Behavioral modification in choice process of Drosophila

Behavioral modification in choice process of Drosophila

  • 期刊名字:中国科学C辑
  • 文件大小:712kb
  • 论文作者:王顺鹏,唐世明,李岩,郭爱克
  • 作者单位:Laboratory of Visual Information Processing,Institute of Neuroscience
  • 更新时间:2020-11-10
  • 下载次数:
论文简介

Vol. 46 No.4SCIENCE IN CHINA (Series C)August 2003Behavioral modification in choice process of DrosophilaWANG Shunpeng (王顺鹏)', TANG Shiming (唐世明)',LI Yan (李岩)'& GUO Aike (郭爱克)1.21. Laboratory of Visual Information Processing, Center for Brain and Cognitive Sciences, Institute of Biophysics,Chinese Academy of Sciences, Beiing 100101, China;2. Institute of Neuroscience, Shanghai Institutes for Biological Sciences, Chinese Academy of Sciences, Shanghai200031, ChinaCorrespondence should be addressed to Guo Aike (email: akguo@ion.ac.cn)Received September 16, 2002Abstract In visual operant conditioning of Drosophila at the flight simulator, only motor output offlies一- yaw torque-- -is recorded, which is involved in the conditioning process. The currentstudy used a newly-designed data analysis method to study the torque distribution of Drosophila.Modification of torque distribution represents the effects of operant conditioning on flies' behavioralmode. Earlier worksl10] showed that, when facing contradictory visual cues, flies could make choicesbased upon the relative weightiness of different cues, and it was demonstrated that mushroombodies might play an important role in such choice behavior. The new“torque-position map' methodwas used to explore the CS-US associative learning and choice behavior in Drosophila from theaspect of its behavioral mode. Finally, this work also discussed various possible neural basesinvolved in visual associative learning, choice processing and modification processing of thebehavioral mode in the visual operant conditioning of Drosophila.Keywords: visual operant conditioning, choice behavior, torque distribution, mushroom bodies.DOI: 10.1360/0 2yc0047Because of its clear genetic and developmental background, diversity of behavioralparadigms and neuroanatomy of the brain, Drosophila has become an important animal model forstudying genetic, molecular and cellular bases of learning and memoryl. Extensive research hasexplored the visual operant conditioning of Drosophila and related molecular bases!2- -8]; recently,researchers began to address cognition-like functions and involved neural substrates91. In thesestudies, behavioral analysis focused on flies' ability to acquire or store the association betweenconditioned stimulus (CS) and unconditioned stimulus (US), while the properties of flies'behavioral output remained less understood. In visual operant conditioned behavior of Drosophila,the motor output (yaw torque) is directly involved in the feedback of conditioning (fig. 1(b)). Thus,to explore the yaw torque flies produce in conditioning may be useful for understanding the visualoperant conditioning of Drosophila.中国煤化工In visual operant conditioning, shape and color C.MYHCNMHGassinglevisualcues or jointly used as combined visual cues to serve as conditioning stimulus. When double cueswere coincident in training and became competitive in test, flies showed decision-like choice400SCIENCE IN CHINA (Series C)Vol.46behavior facing contradictory visual cues. Wild-type flies definitely chose flight pathwaysaccording to the cue that had relative higher weightiness 10); however, mutant flies with miniaturemushroom bodies did not show such performance. The present study used a torque analysismethod to examine torque distributions and behavioral modifications of Drosophila and tounderstand its function in choice behavior.1 Materials and methods1.1 Animal preparationDrosophila melanogaster of wild-type strain WTB (wild-type Berlin) and mutant mbm'(mushroom body miniature) were used. mbm' flies have the same genetic background as WTBflies, and female mbm! flies have almost no mushroom bodies (MBs) structurel2l. Flies weremaintained on standard corm meal/molasses medium'3] at 25"C and 60% humidity with a 14 h/10h light/dark photoperiod. Three- to four- day-old female flies were prepared according to standardprocedures as described in earlier works [21.1.2 Flight simulatorThe flight simulator (fig. 1) was a setup with negative feedback to simulate circumstance forflies' stable flight behaviorl13). A torque meter was the core device of the flight simulator, onwhich individual fly was attached by glue at the head and thorax. The torque meter together withindividual flies was placed in the center of a cylindrical panorama which could turn around itsvertical axis. During stable flight, the fly had the solitary“behavioral module”to choose its .horizontal orientation, and the only recorded component of such flight behavior was the yawControl signal for punishment(a(bYaw torqueTorqueComputerLight bulbmeterSUS heatvisual patternIR-filterelectronicshutterCS-US associationDrosophilabehavioral mode中国煤化工Control signal forMHCNMHGpatterm positionFig.1. (a) Flight simulator setup as used in experiments; (b) basic scheme of the visual operant conditioning paradigm. The“yawtorque”in the elipse is the only motor output of Drosophila that is involved in the feedback in conditioning.No.4BEHAVIORAL MODIFICATION IN CHOICE PROCESS OF DROSOPHILA40torque around the vertical axis. Two T patterns and two⊥patterns on the screen were equallysized and located at the center of each quadrant, and the same patterns were located in oppositequadrants. White light iluminated the panorama from behind. Monochromatic filters can be addedon the screen to adjust the colors of the patterns (blue/green) and intensity of colors (CI: colorintensity). The fly' s yaw torque was continuously measured by the torque meter. Through a .feedback circuit the fly maneuvers the turning movement of the panorama with its yaw torque(coupling coefficient: K=-11° (s* 10-10 Nm)厂). Actually, the fly could not turm around to changeits orientation during stable flight, but it could change the angular position of the panorama withits torque output, and so it could choose its flight direction referring to visual patterns on thescreen. An infrared light beam was180 7Angular porilion Yaw torque是focused on the fly' s rear from above,which functioned as a negative45 2reinforcer in conditioning and could be-45intercepted by a computer-controlled-90--90员shutter. The computer continuously-180recorded on-line the torque of the fly104.and the angular position of thepanorama at a sampling frequency of 20 Fig. 2. Experimental traces from a 45-s experimental course. TheHz. Fig. 2 shows the experimentalangular position varies between -180° and 180° and the yaw torquevariable fluctuates around zero.traces for recorded yaw torque andangular position, with the shaded area showing where the fly was heated by the infrared lightbeam.1.3 Experimental procedureThe experimental procedure for the choice behavior paradigm was divided into successiveperiods (PTE, TR1, TR2, TE1 and TE2) (table 1). All periods were 4 min long with exception ofTE1, which was 2 min. Before each trial, two kinds of visual patterns were matched usingdifferent shapes(T and ⊥) and diff- erent colors (black, blue and green). For example, blue⊥and green T represented two kinds of patterns. From these, one kind of pattern was set as theconditioned stimulus (CS+) negatively reinforced by heat punishment. In the pre-test period (PTE)the preference of the fly towards the two kinds of patterns was tested; in training periods (TR1 andTR2) the fly was trained to avoid the CS+ patterns; in the memory-test period (TE1) the memoryretention of the training was checked. After TE1 the matching of shape and color cues wasreversed and the fly was tested to make a choice between contradictorv visual cues in the choice-中国煤化工、test period (TE2). During training periods, heat was:ver the fly washeading into a quadrant containing the CS+ patterns.YHCN MH Gny was able toform the association between visual patterns and heat punishment, and the fly learned todiscriminate different patterns according to shape or color cues and to choose the non-heated402SCIENCE IN CHINA (Series C)Vol.46pattern to decide flight direction. During the pre-test period, memory-test and choice-test periods,the heat was switched off permanently.Table 1 Experimental procedure to examine choice behavior in DrosophilaPre-testTrainingMemory-testChoice- testPTETR1TR2TE1TB2pattern 1 (CS+): color 1 + shape 1pattern A: color 1 + shape 2pattern 2 (CS ): color 2 + shape 2pattern B: color 2 + shape 11.4 Data analysis1.4.1 Performance index (PI).In earlier works a performance index (PI) was introduced toestimate the fly' s learning and memory abilities[2]. The performance index is also calledpreference index in PTE, avoidance index in TR1 and TR2, memory index in TE1 and choice .index in TE2. Four quadrants of the panorama were divided into two groups: one group was theCS+ domains with the punished visual patterns, and the other group was the CS- domains. Thetime a fly spent in each group was calculated from recorded data sequence of the angular positionof the panorama; the times in the CS- and CS+ domains were recorded as t and t2 respectively.The calculation for the performance index was PI =(1- t2)( 1+ 2);-1≤PI≤1. PI=1 meant thefly completely avoided the heat punishment; PI =-1 meant the fly was staying in the punisheddomains all the time; and PI = 0 meant the fly stayed in the CS+ and CS - domains equally.1.4.2 Torque-position map.The calculation of PI only depended on the variable of theangular position of the panorama, which did not take into account the fly' s yaw torque. Obviously,the PI evaluation method is insufficient to explore the dynamics of the conditioning process.Therefore, a new data analysis method was needed to enhance understanding of the fly' s operantconditioned behavior, which would take into consideration both the angular position and yawtorque.The newly designed torque-position map method was used to explore the torque fliesproduced at each angular position during conditioning. The polarity of the yaw torque is related tothe turning direction of flies in the flight simulator (positive torques represent clockwise turningactions and vice versa). Analysis revealed no significant correlation between the torque polarity(i.e., the turning direction) and the angular position. So the torque-position map method was onlyused to study the relationship between the torque amplitude and the angular position. The analysisincluded a transformation of the angular position variable. The true angular position variablechanges in the range of -180° to 180°. Torque-position map analysis used angular distancesbetween the fly' s orientation and the visual patterns to represent the fly' s relative angular position(fig. 3(a)). Thus, when the fly was heading towards CS-中国煤化工he fly was set as-45°; when heading towards the CS- patterns, the po::[YHC N M H Gheading towardsthe boundary between the CS+ and CS - domains, the position was 0° . The newly defined angularNo.4BEHAVIORAL MODIFICATION IN CHOICE PROCESS OF DROSOPHILA403position enabled the torque-position map method to statistically delineate changes of torquedistribution of flies in conditioning. After the angular position transformation, the 360 -degreepanorama was divided into four equal domains (all ranging from -45° to 45° ). The four domainswere superimposed, and then the torque amplitude was averaged at each angular position andplotted on the torque-position map (fig. 3(b).CS+-10CS-4S°是0%/)°嘶)→-45t1- 45°工0以工/°45°- 45 -30 -151530 45CS-: LCS+: TAngular position/(°)(a(b)Fig. 3. (a) The 360-degree panorama was divided into four 90 degree domains (ranging from -45”to 45"). One of the twopatterns(T or⊥) was set to be punished by heat for each trial, two domains with the punished patterns were CS+ domains andthe others were CS- domains. (b) An example for the torque-position map and domain specific torque distributions. The dotsrepresent the average torque amplitudes flies produced at each angular position. The horizontal“solid lines士dashed lines" showthe average amplitude levels (Mean士S.D.) for the CS+ and CS- domains. Data shown are from one WTB fly during the ex-perimental period of TR1.1.4.3 Domain specific torque distribution.In addition to the analysis of torque production ateach angular position, torque distributions in the CS+ and CS- domains were also studied andquantitatively compared. In conditioning, the reinforcer (heat) was presented throughout the CS+domain and canceled throughout the CS- domain, and it did not depend on specific angularpositions in the CS+ domain. Thus, the torque distributions in both CS+ and CS - domains couldbe computed, and the ANOV A-test method was used to compare both torque distributions. Torquedistribution of the fly was regarded as symmetric between CS+ and CS- domains when ANOVA-test results had no significant difference, otherwise the torque distribution was asymmetric. Asshown in fig. 3(b), the horizontal solid lines indicate the average torque amplitude levels in eachdomain and the dashed lines represent the deviations (S.D.) of torque distribution in each domain.2 Results and analysis2.1 Torque distributions of Drosophila facing a single visual cueThe performance indices and torque distributions of WTB flies (n= 12) were studied inconditioned learning with only shape cue. The two中国煤化工for conditionedstimulus in learning paradigm were different only in as.MYHCNMHGad上on whitebackground). Because only the shape cue was used in conditioning, the memory retention wastested in TE2, which was different from that in choice paradigm as described in sec. 1.3. In the404SCIENCE IN CHINA (Series C)Vol.46training period, flies were trained to choose their flight direction in response to the shapes. Beforeeach trial T or ⊥patterns were set as punished objects (CS+). PIs (fig. 4(a)) revealed that fliesexposed to training period for such learning tasks chose their orientation toward the CS- patternsand avoided the CS+ patterns (positive PImI, PIT2), and such preference was still detectable in10[(a)从(b)CS+CS-合0.5-45-30 -15030 45壬⊥Angular positin/)PTETEITE2() CS+()是-45 -30 -1515304s-45-30 -131330 49c)cs+ 6-30 -i545Angular position(C)Fig. 4. The performance indices and torque distributions of WTB flies (n=12) during conditioning with the pattern shape cue.(a) Performance indices flies gained during each experimental period. During the pre-training period (PTE), flies had no prefer-ence for different pattern shapes; during training periods (TR1 and TR2), flies learned to avoid the CS+ patterns and inclined tothe CS- patterns and the tendency persisted during the memory test periods (TE1 and TE2). (b) In the PTE period, WTB fliesproduced approximately equal torque amplitudes at all angular position中国煤化Itween CS+ and CS-domains were symmetrical. (C) and (d) In the TR1 and TR2 periods, Wher torque amplitudesin the CS+ domain than in the CS- domain, which persisted during TE1fYHCN M H G12 hies. hies werepunished for the T shape and the others were punished for the⊥shape.No.4BEHAVIORAL MODIFICATION IN CHOICE PROCESS OF DROSOPHILA405memory-test periods after the training (positive PITEI, PImE2). At the same time, results fromtorque-position map analysis showed that torque distributions in the CS+ and CS- domainsseemed to be symmetric before training and flies produced approximately equal torque amplitudesat each angular position (fig. 4(b)). In contrast, torque distributions came to be asymmetric intraining periods and the torque level in the CS+ domain was measurably higher than in the CS-domain (fig. 4(c) and (d)). Finally, the asymmetric torque distributions between CS+ andCS- - domains persisted in the memory-test periods after training (fig. 4(e) and (f).Qualitative descriptions of torque-position map analysis revealed the first data on themodification of flies' torque distributions in conditioning, which indicated the flies' process ofbehavioral modification caused by the operant conditioning. Additional quantitative analysis wasconducted for torque distributions of WTB flies for the same learming task, in which the ANOVA-test was used to determine the difference of torque distributions between CS+ and CS - domains(table 2). In PTE, there was no significant difference of torque distributions between CS+ and CS-domains (average士S.D. for CS+ and CS- domains were 3.29土0.19, 3.36士0.22 (10-10 Nm)respectively; F1, 881= 2.75, P= 0.10). In the training periods (TR1 and TR2), results from theANOVA-test showed a significant difference of torque distributions between CS+ and CS-domains (for TR1: F1. 881= 187.15, P < 0.001; and for TR2: F. 88=143.85, P < 0.001); the .averaged torque amplitude level in the CS+ domain was higher than in the CS- domain. Inmemory-test periods after training (TE1 and TE2), a significant difference in torque distributionsbetween CS+ and CS - domains still existed (for TE1: F1. 881= 88.01, P < 0.001; and for TE2:Fr.88]= 107.03, P < 0.001).Table 2 Torque distibutions (average士S.D) (x10-"Nm) of WTB flies (n=12) in experiments with only shape cueANOVA-test resultsExperimental periodCS+ domainCS- domainFu.881PPTE3.29 +0.193.36 +0.222.750.10TR13.92 +0.382.79士0.25187.15<0.001*TR23.66 +0.322.56士0.28143.85<0.001**TE13.21 +0.392.58士0.2388.01TE3.11 +0.182.43 +0.26107.03<0.001* ”“** indicates level of significance P <0.001.The torque-position map analysis was used to reveal changes of torque distributions fromsymmetric status to asymmetric status in operant conditioning, and the ANOVA-test was used toquantitatively detect the significance of such modification. Changes in flies' torque distributionsactually represented the behavioral modifications o中国煤化工perant training.Similar experiments also revealed such behavioral modiMYHCNMHGiliesinthesamelearning tasks with only shape cue.406SCIENCE IN CHINA (Series C)Vol.462.2 Torque distributions of flies facing contradictory visual cuesRecently, Tang and Guo10] developed the choice paradigm to study the choice behavior ofDrosophila facing contradictory visual cues: two kinds of visual patterns used as conditionedstimulus had different shapes(T or ⊥) and colors (blue or green), and before each trial one kindwas set as the punished object (e.g.. blue⊥patterns were punished by heat, while green Tpatterns were not). In the memory-test period (TE1), flies' ability to remember the pattern-heatassociation was tested; but in choice-test period (TE2), the matching of shape and color cues wasreversed, which made the two visual cues become competitive (e.g., blue T and green⊥patterns). Tang and Guol10] found that the choice behavior of flies depended on the relativesalience of shape and color cues, and demonstrated that mushroom bodies might be involved insuch processing. In their work they defined an index to represent the intensity of color (CI), whenthe CI value was changed the relative weightiness of color vs. shape was altered: at CI = 1.0 thecolor had greatest intensity and weightiness of color was greater than that of shape; at CI = 0 thecolor had the least intensity and weightiness of color was less than that of shape. Given differentCI values, flies showed distinct choice behaviors. The present work used the torque analysismethod and the ANOV A-test to carry out behavioral studies on torque distributions of WTB andmbm' flies in choice processing.Heat punishment was assigned to blue⊥patterns in periods of PTE, TR1, TR2 and TE1 inchoice experiments, WTB and mbm! flies were trained to prefer green T patterns and avoid bluepatterns. According to PIs, the conditioned avoidance behavior resembled those shown inlearning tasks with only shape cue, and did not depend on CI values. Torque distribution analysesshowed that torque distributions in such experimental periods were similar to those produced inrelevant periods described in sec. 2.1: in the PTE period, the averaged torque amplitude levels inthe domain with a blue⊥patterm and the domain with a green T pattern were similar, andtorque distributions between the two domains seemed to be symmetric; in periods of TR1, TR2and TE1, torque distributions between the blue⊥and green T domains were asymmetric, andthe averaged torque amplitude level in the punished blue⊥domain was significantly higher thanin the non-punished green T domain. After TE1, the matching of shape and color cues wasreversed, and novel blue T and green⊥patterns were presented. In the choice-test period(TE2) following the reversing of shape- color matching, flies had to choose which pattern to flytoward and which to avoid, avoiding blue T patterns if they made the choice in response to color,or avoiding green ⊥patterns if they made the cho中国煤化工The fllowingCNMH G,sections presented results of torque distribution studiespical CI values.CI= 1.0, CI=0.7 and CI=0.8 represented choices on bases of shape, color and disability in choice-No.4BEHAVIORAL MODIFICATION IN CHOICE PROCESS OF DROSOPHILA407making.2.2.1Torque distributions of flies at CI = 1.0.The first torque analysis for choiceexperiments was conducted using WTB (n =9) and mbml (n=9) flies at CI= 1.0 (table 3). WTBflies showed positive performance indices in both TE1 and TE2 (positive PITE1, PITe2), whichindicated that WTB flies made choices in TE2 according to color cues at CI= 1.0. Torque analysisresults showed that the averaged torque amplitude level of WTB flies in the blue T domain wassignificantly higher than in the green⊥domain for both TE1 and TE2 (for TE1: Fr.881= 68.01, P< 0.001; TE2: Fr.881= 30.79, P < 0.01). Torque analysis also revealed WTB flies' choice behaviorin response to color cues. Thus, flies’torque distribution modifications had common propertieswith PIs in such choice processing: they all connected to the color cue that was used for choices.AtCI= 1.0 mbm' flies also showed choice behavior similar to WTB flies. Unlike the performanceof WTB flies in TE1 and TE2, the PI of mbm' flies in TE2 was lower than in TE1 (PITE1= 0.28士0.05, PInE2=0.15土0.06; P < 0.01), and the significance level of torque distributions between blueT and green⊥domains was also reduced (for TE1: F1.88= 52.08, P < 0.001; for TE2: F188=6.95, P < 0.01). But analysis still demonstrated that mbm' flies made definite choices according tocolor cue when facing contradictory shape and color cues.Table 3 Performance indices and torque distributions (average土S.D.) (xl0-°Nm) of WTB and mbm' fliesduring choice processesANOVA-test re-FlyExperi-CIstrainmentalPIs士SE.M.Domain 1Domain 2sultsperiodFu.88PTEI0.35 +0.05blue⊥: 3.10士green T: 2.56土68.0 < 0.001WTB0.340.21*(n=9)TE20.33 +0.06blue T :3.03土green⊥: 2.51 土30.7< 0.001CI=0.370.2710TE10.28 +0.05blue⊥:2.81士green T: 2.33土52.0*=0.001mbm'0.290.22(n =9)0.15 +0.06blue T :2.66土green6.95 < 0.01 *0.330.250.37士0.06blue⊥:3.32士T:2.76土89.6**0.31(n = 10)-0.04 +0.07blueT:2.93士green⊥: 2.81 士4.09 0.05CI =.400.34.blue工:2.86+green T:2.40士29.20.22士0.040.350.240(n=12)0.02士0.05green⊥: 2.57士3.01 0.090.38).350.41士0.03blue 工:3.21士green T : 2.67士76.0 < 0.0010.320.20(n=11)preen.16.-0.34士+0.07中国煤化工10.7blue工:2.8mbm' .0.31 +0.06IYHCNMHG2.2*<(n= 13)- -0.02 +0.04blue T : 2.60+Green⊥: 2.53土2.26 0.140.30SCIENCE IN CHINA (Series C)Vol.46“*”and *** indicate levels of signifcance P <0.01 and P < 0.001 respectively.Thus, at CI= 1.0 both WTB and mbm' flies could make choices according to color cue on thebasis of its greater weightiness. In choice processing, modifications of flies' s torque distributionsdepended on the color cue. Choice behaviors of WTB and mbm' flies in experiments indicated thatthe knowledge of“visual pattern (CS)-heat (US)" association acquired using combined visual cuesof shape and color could be detected through a single cue in the choice-test period, although thereversed visual cues were competitive. Modifications of torque distributions in choice behaviorwere connected with the color cue. This fact together with the earlier results that the choice indi-ces of flies were connected to the color cue at CI = 1.0l10] elucidated that choice behaviors in fliescould be observed through both performance indices and torque distributions.2.2.2 Torque distribution of flies at CI = 0.8.At CI = 0.8, memory and choice indices andtorque distributions of WTB (n= 10) and mbm' (n= 12) flies in TE1 and TE2 were revealed (table3). Both WTB and mbm' flies showed similar performances in TE1 as at CI = 1.0. They both hadpositive memory indices (WTB, PI = 0.37土0.06; mbm' , PI = 0.22土0.04). This fact indicated thatboth strains had normal associative learning and memory in training and memory-test periods.Torque analysis for both strains in TE1 indicated they had a significantly higher level of averagedtorque ampltudes in the blue⊥domain than in the green T domain (WTB: Fr188= 89.67, P <0.001; mbm': F1r.88= 29.20, P < 0.001). However, in the choce-test period (TE2), performancesof both WTB and mbm' flies were distinctly different from those of CI = 1.0. Choice indices ofboth strains were near zero (WTB, PI= -0.04士0.07; mbm', PI = 0.02士0.06), which indicatedthat they could not decide which pattern to pursue and which pattern to escape from, and theychose their flight direction on the basis of color or shape equally. At the same time, torqueanalysis showed that for both strains in TE2 the difference of torque distributions between the blueT and the green ⊥domains was not significant (WTB: F1.88]= 4.09, P = 0.05; mbm' : F1.88=3.01, P = 0.09), and the torque distribution in the two domains was symmetric.Color and shape cues have equal relative weightiness at CI = 0.8101. The results showed thatboth WTB and mbm' flies failed to make choices when facing contradictory visual cues that hadequal relative weightiness, and this disability in choice processing was confirmed by PI and torquedistribution data.2.2.3 Torque distribution of flies atCI = 0.7.Choice behaviors of WTB (n= 11) and mbm'(n= 13) flies at CI = 0.7 (table 3) were also studied. In contrast to choice behavior at CI = 1.0,WTB flies displayed choice behavior relying on shape at CI = 0.7, while mbm' flies were unableto make definite choices when facing contradictory visu中国煤化工In the memory-test period (TE1), both WTB andHHCNMH G_PIs and torquedistributions at CI =0.7 as atCI= 1.0 and CI = 0.7. In tne cnorce-test pernoa ( 1E2), WTB fliesshowed negative PI (PITE2 = -0.34土0.07), which indicated that WTB flies chose to avoid greenNo.4BEHAVIORAL MODIFICATION IN CHOICE PROCESS OF DROSOPHILA409⊥patterns and had preferences toward blue T patterns. Torque analysis showed that the torquedistribution of WTB flies in TE2 was also asymmetric, but the averaged torque amplitude level inthe green⊥domain was distinctly higher than in the blue T domain, which was different fromthe situation in TE1. Thus, due to the greater relative weightiness of shape cue at CI = 0.7, thechoice behavior of WTB flies was based upon the shape but not color, and the choice behavior ofWTB flies atCI = 0.7 was revealed from aspects of PIs and torque distributions. Compared to thechoice behavior of WTB flies, mutant mbm' flies' PI in TE2 was near zero (PITE2 = -0.02土0.04),and the difference of torque distributions between green ⊥and blue T domains was notsignificant (Fr.88]= 2.26, P= 0.14). The results indicated that although shape had relative salienceto color, mbm! flies could not make definite choices, and this disability was manifested in bothchoice indices and torque distributions.In conclusion, at CI=0.7 WTB and mbm! flies showed different choice behaviors that wereapparent both in PIs and in torque distributions: WTB flies exhibited the ability to discriminatesmall differences of relative salience between competitive color-shape cues and to make definitechoices. In choice processing, the asymmetry in torque distribution persisted and modifications ofWTB flies’torque distributions were connected to shape cue. In contrast, mbm' flies lacked thisdecision-making abilty, and could not make distinct choices when facing competitive visual cues;their torque distributions appeared to be symmetric in the choice-test period and there was nonoticeable modification of torque distributions.2.2.4 Torque distribution of flies around the choice point (CI=0.8).At CI = 0.8, color andshape cues have close relative salience and they affected flies' choice behavior equally. To furtherunderstand flies' choice behavior depending on CI and the difference of choice behaviors betweenWTB and mbm' strains, this study explored the choice behaviors of WTB and mbm' flies aroundthe choice point (CI = 0.8).In training and memory-test periods CI was set as 0.8, after which the matching of color andshape cues was reversed and CI was changed to 0.84 or 0.76. Thus the relative weightiness ofcolor and shape was changed. As shown in plate I, the PIs and torque distributions of two strainsin TE1 and TE2 were studied. In TE1, both kinds of flies showed preferences for green Tpatterns (positive PIre). ANOVA-test results indicated that in TE1 both strains had significantlyhigher torque amplitude levels in the green T domain (P < 0.001). When the CI was changed to0.84 in TE2, WTB flies made choices according to color cue and had preferences for green ⊥patterns, whereas mbm' flies flew toward two patterns eaualv (WTB. PLr= 0.19土0.07; mbm':PIne2= 0.03土0.06). Torque analysis results showed thal中国煤化工:ed sifcantlyhigher torque level in the blue T domain than itCNM HCGeas the torquedistributions of mbm' flies in the two domains had no significant difference (WTB: F1.881= 32.60,410SCIENCE IN CHINA (Series C)Vol.46P < 0.001; mbm': F1.881= 2.44, P= 0.12). When the CI was reduced to 0.76 in TE2, WTB fliesbegan to make choices according to shape cue and had preferences for blue T patterns; however,mbm' flies chose green⊥and blue T patterns equally (WTB: PITE2= -0.15士0.06; mbm':PInE2= -0.05土0.07; negative PITE2 indicated a preference for blue color). Torque analysis showedthat WTB flies' level of torque amplitude was significantly higher in the green⊥domain than inthe blue T domain, whereas mbm' flies produced approximately equal torque amplitude levels inboth domains (WTB: F1.88= 13.79, P < 0.001; mbm': F1.88 = 0.26, P = 0.62).From the results of the present analysis of WTB and mbm' flies and the choice behaviors ofboth strains at CI = 0.8, we can conclude that when the matching of shape and color cues waschanged and the CI value was not altered (CI = 0.8), WTB and mbm' flies both showed disabilitiesin making choices when facing competitive visual cues. When the matching of shape and colorcues was changed and the CI was altered (from 0.8 to 0.84 or 0.76), WTB flies could discriminatethe slightly different salience between competitive visual cues and choose the noticeable one formaking choices, but mbm' flies lacked the acute ability to fulfill the choice behavior. In all cases,the differences of choice behavior between WTB and mbm' flies were confirmed by both aspectsof PIs and torque distributions.3 DiscussionA new torque analysis method was used to study the conditioned behavior of Drosophila invisual operant conditioning at the flight simulator. Results showed that with the formation of aCS-US association, a process relating to flies' behavioral mode modifications was involved inconditioning. The torque output was the only behavioral output recorded in the flight simulator, asthe torque distribution mode represents the fly' s behavioral mode. Before operant training, thebehavioral mode appeared to be similar torque amplitudes produced at each angular position.Through operant training, the behavioral mode was modified and data exhibited a significantdifference between the CS+ and CS- domains. Flies produced distinctly higher torque amplitudesin the CS+ domain than in the CS- domain. Such behavioral modifications could help the flies toefficiently avoid the punishment pathway and pursue the unpunished flight pathway, which can beuseful for the formation of CS-US association. Due to the motor output variables of flies recordedon-line in the flight simulator, this work tried to study the operant conditioned behavior andchoice behavior of flies through focusing on a novel aspect of the behavioral mode. Unlike PIs,the traditional method used to describe flies' behavior, the torque analysis method first presentedin this paper offered a new method for dynamic behavioral study of flies' conditioned behavior.The most marked difference between operant cond中国煤化Iditioning is thatan animal's motor output is actively involved in:YHCNMH Gm the CS-USassociation,whereas in classical conditioning no motor output is involved and the CS-USassociation is formed passively. Heisenberg et al.l7,8] explored both operant and classicalNo.4BEHAVIORAL MODIFICATION IN CHOICE PROCESS OF DROSOPHILA411conditioned behavior of flies at the flight simulator and found that, because of the involvement ofthe torque output in operant conditioning, flies had better learning performance in operantconditioning than in classical conditioning. The theoretical explanation was also proposed thatclassical conditioning was based upon a two-term contingency of CS -US or Behavior- US, whileoperant conditioning was based upon a three -term contingency of CS- US- Behavior. Therefore,studying flies’ torque output in operant conditioning can increase knowledge about theirconditioned behavior. The results of the present study showed that the behavioral modificationprocess was closely related with the CS-US association process in the learning paradigm or withthe choice process in the choice paradigm. Modifications of torque distributions actually revealedthe direct effects of operant training on flies' behavior, while CS-US association ability andcognition-like choice ability were shown to be learned “knowledge" or“experience" accumulatedthrough repetitionl4, 14.Scientists recently have begun to explore choice behavior of Drosophila in facingcontradictory visual cuesl101. The newly designed torque analysis method offers new ways toadvance understanding of such choice behavior. The torque analysis of the choice behavior of fliesindicated that the decision made choice behavior and the process of behavioral modifications wereclosely related, both for wild-type WTB flies and for mutant mbm! flies. When flies could notmake definite decisions in choice behavior, the torque distributions looked the same as beforeoperant training. An example of this is the mbm' flies’ disability at CI = 0.7 and both WTB andmbm' flies's disabilities at CI = 0.8. In contrast, the torque distributions showed a significantlyasymmetry tendency when flies made definite decisions in the choice processes.Milner et al. [15]1 theorized that nervous systems contained multi-memory systems and animalscould have kinds of learning and memory behavior in different behavioral paradigms; studies onvertebrates demonstrated that different forms or different orders of behavioral performance inlearning paradigms might be connected with different neural substrates or signal systems. With acombination of behavioral studies and molecular genetics and other research methods, learningand memory of Drosophila in olfactory classical conditioning has been widely investigated16l.Results have shown that mushroom bodies played an important role in such learning task, andneurotransmission in mushroom bodies was necessary for memory retrieval but not for acquisitionor storage of memoryl17, 181. Works on visual operant conditioning of Drosophila proved thatmushroom bodies were dispensable for visual associative learning51 but were necessary for somecognition-like functionsI9, 10], while evidence is still lacking to elucidate the cellular and molecularmechanisms in mushroom bodies and some other brain structures that are involved in visualassociative learning or cognition-like behaviors. Similarly, the present behavioral studies withwild-type and mutant flies were insufficient to el中国煤化工underlying thebehavioral modifications of flies. Martin et al.[19] reveaHC N M H G could suppressthe locomotor activities. In contrast, the torque analysis in this study showed that the loss ofmushroom bodies in flies did not noticeably interfere in their torque production or otherSCIENCE IN CHINA (Series C)Vol.46behavioral aspects; it did not affect the CS-US association, but ultimately affected their choiceability. Thus, mushroom bodies seem to serve as control units in flies' nervous systems toparticipate in the integration process of competitive multi information pathways; in choicebehavior between contradictory visual cues, its function may be to emphasize one cue but supressanother.Behavioral analysis and results from molecular genetic and neuroscience studies are stillinsufficient to clarify the differences or relationships among behavioral modification process,visual associative learning and choice behavior in Drosophila visual operant conditioning. In-depth and thorough studies of their molecular, cellular and neural bases for them are stillnecessary. The present work is the first behavioral study on torque distributions of flies in visualoperant conditioning. Results from torque analysis of wild-type WTB and mutant mbm' fliesshowed that the behavioral modification process was closely related to the CS-US associativelearning process and choice process, which indicated that they might have related neural bases.The clear genetic and developmental background of Drosophila and modern moleculargenetic techniques have led to the practice of obtaining single gene mutants of Drosophila thathave certain functional phenotype or behavioral phenotypesl'"l, making them an important model inneuroscience studies of the route of“Gene- Brain- Behavior. Multi-behavioral paradigms andanalysis methods enable the study learning/memory and cognition-like functions from variousaspects, which undoubtedly will lead to an understanding of neural functions in Drosophila andeven in mammals.Acknowledgements This work was supported by the National Natural Science Foundation of China Grant Nos.39770187 & 69835020), the Multidisciplinary Research Program of Chinese Academy of Sciences, and the Major State BasicResearch Program (G2000077800).References1. Wadell, S, Quinn, W. G, What can we teach Drosophila ? What can they teach us? Trends Genet., 2001, 17:719- 726.2. Wolf, R, Heisenberg, M, Basic organization of operant behavior as revealed in Drosophila flight orientation, J. Comp.Physiol. A. 1991, 169: 699- -705.3. Guo, A. K.. Liu, L, Xia, S. Z et al, Conditioned visual flight orientation in Drosophila: Dependence on age, practice anddiet, Learm Mem., 1996, 3:49- -59.4. Xia, S. Z, Liu, L., Feng C. H. et al, Memory consolidation in Drosophila operant visual learning Lear Mem, 1997, 4:205- 218.5. Wolf, R., Wittig, T, Liu, L. et al, Drosophila mushroom bodies are dispensable for visual, tactile, and motor learmingLearn Mem, 1998,5: 166- -178.6. Wang X, Liu, L., Xia S. Z. et al., Relationship between visual leaming/memory ability and brain camp level in Droso-phila, Sci. in China, Ser. C, 1998, 41(5): 503- 511.7. Brembs, B, Heisenberg. M., The operant and the classical in con中国煤化I melanogaster, at theflight sim ulator, Learn Mem, 2000, 7: 104- -115.CNMHG8. Heisenberg M.. Wolf, R, Brembs, B.. Flexibility in a single behavMem, 2001, 8: 1-9. Liu, L, Wolf, R., Emnst, R. et al, Context generalization in Drosophila visual learmning requires mushroom bodies, Nature,No.4BEHAVIORAL MODIFICATION IN CHOICE PROCESS OF DROSOPHILA4131999, 400: 753- -756.10. Tang S. M., Guo, A. K. Choice behavior of Drosophila facing contradictory visual cues, Science, 2001, 294: 1543-1547.11. Wu,Z. H, Gong Z. F, Feng C. H. et al, An emergent mechanism of selective visual atention in Drosophila, Biol Cy-bern, 2000, 82:61- -68.12. de Belle, J. S, Heisenberg M., Expression of Drosophila mushroom body mutations in altemative genetic background: Acase study of the mushroom body miniature gene (mbm), Proc. Natl. Acad Sci, 1996, 93: 9875- -9880.13. Heisenberg M., Wolf, R, Vision in Drosophila : Genetics of Microbehavior, Berin : Springer, 1984.4. DeZazzo, J, Tully, T, Dissection of memory formation: From behavioral pharmacology to molecular genetics, TrendsNeurosci, 1995, 18:212 218.15. Milner, B., Squire, L. R, Kandel, E. R, Cognitive neuroscience and the study of memory, Neuron, 1998, 20: 445- 468.6. Tully, T, Quinn, W. G, Classical conditioning and retention in normal and mutant Drosophila melanogaster, J. Comp.Physiol. A, 1985, 157: 263- -277.17. Dubnau, J, Grady, L, Kitamoto, T. et al, Disruption of neurotransmission in Drosophila mushroom body blocks retrievalbut not acquisition of memory, Nature, 2001, 411: 476- -480.18. McGuire, S. E, Le,P. T, Davis, R. L,The role of Drosophila mushroom body signaling in olfactory memory, Science,2001, 293(553): 1330- 1333.19. Martin,J. R, Emst, R, Heisenberg M., Mushroom bodies suppress locomotor activity in Drosophila melanogaster, LearnMem, 1998, 5: 179- -191.中国煤化工MHCNMHG414SCIENCE IN CHINA (Series C)Vol.46Wang Shunpeng et al.: Behavioral modification in choice process of DrosophilaPlate I(1) (C1-0.80)(C1=0.84)(4(C1-0.80) .(C1=0.76)★西。TEITE2wTBte”口wTB(2)(5TE2:WTBTE2: WTBt230 4545-30-1515 30 45(3(6TE2:mbon'TE2; mbm'T-45 -30-150-45-30-15015 30 45Angular positionvt"When the shape and color cues were reversed and the value of CI was changed during TE2, WTB and mbm' flies showed differ-ent choice behavior when facing contradictory visual cues. For training (TR1 and TR2) and memory test (TE1) periods CI=0.8,flies were punished for the visual pattern of blue⊥and flies definitely flew toward the green T pattern. When the CI waschanged to 0.84 in the choice test period (TE2), the color cue had relatively high weight: WTB flies (n= 16) made the choiceaccording to color cue (positive PIm2 indicates the preference for the green⊥patterm), but the mbm' fies (n=16) did not makea definite choice (PITE2 ~ 0)(1). During the choice test period (TE2), the average torque amplitude level in the blue Tdomain was noticeably higher than in the green⊥domain for WTB flies (2), but the average torque amplitude levels of mbm'flies in blue T and green⊥domains were close to each other (3) . When the CI was changed to 0.76 during TE2, the shapecue had relatively high weight: WTB flies (1=14) made the choice according to shape cue (negative PI2 means a preference forthe blue T pattern), however, the mbm' flies (n=13) had no definite preference (PIπE2~ 0) (4). During TE2,the averagetorque amplitude level of WTB flies in the green⊥domain was measurably higher than in the blue T domain (5); whereas,the averaged torque amplitade levels of mbm' flies in the blue T and g中国煤化工h other(6).YHCNMHG

论文截图
版权:如无特殊注明,文章转载自网络,侵权请联系cnmhg168#163.com删除!文件均为网友上传,仅供研究和学习使用,务必24小时内删除。