This article was downloaded by:[Yue, Stephanie]On: 29 January 2008Access Details: [subscription number 790001677]Publisher: RoutledgeInforma Ltd Registered in England and Wales Registered Number: 1072954Registered office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK

Journal of Applied Animal WelfareSciencePublication details, including instructions for authors and subscription information:http://www.informaworld.com/smpp/title~content=t775648083

Investigating Fear in Rainbow Trout (Oncorhynchusmykiss) Using the Conditioned-Suppression ParadigmStephanie Yue a; Ian J. H. Duncan a; Richard D. Moccia aa Department of Animal and Poultry Science, University of Guelph, Canada

Online Publication Date: 01 January 2008To cite this Article: Yue, Stephanie, Duncan, Ian J. H. and Moccia, Richard D.(2008) 'Investigating Fear in Rainbow Trout (Oncorhynchus mykiss) Using theConditioned-Suppression Paradigm', Journal of Applied Animal Welfare Science,11:1, 14 - 27To link to this article: DOI: 10.1080/10888700701729106

URL: http://dx.doi.org/10.1080/10888700701729106

PLEASE SCROLL DOWN FOR ARTICLE

Full terms and conditions of use: http://www.informaworld.com/terms-and-conditions-of-access.pdf

This article maybe used for research, teaching and private study purposes. Any substantial or systematic reproduction,re-distribution, re-selling, loan or sub-licensing, systematic supply or distribution in any form to anyone is expresslyforbidden.

The publisher does not give any warranty express or implied or make any representation that the contents will becomplete or accurate or up to date. The accuracy of any instructions, formulae and drug doses should beindependently verified with primary sources. The publisher shall not be liable for any loss, actions, claims, proceedings,demand or costs or damages whatsoever or howsoever caused arising directly or indirectly in connection with orarising out of the use of this material.

Dow

nloa

ded

By:

[Yue

, Ste

phan

ie] A

t: 04

:04

29 J

anua

ry 2

008

Investigating Fear in Rainbow Trout(Oncorhynchus mykiss) Using the

Conditioned-Suppression Paradigm

Stephanie Yue, Ian J. H. Duncan, and Richard D. MocciaDepartment of Animal and Poultry Science

University of Guelph, Canada

Trout learned the operant task of pendulum-pressing for a food-reward in a mean of4.3 sessions lasting 1 hr. In a separate phase, fish also learned—through classical con-ditioning—to associate a neutral light cue with an aversive stimulus. When again al-lowed to pendulum-press for food, after aversive classical conditioning, there was adrop in the rate of responding. The mean rate dropped from 3.6–2.9 responses permin. Most important, when the light-stimulus was superimposed on a steady bout ofpendulum-pressing, trout ceased to press the pendulum and did not resume activityuntil termination of the light-stimulus (mean number of responses during a 3-min in-terval immediately prior to light-stimulus = 14.3 vs. during 3-min light-stimulus =0.1). Psychologists have used this decrease in operant responding, or “conditionedemotional response,” as a tool to examine the psychological nature of this type ofaversive conditioning. In this study, the fish demonstrated various results under thisparadigm similar to those shown by “higher” nonhuman animals, therefore challeng-ing the view of fish as unconscious, nonsentient animals.

Conflicting arguments in the research literature maintain the current debateabout whether fish have the capacity for subjective experiences or consciousfeelings. One side of the argument claims that fish lack the necessary neuroana-tomical structures (the neocortex) that allow for the phenomenon of subjectiveexperiences such as pain (Rose, 2002). Another side has shown that fish, like therainbow trout (Oncorhynchus mykiss), possess neurological features(nociceptors) that have the biological function of sensing noxious stimuli and itsaccompanying negative subjective experience: pain (Sneddon, Braithwaite &

JOURNAL OF APPLIED ANIMAL WELFARE SCIENCE, 11:14–27, 2008Copyright © Taylor & Francis Group, LLCISSN: 1088-8705 print/1532-7604 onlineDOI: 10.1080/10888700701729106

Correspondence should be sent to Stephanie Yue, Department of Animal and Poultry Science, Uni-versity of Guelph, Guelph, Ontario, Canada N1G 2W1. Email: stephyue@gmail.com

Dow

nloa

ded

By:

[Yue

, Ste

phan

ie] A

t: 04

:04

29 J

anua

ry 2

008

Gentle, 2003). Somewhere in between these two ideas is a plethora of literaturethat covers categories from cognitive capacity (Kieffer & Colgan, 1992;Overmier & Hollis, 1990) to avoidance behavior (Brown, 2001; Gallon, 1972;Pinckney, 1967) to physiological stress responses (Barton, Schreck, &Sigismondi, 1986; Mazeaud, Mazeaud, & Donaldson, 1977; Schreck, 1981;Sloman, Metcalfe, Taylor, & Gilmour, 2001).

Each category is frequently referenced by those interested in investigating thepossibility of sentience (the ability to consciously feel) in fish. There are problemsof interpreting sentience within each of these categories. It has been suggested thatmore complex, higher cognitive function comes with a higher level of sentience(Århem & Liljenström, 1997; Rose, 2002).

The opposing argument is that a high level of cognitive function is not neededto experience certain primitive feelings like fear (Dawkins, 2000); therefore, theability to feel may not necessarily be dependent on cognitive capacity. Likewise,the biological events associated with the physiological stress response can be eas-ily misinterpreted. The association between stressors and endocrine changes is notdirect. Sometimes the elevation of “stress hormones” does not indicate an animalin distress. Colborn, Thompson, Roth, Capehart, and White (1991) found thatcortisol levels increase in stallions during the act of mating (generally regarded asa rewarding activity). Finally, interpreting simple avoidance behavior as fear is notalways correct. An animal may take flight from a putatively frightening stimulusnot out of fear but simply as a result of a reflexive response, much like the humanknee-jerk reaction. Although the avoidance of a stimulus indicates that the animalfinds the stimulus aversive, a more convincing case of avoidance behavior with anemotional component consists of the response being shown to have been con-sciously performed (Dunlop, Millsopp, & Laming. 2006). In this way—instead ofbeing evoked by a reflexive, nonconscious mechanism—an intense avoidance re-sponse likely would be motivated by fear or some other emotion.

The conditioned-suppression approach has been commonly used to assess lev-els of anxiety or fear in animal subjects (Estes & Skinner, 1941). This proceduralparadigm consists of the suppression of a stable and repetitive operant behavior(lever-pressing) that is maintained by positive reinforcement (food-reward) and bya fear-evoking, conditioned stimulus (bell-ring). The aversive quality of the condi-tioned stimulus (CS) is acquired during the course of classical conditioning, wherethe subject learns to associate an initially neutral stimulus with a brief inescapablefright-stimulus (electric shock). The resulting suppression of the stable behavior isempirically measured by calculating its rate of change in the presence of the CS.The suppression ratio is commonly used to measure the strength of the suppressiveeffect. The ratio takes the form of A/(A+B), where A is the response rate during theCS and B is the response rate in the absence of the CS (measured immediately be-fore the CS presentation). According to this formula, a CS that completely sup-presses the animal’s responses will score 0.0; one that has no effect will score 0.5;

INVESTIGATING FEAR IN RAINBOW TROUT 15

Dow

nloa

ded

By:

[Yue

, Ste

phan

ie] A

t: 04

:04

29 J

anua

ry 2

008

and, in some rare instances, a stimulus that elevates responding will score between0.5 and 1.0. However, the true interest lies not in the rate of change of the ongoingbehavior but in the reason for the change. The level of suppression of the behav-ioral response is assumed to be motivated by fear, which is why this suppression isalso known as the “conditioned emotional response”). Estes and Skinner (1941)recognized early on that treating emotion purely as a response tended to overlookthe motivation behind behavioral displays.

A stimulus giving rise to “fear,” for example, may lead to muscular reactions (includ-ing facial expression, startle, and so on) and a wide-spread autonomic reaction of thesort commonly emphasized in the study of emotion; but of greater importance in cer-tain respects is the considerable change in the tendencies of the organism to react invarious other ways … our concern is most often with anxiety observed in this way, asan effect upon the normal behavior of the organism, rather than with a specific supple-mentary response in the strict sense of the term. (pp. 390–391)

Therefore, if carefully designed, an experiment may be able to give us a glimpseof the black box that lies between inner emotions and outward responses. Theconditioned- suppression method has never been tested in domesticated fish forthe purpose of investigating the phenomenon of fear.

Because this negative feeling can greatly reduce an animal’s welfare, fear is animportant area of research in aquaculture. If evidence suggests that domesticatedfish like the rainbow trout are capable of suffering from fear, we may need to re-think some current, husbandry protocols and management practices. However, ifthe traditional, hardwired view of fish remains difficult to contest, we may onlyneed to be concerned with minimizing stress for the purpose of maintaining bio-logical integrity.

MATERIALS AND METHODS

Subjects and Apparatus

Twenty-four domestic rainbow trout (970 ± 60 g, mean weight ± SE) served asthe treatment group and were obtained from Cedar Crest Trout Farm (Hanover,Ontario, Canada). The 10 rainbow trout (252 ± 11.7 g, mean weight ± SE) thatserved in the control group were obtained from the Alma Aquaculture ResearchStation (University of Guelph, Ontario, Canada). Fish were held in small groupsin 1000 L tanks (104 cm long � 104 cm wide � 31 cm height) with aerated,continuously flowing water at 9º C. They were each tagged for individual identi-fication and were fed every other day at a rate of 0.8% body weight with a com-mercial trout diet (Martin Mills, Elmira, Ontario, Canada).

16 YUE, DUNCAN, MOCCIA

Dow

nloa

ded

By:

[Yue

, Ste

phan

ie] A

t: 04

:04

29 J

anua

ry 2

008

The operant behavior chosen for this experiment was pendulum-pressing froma demand-feeder for a food-reward (Figure 1). A demand-feeder typically consistsof a food-holder with an outlet in which opening is controlled by a swing-gate. At-tached to this swing-gate is a pendulum (a long, thin rod) that extends into the wa-ter; its tip sits just below the water’s surface. Displacement of the pendulum fromits vertical position causes the swing-gate to open and release food into the water.However, demand-feeders have several disadvantages: jamming of the swing-gateor accidental food release caused by wave-action induced pendulum movement.Therefore, for the purpose of this study, the food-holder was left empty, and theexperimenter manually introduced food pellets into the tank only when the fish re-sponded by pressing the pendulum with their snouts.

The negative stimulus chosen to represent the unconditioned stimulus (US) wasa plunging dip net. Rainbow trout find this stimulus aversive, and they activelyavoid it if given the chance (Yue, Moccia, & Duncan, 2004). Last, the neutral cue

INVESTIGATING FEAR IN RAINBOW TROUT 17

FIGURE 1 A rainbow trout touches the tip of the pendulum with its mouth. A correct re-sponse to the operant task was defined as: a fish using its mouth or snout to displace the pendu-lum from its vertical position. In addition, the fish must then immediately retrieve and ingest thefood pellet.

Dow

nloa

ded

By:

[Yue

, Ste

phan

ie] A

t: 04

:04

29 J

anua

ry 2

008

(later known as the CS) used to signal the oncoming US was the illumination of ablue light.

Procedure

In Phase 1 (group learning), food pellets (0.14 g pellet) were tossed into the wa-ter as close as possible to the pendulum of the demand-feeder. This was done tofamiliarize the fish with feeding around the location of the pendulum. The pen-dulum was placed near one corner of the tank. Fish were exposed to this situa-tion as a group to encourage exploratory behavior and facilitate the feeding re-sponse. Food (ration of 0.8% body weight) was delivered in this manner everyother day for a total of 3 days.

In Phase 2 (individual training), each fish was separated from tank-mates by asolid, perforated partition that bisected the holding tank. This allowed each subjectto be trained and tested individually without having to be removed from the hometanks during any part of the experiment. Using the behavioral-shaping technique(Skinner, 1953), each fish underwent pendulum-pressing training. Initially, iso-lated fish remained motionless and unresponsive to food pellets introduced intothe tank. When swimming or movement was regained, food pellets were tossednext to the pendulum. Fish would typically dart toward the pellet, take it, and thenretreat to the tank sides or corners. This would progress until fish received pelletsreadily and then stayed within the vicinity of the pendulum. Fish would eventuallybegin to approach the pendulum on their own and were rewarded with pellets fordoing so. At this point, a pellet would only be delivered each time the fish madephysical contact with the pendulum. Finally, a pellet would be delivered only if thefish used the snout to displace the pendulum. Each fish received a 1-hr session ofindividual training every other day for 11 sessions. In this way, fish eventuallylearned to feed themselves.

All pendulum-presses were rewarded. The fish were self-feeding in this manneronly during training/experimental trials. Because they were trained and/or testedevery other day, they were consequently fed every other day. It is generally recom-mended that trout of this weight class be fed at 0.8 % body weight per day. Troutwere initially fed at this rate every other day, so satiation was not a concern. Asthey eventually learned that each pendulum-press yielded a pellet, some fishwould pendulum-press intensely and eat more than their usual ration—up to ap-proximately 2% body weight in one session. This still works out to about 1% perday, which is coincidentally close to what is recommended. Therefore, again, sati-ation was never a concern. On the final session, each fish’s prime rate (preclassicalconditioning rate of pendulum-pressing) was measured.

In Phase 3 (classical conditioning), each fish—much as in Phase 2—was sepa-rated from tank-mates by a solid partition. Here, fish learned to make an associa-

18 YUE, DUNCAN, MOCCIA

Dow

nloa

ded

By:

[Yue

, Ste

phan

ie] A

t: 04

:04

29 J

anua

ry 2

008

tion between a neutral cue and an aversive stimulus. A blue light (neutral cue) wasilluminated for an interval of 10 s, at the end of which a dip net (aversive stimulus)was plunged into the water in the middle of the tank. Each fish received 10 trials oflight-net pairings on each of 3 days. The intertrial intervals varied from 1 to 5 min.

In Phase 4 (to establish baseline rate), each fish received one 1-hr session wherethe fish was allowed to pendulum-press for food-reward. At the end of the session,the rate of response for each fish was measured because this baseline rate (post-classical conditioning rate of responding) was expected to differ from the primerate.

In Phase 5, the conditioned-suppression test was applied. Each fish was allowedto pendulum-press for food-reward for approximately 30 min. At the 30th min, theblue light was illuminated for a period of 3 min (the phase of negative conditioningwas superimposed on the phase of positive operant conditioning). The rates of re-sponding during the 3-min interval prior to the light illumination and the 3-min in-terval during the light illumination were used to generate the suppression ratio.

In addition, a control group was included to validate the light-stimulus as a trulyneutral cue. These fish were trained and tested according to the aforementionedmethod but were exempted from Phases 3 and 4. The control group had no priorexperience of being exposed to the light-stimulus.

RESULTS

Of the 10 fish in the control group, 6 successfully acquired the task of pendu-lum-pressing for food-reward. The four fish who were unable to learn the taskwere omitted from the study. Results showed that the number of sessions re-quired by each fish to acquire the pendulum-pressing task ranged from 2 to 5sessions with a mean of 3.3 sessions (Figure 2a). After completing Session 11,fish were tested for neutrality to the light-stimulus. The mean number of pendu-lum-presses before and during the light-stimulus did not differ (t = –1.02, p >.05). The mean number of pendulum-presses in the 3 min prior to illuminationwas 12.6 ± 1.38 (mean ± SE), whereas the number of responses during the 3-minlight-stimulus was 11.3 ± 1.49 (mean ± SE).

Of the 24 fish who served in the experimental treatment, data were generatedfrom 16 subjects. As in the control group, some fish did not pass Phase 2 for rea-sons ranging from health problems to the inability to learn the operant task tononresponsiveness during training. Therefore, these fish were omitted from thestudy.

The number of sessions required by each fish to acquire the pendulum-pressingtask varied greatly. The rate of learning in Phase 2 ranged from 1 to 7 sessions,with an overall mean of 4.3 sessions. Regardless of the number of sessions it tookfish to master the pendulum-pressing task, all subjects were required to continue

INVESTIGATING FEAR IN RAINBOW TROUT 19

Dow

nloa

ded

By:

[Yue

, Ste

phan

ie] A

t: 04

:04

29 J

anua

ry 2

008

with it until the end of Session 11 (Figure 2b). On the last day of Phase 2, prior toundergoing negative conditioning, the response rate was calculated for each fish;the mean prime rate for all subjects was 3.6 ± 0.4 (mean SE) responses per min.The mean baseline rate was 2.9 ± 0.3 (mean SE) responses per min. The rate ofpendulum-pressing after negative conditioning was significantly lower than be-fore it (t = –2.61, p < .05). The lower baseline rate may have been due to an initialdelay of behavioral responding. Fish started pendulum-pressing within an averageof 1.7 ± 0.8 (mean –) min before the classical conditioning phase; however, afterthe exposure to the light/dip-net pairings, fish took an average of 4.8 ± 1.6 (meanSE) min. to start responding. The average time latency to first response between

20 YUE, DUNCAN, MOCCIA

FIGURE 2a The number of sessions needed for control fish (N = 6) to learn the operant taskof pendulum-pressing.

FIGURE 2b The number of sessions needed for treatment fish (N = 16) to learn the operanttask of pendulum-pressing.

Dow

nloa

ded

By:

[Yue

, Ste

phan

ie] A

t: 04

:04

29 J

anua

ry 2

008

prime and baseline rates significantly differed (t = 2.15, p < .05). Figure 3 showsthe typical behavior of fish who are placed into their operant chamber containingthe pendulum.

The light-stimulus given during the conditioned-suppression test had a disrup-tive effect on the rate of responding, as the number of pendulum-presses shown be-fore and during the conditioned stimulus (Table 1) significantly differ (t = 16.74, p< .0001). The mean number of pendulum-presses in the 3 min prior to the illumina-tion of the light was 14.3 ± 0.9 (mean SE), whereas the mean number of responsesduring the 3-min light-stimulus was 0.1 ± 0.1 (mean SE). Therefore, the presenta-tion of the conditioned light-stimulus elicited complete behavioral suppression

INVESTIGATING FEAR IN RAINBOW TROUT 21

FIGURE 3 The rate of pendulum-pressing of a typical subject (T1G) before and after classi-cal conditioning. T1G’s prime rate of response was 4.9 presses/min (upper graph) whereas itsbaseline rate of response was 3.5 presses/min (lower graph). Note the time lag before the firstresponse is approximately 1 min before negative conditioning and 5 min afterward. A, Rate ofresponding before classical conditioning (prime rate); B, Rate of responding after classical con-ditioning (baseline rate).

Dow

nloa

ded

By:

[Yue

, Ste

phan

ie] A

t: 04

:04

29 J

anua

ry 2

008

(mean suppression ratio = 0.0). Finally, upon termination of the light-stimulus,fish were found to recover their pendulum-pressing behavior. The time taken to thefirst response varied greatly among fish (ranging from nearly 30 s to more than 13min), but all fish regained pendulum-pressing for food-reward within 14 min.

DISCUSSION

To study avoidance behavior in more depth, Coble, Farabee, and Anderson(1985) suggested the use of methods based both on punishment and reward; theyfurther suggested that these methods require the performance of more complextasks that involve more than simple reflexive avoidance responses or flight reac-tions. The present study fits all the aforementioned criteria. The results of thepresent study show that an initially neutral stimulus that had been consistentlypaired with an aversive US caused the suppression of an ongoing, appetitivelyreinforced operant behavior. When the CS (light) was presented, it disrupted thefish’s steady pendulum-pressing. Although the fish had never experienced thesetwo contexts simultaneously in one environment, their behavior was consistent

22 YUE, DUNCAN, MOCCIA

TABLE 1Number of Responses in the 3-Min Interval Before and During the CS,

Suppression Ratios, and the Time Taken to First Response AfterTermination of the CS.

SubjectAnimal

ResponsesBefore CS

ResponsesDuring CS

SuppressionRatio

Time to ResponseRecovery (min:s)

T1Y 19 1 0.05 0:25T1W 10 0 0.0 1:08T1G 12 0 0.0 7:59T1LO 15 0 0.0 2:28T1P 10 0 0.0 1:34T2Y 14 0 0.0 6:07T2W 21 0 0.0 0:30T2G 10 0 0.0 0:25T2LO 13 0 0.0 1:33T2P 16 0 0.0 0:23T3W 16 0 0.0 1:08T3G 18 0 0.0 4:24T3LO 12 0 0.0 2:22T4W 11 0 0.0 3:26T4G 13 0 0.0 13:40T4LO 18 0 0.0 1:38M ± SE 14.3 ± 0.9 0.1 ± 0.1 0.0 N/A

Dow

nloa

ded

By:

[Yue

, Ste

phan

ie] A

t: 04

:04

29 J

anua

ry 2

008

with what learning theories predict. There are two main results that could havearisen from the presentation of the CS:

1. No effect of the CS resulting in the continuation of pendulum-pressing; or2. The suppression of the operant behavior during CS presentation.

In this case, the fish completely suppressed pendulum-pressing during the CSpresentation. Presumably, the CS evoked the memory of the dip net along withthe subjective experiences that accompanied its presentation. It could be arguedthat the suppression of the operant behavior was motivated by fear. Experi-encing fear can disrupt an animal’s current behavior because the animal is bio-logically adaptive to cease the behavior being performing and attend to thethreatening situation. According to the conflicting motivation hypothesis(Davey, 1981), conditioned suppression can be explained in terms of underlyingmotivation (CS-US pairings induces a negative motivational state that conflictswith the appetitive motivational state maintaining operant responding). It is sug-gested that the negative motivation brought forth by the CS subtracts from thepositive motivation supporting operant responding, therefore decreasing the ani-mal’s incentive to respond (Davey, 1981; Mackintosh, 1974).

In contrast to this explanation, it is sometimes suggested that an equally likelyreason for the conditioned-suppression behavior is one of simple motor responseconflict—that the CS would cause behavior like freezing, therefore interferingwith instrumental responding. However, more recent neurophysiological studiesare showing evidence of this to be untrue. For example, Amorapanth, Nader, &LeDoux (1999) found that in rats trained to the typical conditioned-suppressiontask, lesions of the periaquaductal gray (PAG) in the midbrain can dissociate con-ditioned freezing (freezing refers to loss of all movement) from conditioned sup-pression (conditioned suppression in this case is merely the suppression of leverpressing); lesions of the PAG that block conditioned freezing leave other condi-tioned responses unaffected. In addition, although one might argue that the evi-dence might not apply to fish because the neuroanatomy of fish and mammalsdiffers, new evidence points to fish indeed having the homologous and function-ally analogous brain structures that are needed for emotional learning and behavior(Portavella & Vargas, 2005; Salas et al., 2006).

We contend that these new findings and line of reasoning should provide a solidargument against the motor conflict hypothesis and in favor of the motivationalconflict hypothesis. It seems reasonable to conclude that if the fish had no abilityfor subjective experiences, the CS presentation would have little effect on pendu-lum-pressing. The control group confirmed that the fear responses evoked by theCS were due to its being a learned signal and that there was no inherent aversion tothe light-stimulus on its own. Our results are consistent with Moreira and Volpato(2004), who found that tilapia do not find light cues aversive unless they have been

INVESTIGATING FEAR IN RAINBOW TROUT 23

Dow

nloa

ded

By:

[Yue

, Ste

phan

ie] A

t: 04

:04

29 J

anua

ry 2

008

trained to associate the cues with something noxious. Nonetheless, the CS had aneffect and in a manner consistent with results obtained from other animals sub-jected to this paradigm.

Fish and mice, for example, follow identical behavioral fear patterns. Miceshow flight by freezing or becoming immobile, scanning surroundings, inhibitingnondefensive behaviors, then gradually resuming normal activities (Blanchard,Briebel, & Blanchard, 2001). Mice are deemed to be sentient animals with the ca-pacity for a range of subjective experiences. Why then can these same behavioralpatterns, under the same paradigm, not be employed as evidence of the possibilityfor subjective experiences in fish?

In this study, the fish have demonstrated various interesting things also demon-strated by many “more derived” animals during conditioned suppression. For ex-ample, it was originally assumed that the rate of operant responding would bemaintained at a constant level despite the introduction of CS-US pairing. How-ever, many authors have found this to be untrue (Baker & Mercier, 1982;Millenson & de Villiers, 1972). As Hurwitz & Davis (1983) noted, there are threeprimary, dependent measures of conditioned suppression:

1. The pre-CS-US or prime rate of operant responding;2. The post-CS-US or baseline rate of responding; and3. The rate of responding during the CS presentation.

Many have found that the baseline response rates of animals are usually lowerthan their prime rates. Likewise, the response rates of the fish were signifi-cantly reduced from their prime rates. This may be explained, in part, by con-textual conditioning. The system that we used was a positive contingency (ifCS—then US). This means that the probability of the US is higher in the pres-ence of the CS than in its absence. However, the background cues, or context,can act as a type of weak and vague CS and take on ambiguous meanings be-cause the cues are present not only during negative conditioning but also dur-ing positive reinforcement. Because the classical conditioning was done in thesame environment as the operant conditioning, it is possible that backgroundcues (tank walls) from one situation carried over into the other. Some back-ground factors related to the negative conditioning were still present when re-turned to the positive operant situation. It is believed that subjects experiencegeneral apprehension in situations that become associated with fear cues(Balleine, 2000) or that fear subtly generalizes to the entire experimental situa-tion. Therefore, it is not surprising that the level of operant responding in thefish decreased after consecutive sessions of aversive conditioning. This appre-hension may also explain why the latency to first pendulum-press when ini-tially exposed to the feeding apparatus differed in sessions before and afterlight-net conditioning. Fish took significantly longer to start operant respond-

24 YUE, DUNCAN, MOCCIA

Dow

nloa

ded

By:

[Yue

, Ste

phan

ie] A

t: 04

:04

29 J

anua

ry 2

008

INVESTIGATING FEAR IN RAINBOW TROUT 25

ing, presumably because the pervasive negative state had an inhibitory or de-laying effect.

In the conditioned-suppression test, pendulum-pressing was completely sup-pressed (responses fell to zero) during the entire 3-min CS interval in every fishwith the exception of one (who responded once during the CS). It is also notewor-thy that pendulum-pressing recovered upon CS termination in all fish in a mannersimilar to that noted in other animal species undergoing conditioned suppression(Estes & Skinner, 1941; Spevack, Schulman, & Cotton, 1974). Most interesting isthe wide range in response-recovery times between fish. The time taken to first re-spond after CS termination ranged from under 30 s to almost 14 min. Through theprocess of contextual conditioning, the stimuli in the background environment cansometimes elicit conditioned responses, and the feedback from these responsesserves to motivate ongoing behavior (Petri & Govern, 2004). In this study, how-ever, pendulum-pressing and CS-US pairing were done in the same tank. There-fore, even if the context—through negative conditioning—served to suppresspendulum-pressing, the same background cues also motivated pendulum-pressing(and the motivation strengthens as the period of CS absence lengthens). Therefore,fish start to respond again. However, the individual variation contradicts the tradi-tional view of fish as hardwired organisms. Hardwired or reflexive behavior wouldpredict that all fish would recover operant responding at approximately the sametime. It would also predict that fish would recover pendulum-pressing behaviorimmediately upon CS termination. This was not the case. Although an alternativeviewpoint could argue that even if fish were rigidly programmed, recovery ofstress-induced physiological changes could differ between individuals and therebyaffect how they resume responding.

Some researchers believe that conditioning of affective states influences in-strumental performance (Balleine, 2000). In a study of conditioned suppression intwo genetic lines of chickens (Spevack et al., 1974), both lines of chickens couldlearn the operant task of key pecking for water at comparable levels, but thehigh-weight birds showed significantly more freezing in the presence of the CSthan did the low-weight birds. We attributed the differences in performance to dif-ferences in emotionality. Although we have no way of measuring the level of emo-tionality in the trout, this state is thought to be promoted by the presentation of theaversive stimulus during conditioned-suppression training (Brady & Hunt, 1955).

Finally, one point that deserves mention is that a (significant) proportion of fishwere eliminated from the study because they were not able to acquire the operanttask. However, this percentage of failure is commonly seen in experiments whereanimal training is part of the protocol. In many experiments involving train-ing—depending on the difficulty of the task—a small-to-large proportion of fishfail (Dörr & Neumeyer, 2000; Miller & Janzow, 1979). In a review of thepsychobiology of novelty-seeking, Bardo, Donohew, and Harrington (1996) notedthat individual differences in novelty-seeking are partially under genetic control;individuals within any species differ widely in the nature of their exploratory-

Dow

nloa

ded

By:

[Yue

, Ste

phan

ie] A

t: 04

:04

29 J

anua

ry 2

008

26 YUE, DUNCAN, MOCCIA

approach behavior. Behavioral reactivity to novel objects in their environment in-fluences learning of an operant task. Some fish learned the task much more slowlythan others or not at all, probably because they were initially apprehensive of thenovel pendulum and therefore less likely to make contact with it. There were somefish who appeared too stressed to learn the task, some who responded very well tofood-reward but did not understand the task, and some who were unresponsive totraining for unknown reasons.

However, upon successful operant conditioning, the ensuing pattern in the fishof avoidance responses suggests some level of conscious (therefore sentient) be-havior. Conditioned suppression has been considered a potentially valuable tech-nique for the study of fear (Estes & Skinner, 1941; Spevack et al., 1974). Theresults from this study support this suggestion. The type of behavioral variabilityshown in the operant-learning process—in addition to the results in the condi-tioned-suppression test—challenges the view of fish as unconscious, nonsentientanimals.

REFERENCES

Amorapanth, P., Nader, K., & LeDoux, J. (1999). Lesions of periaquaductal gray dissociate-conditionedfreezing from conditioned suppression behavior in rats. Learning & Memory, 6, 491–499.

Århem, P., & Liljenström, H. (1997). On the coevolution of cognition and consciousness. Journal ofTheoretical Biology, 187, 601–612.

Baker, A. G., & Mercier, P. (1982). Manipulation of the apparatus and response context may reduce theUS pre-exposure interference effect. Quarterly Journal of Experimental Psychology, 34B, 221–234.

Balleine, B. W. (2000). Incentive process in instrumental conditioning. In R. Mowrer & S. Klein (Eds.),Handbook of contemporary learning theories (pp. 307–366). Hillsdale, NJ: Erlbaum.

Bardo, M. T., Donohew, R. L., & Harrington, N. G. (1996). Psychobiology of novelty seeking and drugseeking behavior. Behavioral Brain Research, 77, 23–43.

Barton, B. A., Schreck, C. B., & Sigismondi, L. A. (1986). Multiple acute disturbances evoke cumula-tive physiological stress responses in juvenile chinook salmon. Transactions of the American Fish-eries Society, 115, 245–251.

Blanchard, D. C., Briebel, G., & Blanchard, R. J. (2001). Mouse defensive behaviors: Pharmacologicaland behavioral assays for anxiety and panic. Neuroscience and Biobehavioral Reviews, 25,205–218.

Brady, J. V., & Hunt, H. F. (1955). An experimental approach to the analysis of emotional behavior. TheJournal of Psychology, 40, 313–324.

Brown, C. (2001). Familiarity with the test environment improves escape responses in the crimson spot-ted rainbowfish (Melanotaenia duboulai). Animal Cognition, 4, 109–113.

Coble, D. W., Farabee, G. B., & Anderson, R. O. (1985). Comparative learning ability of selected fishes.Canadian Journal of Fisheries and Aquatic Sciences, 42, 791–796.

Colborn, D. R., Thompson, D. L., Roth, T. L., Capehart, J. S., & White, K. L. (1991). Responses ofcortisol and prolactin to sexual excitement and stress in stallions and geldings. Journal of AnimalScience, 69, 2556–2562.

Davey, G. (1981). Animal learning and conditioning. London: Macmillan.Dawkins, M. S. (2000). Animal minds and animal emotions. American Zoologist, 40, 883–888.

Dow

nloa

ded

By:

[Yue

, Ste

phan

ie] A

t: 04

:04

29 J

anua

ry 2

008

Dörr, S., & Neumeyer, C. (2000). Color constancy in goldfish: The limits. Journal of ComparativePhysiology, A, 186, 885–896.

Dunlop, R., Millsopp, S., & Laming, P. (2006). Avoidance learning in goldfish (Carassius auratus) andtrout (Oncorhynchus mykiss) and implications for pain perception. Applied Animal Behavior Sci-ence, 97, 255–271.

Estes, W. K., & Skinner, B. F. (1941). Some quantitative properties of anxiety. Journal of ExperimentalPsychology, 29, 390–400.

Gallon, R. L. (1972). Effects of pretraining with fear and escape conditioning on shuttlebox avoidanceacquisition by goldfish. Psychological Reports, 31, 919–924.

Hurwitz, H. M. B., & Davis, H. (1983). The description and analysis of conditioned suppression: A cri-tique of the conventional suppression ratio. Animal Learning and Behavior, 11, 383–390.

Kieffer, J. D., & Colgan, P. W. (1992). The role of learning in fish behavior. Reviews in Fish Biology andFisheries, 2, 125–143.

Mackintosh, N. J. (1974). The psychology of animal learning. London: Academic.Mazeaud, M. M., Mazeaud, F., & Donaldson, E. M. (1977). Primary and secondary effects of stress in

fish: Some new data with a general review. Transactions of the American Fisheries Society, 106,201–212.

Millenson, J. R., & de Villiers, P. A. (1972). Motivational properties of conditioned suppression.Learning and Motivation, 3, 125–137.

Miller, R. J., & Janzow, F. T. (1979). An experiment on visual discrimination in the largemouth bass,Micropterus salmoides. Proceedings of the Oklahoma Academy of Science, 59, 34–40.

Moreira, P. S. A., & Volpato, G. L. (2004). Conditioning of stress in Nile tilapia. Journal of Fish Biol-ogy, 64, 961–969.

Overmier, J. B., & Hollis, K. L. (1990). Fish in the think tank: Learning, memory, and integrated behav-ior. In R. P. Kesner & D. S. Olson (Eds.), Neurobiology of comparative cognition (pp. 205–236).Hillsdale, NJ: Erlbaum.

Petri, H. L., & Govern, J. M. (2004). Motivation: Theory, research and applications (5th ed.). Belmont,CA: Wadsworth.

Pinckney, G. A. (1967). Avoidance learning in fish as a function of prior fear conditioning. Psychologi-cal Reports, 20, 71–74.

Portavella, M., & Vargas, J. P. (2005). Emotional and spatial learning in goldfish is dependent on differ-ential telencephalic pallial systems. The European Journal of Neuroscience, 21, 2800–2806.

Rose, J. D. (2002). The neurobehavioral nature of fishes and the question of awareness and pain. Re-views in Fisheries Science, 10, 1–38.

Salas, C., Broglio, C., Durán, E., Gómez, A., Ocana, F. M., Jiménez-Moya, F., et al. (2006).Neuropsychology of learning and memory in teleost fish. Zebrafish, 3, 157–171.

Schreck, C. B. (1981). Stress and compensation in teleostean fishes: Response to social and physicalfactors. In A. D. Pickering (Ed.), Stress and fish (pp. 295–321). Cumbria, UK: Academic.

Skinner, B. F. (1953). Science and human behavior. New York: Macmillan.Sloman, K. A., Metcalfe, N. B., Taylor, A. C., & Gilmour, K. M. (2001). Plasma cortisol concentrations

before and after social stress in rainbow trout and brown trout. Physiological and Biochemical Zool-ogy, 74, 383–389.

Sneddon, L. U., Braithwaite, V. A., & Gentle, M. J. (2003). Do fishes have nociceptors? Evidence for theevolution of a vertebrate sensory system. Proceedings of the Royal Society of London B, 270,1115–1121.

Spevack, A. A., Schulman, A. H., & Cotton, B. (1974). Unconditioned and conditioned suppression intwo closely related genetic lines of chickens. Poultry Science, 53, 2020–2025.

Yue, S., Moccia, R. D., & Duncan, I. J. H. (2004). Investigating fear in domestic rainbow trout(Oncorhynchus mykiss), using an avoidance learning task. Applied Animal Behavior Science, 87,343–354.

INVESTIGATING FEAR IN RAINBOW TROUT 27