Two-Alternative Forced Choice test (2-AFC) with photographs of congeners as S+

Get Complete Project Material File(s) Now! »

Trigger activation training

After having been tested in a spatial learning test in a preliminary experiment (Peyrafort, 2018), 15 trout were selected on the basis of their speed learning performance. They were then trained to get used to the apparatus and to activate the self-feeder triggers. Three triggers were submerged for this habituation phase in order to increase their chance to activate one. During this habituation phase, the monitor displayed a grey screen (RGB code: #EEECEC) to reduce contrast with the tank colour and no matter which trigger was activated, the fish received a food reward. Furthermore, at a fish entrance in the test area, a food reward was given to make them associate netting as a positive event. The first session consisted in 5 trials of 10 minutes each. If a trigger was activated before the 10 min delay, the fish received a reward and the screen turned green (RGB code: #0F790F) during 60 seconds, which acts as a secondary positive reinforcement (Zerbolio and Royalty, 1983), the green colour having been shown to be appreciated by rainbow trout (Luchiari and Pirhonen, 2008). After this time, a 10-second Inter-Trial-Interval (ITI) with a grey screen occurred and a new trial began. If the fish did not activate the trigger within the 10 minutes delay (Cut-Off), a food reward was still given but the trial stopped. Then a new trial started after the 10s ITI.
In the next sessions, the procedure was the same as in session 1 but they lasted after 10 trials and no more reward was given after a Cut-Off. After each session the trout was netted and joined its congeners. Fish were tested randomly. An individual was considered to have successfully learned to activate the trigger once an activation was significantly different from chance level (33%) in three consecutive sessions (≥ 70%; binomial: P-value = 4.347e-05, N = 30 trials), ensuring the correct acquisition of the operant conditioning device for all subsequent exercises. Nine individuals (out of 15) correctly learned trigger activation in around 2 months (150 trials).

Training to trigger activation associated to a stimulus

After having correctly learned the operant conditioning device, the 9 remaining trout were trained for the activation of a trigger associated to the presence of a visual stimulus displayed on the screen (Fig.4). This exercise adds a difficulty because a fish is rewarded only if it chooses the trigger upper the stimulus image. One unique stimulus (S+) was displayed on the screen at each trial. Individuals were first given 3 triggers (one being rewarded, the screen behind the 2 others remaining blank). Then, only 2 triggers were still available, one being rewarded (right or left position), the screen behind the other remaining blank. If the incorrect trigger was activated, no food reward was given, and a red screen appeared (RGB code: #E90808) followed by an ITI before turning up to a new trial (Fig.4). Rainbow trout is known to avoid the red colour (Luchiari and Pirohnen, 2008) and so this colour acts as a secondary negative reinforcement. After a Cut-Off, no food reward was given. For this exercise, 10 trials were ran per session. For the 3-trigger training, different kinds of images were displayed as the positive stimulus (4 multi-coloured shapes, black shapes, blue shapes). An individual was considered to have successfully learned the task once a S+ selection was significantly different from chance level (33%) in two consecutive sessions (≥60%; binomial: P = 0.0119, N = 20 trials). For the 2-trigger training, the S+ represented a photograph of 4 different trout (Fig.4). Indeed, we expected better results with biologically relevant images. An individual was considered to have successfully learned the task once a S+ selection was significantly different from chance (50%) in two consecutive sessions (≥80%; binomial: P = 0.01182, N = 20 trials). Training for the activation of a trigger associated to a stimulus lasted 4 months. Fish were trained until a total of 520 trials before starting with the 2-AFC paradigm.

Two-Alternative Forced Choice test (2-AFC)

Only 5 of the 9 remaining individuals were used for the 2-AFC exercise because two individuals died, and the other stopped being motivated, performing only cut-offs. As the previous exercises were not as successful as expected, a new exercise was proposed to the 5-remaining fish: the 2-AFC test. It consists in presenting two contrasted stimuli simultaneously. Fish must choose between the two stimuli, one being food-rewarded and secondary reinforced by a green screen (S+). For this exercise, S+ represented a trout congener from a pool of 4 recurrent photographs. A distracting stimulus (S-) was presented at 40 cm (horizontally) from the S+. Trout were not rewarded if they activated the second trigger corresponding to the S- and a red screen appeared. The ITI was still 10 seconds. The distracting stimuli were: (stage 1) blue (RGB code: #1E90FF) abstract shapes (star, triangle, square or circle), (stage 2) a unique black abstract shape (star), (stage 3) the photograph of an object (among a pool of 60 objects), (stage 4) the photograph of another fish species (from a pool of 60 photographs), (stage 5) the photograph of unknown congeners (from a pool of 141 photographs) (Fig.5). Therefore, the discrimination difficulty between S- and S+ increased from stage 2 to stage 5. The first stage (stage 1) was proposed using a blue distracting stimulus since this colour is known to be appreciated by rainbow trout (Luchiari and Pirhonen, 2008). Due to poor results in stage 1, we chose to accentuate the contrast between S+ and S- by using a black coloured shape as S- (stage 2), the black colour being usually avoided by rainbow trout (Luchiari and Pirhonen, 2008). A total of 12 trials per session were run. Fish performed these stages in this order and moved on to the next stage as soon as they validated the learning criterion of 75% correct answers, during two consecutive sessions (different from 50%; binomial: P = 0.02266, N = 24 trials). A maximum number of 30 sessions was fixed for stages 1 to 3 (blue shape, black shape and objects as S-), a period where success is supposed to occur (Gierszewski et al., 2013). For stages 4 and 5 (other fish species and unknown trout as S-), the maximum number of sessions was fixed to 10 sessions, since we considered these last instructions too ambiguous and then impossible to learn if the two stimuli were not spontaneously discriminated. If the maximum number of sessions was reached by the subject without validating the learning criteria, the stage 2 (black shape as S-) was re-experimented to control for the initial performance of the trout. Therefore, we registered until 3 steps for stage 2. To prevent any side bias, the S+ and S- never appeared more than twice consecutively on the same side and in the case of the stage 2, a correction trial (CT) .was given on the choice of the incorrect response (S-). It allowed re-experimenting the exact same set of stimuli until a correct response was made. However, in subsequent stages, no CT were given since these stages (objects, species and trout as S-) were not training stages (like stage 2) but testing stages with increasing difficulty. Thus, if a side bias was observed, only a forced test was ran to delete it. It consisted in a unique S+ presentation on the screen on the opposite side from the preferred one.

READ  column subset selection using projection dpps 

Stage 1: Blue shapes as S-

The first stage of the 2-AFC test consisted in displaying 4 blue shapes as S- and 4 photographs of congeners as S+. Even if the scores recorded were mainly at or above chance-level and reached for a few sessions more than 75% of correct responses, none of the five individuals were able to meet the learning criterion (2 consecutive sessions) in 30 sessions completed (Fig.6). A unique black shape (star) instead of blue ones was then proposed as S- (stage 2).

Stage 2: Black shape as S-

The learning criterion was reached by the five trout when displaying a black shape as S- and the 4 recurrent photographs of congeners as S+ (stage 2-step 1). Trout N°0080 reached the criterion in 7 sessions (binomial: P=0,011 N=24 trials), N°0524 in 10 sessions (P<0.01, N=24 trials), N°0475 in 11 sessions (P=0,011 , N=24 trials), N°4494 needed 13 sessions (P=0,003 , N=24 trials) and N°4474 learned the task in 14 sessions (P=0,011, N=24 trials) (Fig.6 and Fig.7).
For step 2, four individuals succeeded in reaching the learning criterion, after only 2 sessions for N°0524 (binomial: P=0,003, N=24 trials), 3 sessions for N°0080 (P=0,003, N=24 trials), 9 sessions for N°4474 (P=0,006, N=17 trials) and 15 sessions for N°0475 (P=0,011, N=24 trials) (Fig.6). Trout N°4494 completed 30 sessions without meeting the learning criterion. However, it succeeded 4 times at 75% level, but in non-consecutive sessions.
Three trout passed step 3, and two of them met the learning criterion. Trout N°0080 succeeded in 2 sessions (binomial: P=0,011, N=24 trials) and N°4474 in 7 sessions (P=0,005, N=23 trials) (Fig.6). Trout N°4494 failed a second time to reach the learning criterion in 30 sessions done. This trout reached the 75% correct responses level only once in this step.

Stage 3: Objects as S-

When displaying objects as distracting stimuli, four individuals reached a statistically significant S+ selection frequency in 2 sessions for trout N°0080 (binomial: P=0,003, N=24 trials), 9 sessions for N°0475 (P=0,011, N=24 trials), 10 sessions for N°4494 (P=0,003 , N=24 trials) and 12 sessions for N°0524 (P=0,001 , N=36 trials) (Fig.6). Trout N°4474 reached the maximum number of 30 sessions without meeting the learning criterion. This trout reached the 75% correct responses level only once.

Global fish performances: stages comparisons

We compared the percent of correct responses between the first session and the last session (when learning criterion was met) for each exercise (Fig.9). The results indicate a significant increase between first and last sessions for the trigger habituation stage (Wilcoxon: V=0, P=0.031, N=5 individuals), for stage 2 (black shape as S-), step 1 (V=0, P=0.029, N=5) and step 2 (trend: V=1, P=0.063, N=5) and for stage 3 (objects as S-) (V=0, P=0.029, N=5). However, no differences were found for stage 1 (blue shapes as S-) (V=9, P=0.938, N=4), stage 4 (other fish species as S-) (V=8, P=0.607, N=5) and stage 2-step 3 (V=1, P=0.977, N=3). Such Wilcoxon analyses could not be performed within stage 5 since it concerned only one individual (N°0080).
The figure 10 shows that once a black shape was displayed as S- (stage 2-step 1) in the two-trigger exercises, the average percentage of correct responses (60.52%) started to exceed the chance level (binomial: P<0.05, 50%).

Table of contents :

I. INTRODUCTION
II. MATERIALS AND METHODS
1. Ethics statement
2. Experimental animals
3. Apparatus
4. General Procedure
4.1. Trigger activation training
4.2. Training to trigger activation associated to a stimulus
4.3. Two-Alternative Forced Choice test (2-AFC)
5. Stimuli
6. Statistical analyses
III. RESULTS
1. Trigger activation training
2. Training to trigger activation associated to a stimulus
3. Two-Alternative Forced Choice test (2-AFC) with photographs of congeners as S+
3.1. Fish performances within stages
3.2. Global fish performances: stages comparisons
3.3. Fish responses analyses
IV. DISCUSSION
1. Rainbow trout’s cognitive abilities
2. Individuals’ learning strategies
V. CONCLUSIONS

GET THE COMPLETE PROJECT

Related Posts