Assessment of Replicability and Efforts to Refine an Operant Conditioning Procedure for Larval Zebrafish

Abstract

The zebrafish model is rapidly advancing numerous areas of basic and translational research, including toxicology, drug discovery, molecular genetics, genomic research, developmental biology, and the study of human diseases (reviewed in [...].

Keywords: animal cognition; behavioral neuroscience; discrimination learning; replication study.

Figure 1

Experimental apparatus (right) and stimuli (left) used in Experiment 1.

Figure 2

Virtual areas used to score stimulus preference in Experiment 2. Pink and green: stimulus areas, blue: no-choice area.

Figure 3

Apparatus used for post-hatching, pre-training and training periods in Experiment 3.

Figure 4

Timeline of a trial procedure (for the morning trial). (A) Subject spent inter-trial night period in a tank without stimuli. (B) At the start of the trial, two different stimuli were introduced, and preference was measured for 30 min. (C) Two different stimuli were removed and replaced with two identical stimuli (both of the reinforced type), then food was released in both compartments; larvae were left undisturbed for an hour. (D) Stimuli were removed for the duration of the next inter-trial period. A second trial was performed in the afternoon from 18:00 to 19:30 with an identical procedure.

Figure 5

Experiment 1. Performance as a function of stringency of inclusion criteria. Fish performance improved when only subjects with high inter-compartment mobility were included in the analysis. Each point represents the mean of the two daily trials.

Figure 6

Experiment 1. Performance trend over three days of individual training in this experiment (solid line) and in identical experiment of previous study (dotted line; Santacà et al. [27]).

Figure 7

Experiment 2. Learning performance as a function of stringency of inclusion criteria. Inclusion criterion was percentage of trials in which a subject performed at least one passage between the two compartments (i.e., active trials). Performance improves as more selective criteria for inclusion are applied, suggesting that less active larvae may not have learned the task. Given the tradeoff between selectivity and sample size, we used a 70% threshold (highlighted in blue) for further statistical analysis.

Figure 8

Experiment 2. Performance trend over five days of individual training. A significant preference for the reinforced stimulus emerged in the second part of the training.

Figure 9

Experiment 3. Performance trend over six days of individual training.