Differential reinforcement of an approach response in zebrafish (Danio rerio).

Five zebrafish were trained to approach a target using a fully automated training procedure. During a training session, if the distance between the fish and the target was closer than an arbitrarily set distance, the approach response was reinforced by food. The fish continued to respond under this reinforcement contingency and the distance criterion could be shortened up to eighty times within a 1h session. The initial distance limit was then shortened for the next test training session. Once the initial distance criterion was reduced to a final minimum distance, the distance criterion was fixed at this value for the next nine successive sessions. In a second experiment using different fish, we manipulated approach distances in three conditions. The first condition was identical to the changing criterion training as in Experiment 1. In the second condition, only response distances under a distance criterion were reinforced. And in the last condition, only response distances over the distance criterion were reinforced. Results show that zebrafish can control the distance between themselves and a target. In other words, zebrafish are sensitive to the spatial consequences of their behavior. The present results show that a differential reinforcement paradigm can be successfully applied to zebrafish which therefore enhances their value as a vertebrate model for studies of complex behavior including visuomotor learning.

An automated device for appetitive conditioning in zebrafish (Danio rerio).

An automated device and a procedure for the operant conditioning individual zebrafish were developed. The key feature of this procedure was the construction of a simple, inexpensive feeder that can deliver extremely small amounts of food, thus preventing rapid satiation. This allows the experimenter to run multiple trails in a single test session and multiple sessions in one day. In addition, small response keys made from acryl rods and fiber sensors were developed that were sufficiently sensitive to detect fish contact. To illustrate the efficiency and utility of the device for traditional learning paradigms, we trained zebrafish in a fixed ratio schedule where subjects were reinforced with food after 10 responses. Zebrafish reliably responded on the response key for sessions that lasted as long 80-reinforcements. They also showed the traditional "break and run" response pattern that has been found in many species. These results show that this system will be valuable for behavioral studies with zebrafish, especially for experiments that need many repeated trials using food reinforcer in a session. The present system can be used for sensory and learning investigations, as well applications in behavioral pharmacology, behavioral genetics, and toxicology where the zebrafish is becoming the vertebrate model of choice.