On the use of an automated learning device by group-housed dwarf goats: Do goats seek cognitive challenges?

It has been postulated that implementing appropriate cognitive challenges in the housing of captive animals allows them to use their cognitive skills, and to learn to control certain aspects of their environment. This cognitive activity is thought to be a major source of positive emotions and a prerequisite for psychological well-being of the animal. We studied the use of an automated learning device (LD) by group-housed dwarf goats. To determine if the animals seek cognitive challenges, we investigated whether the goats continued to operate the LD to get a reward when the same amount of reward was available concurrently without additional cognitive effort. In a first training period (W), a group of 10 dwarf goats were trained to press a button at a waterer (WA) to get 35 ml of drinking water. In two following periods (P1 and P2), the goats learned to discriminate two different sets of visual shapes at the LD to get the same amount of water as a reward. The WA was closed in these periods. Finally, in two experimental periods (P2r and P3), water was concurrently available at the WA and the LD for five days. We analysed the use of the LD when the WA was accessible: (a) while the goats were presented with the discrimination problem that they had learned previously (P2r); and (b) while they were given a new problem to learn (P3). The total number of button presses that triggers the delivery of water (presses at the WA and/or successful presses at the LD) was stable throughout the study, whether water was delivered only at the waterer (WA) or, after successful shape discrimination, at the LD, or if it was concurrently available at both devices. When water was concurrently available at the WA and the LD (P2r, P3), the goats constantly directed 1/3 of all daily button presses to the LD. By doing so, they gained 23.0% (P2r) and 13.8% (P3) of daily drinking water at the LD because of differing learning success. Goats showed large individual differences in the number of button presses at the LD in P2r and P3. We found a strong correlation between individual learning success and the use of the LD in P2r (rs = 0.75; P = 0.025) and P3 (rs = 0.76; P = 0.036). Results indicate that cognitive challenges, and especially successful coping with the challenge, could have at least moderate intrinsic reinforcing properties beyond the reward.