Social factors involved in the control of impulsiveness were examined in domestic chicks. In binary choices between a large/long-delay option (LL) and a small/short-delay alternative (SS), chicks that had been competitively trained in groups of three individuals showed fewer choices of LL than did those trained in isolation (experiment 1), suggesting that competition causes impulsive choice. In experiment 2, in order to identify the critical factor involved, we tested the effects of perceived competition (coincident feeding without interruption) and scrounging (gaining food without pecking bead) separately. To examine the effects of risk/noise that individual chicks experienced in competition, the food amount varied randomly in trials according to a binomial distribution around the expected mean. Perceived competition primarily contributed to the influence on the impulsive choice, whereas the contribution of scrounging was weaker. Collection risk did not explain the social influences since the perceived competition was not accompanied by actual interruption of the delayed food reward. The risk owing to variable food per se did not cause impulsive choices. Coincident foraging during competition is thought to play a critical role.
Studies on the foraging choices are reviewed, with an emphasis on the neural representations of elementary factors of food (i.e., amount, delay and consumption time) in the avian brain. Domestic chicks serve as an ideal animal model in this respect, as they quickly associate cue colors with subsequently supplied food rewards, and their choices are quantitatively linked with the rewards. When a pair of such color cues was simultaneously presented, the trained chicks reliably made choices according to the profitability of food associated with each color. Two forebrain regions are involved in distinct aspects of choices; i.e., nucleus accumbens-medial striatum (Ac-MSt) and arcopallium intermedium (AI), an association area in the lateral forebrain. Localized lesions of Ac-MSt enhanced delay aversion, and the ablated chicks made impulsive choices of immediate reward more frequently than sham controls. On the other hand, lesions of AI enhanced consumption-time aversion, and the ablated chicks shifted their choices toward easily consumable reward with their impulsiveness unchanged; delay and consumption time are thus doubly dissociated. Furthermore, chicks showed distinct patterns of risk-sensitive choices depending on the factor that varied at trials. Risk aversion occurred when food amount varied, whereas consistent risk sensitivity was not found when the delay varied; amount and delay were not interchangeable. Choices are thus deviated from those predicted as optima. Instead, factors such as amount, delay and consumption time could be separately represented and processed to yield economically sub-optimal choices.