In many animal groups, certain individuals consistently appear at the forefront of coordinated movements [1-4]. How such leaders emerge is poorly understood [5, 6]. Here, we show that in pairs of sticklebacks, Gasterosteus aculeatus, leadership arises from individual differences in the way that fish respond to their partner's movements. Having first established that individuals differed in their propensity to leave cover in order to look for food, we randomly paired fish of varying boldness, and we used a Markov Chain model to infer the individual rules underlying their joint behavior. Both fish in a pair responded to each other's movements-each was more likely to leave cover if the other was already out and to return if the other had already returned. However, we found that bolder individuals displayed greater initiative and were less responsive to their partners, whereas shyer individuals displayed less initiative but followed their partners more faithfully; they also, as followers, elicited greater leadership tendencies in their bold partners. We conclude that leadership in this case is reinforced by positive social feedback.

No abstract

When individuals stand to gain by interacting with one another, but disagree over their preferred course of collective action, coordination can be hard to achieve [1-4]. In previous work, we found that pairs of stickleback fish prefer to synchronize their trips out of cover to look for food [5], possibly because this reduces perceived predation risk [6]. To create a degree of conflict over group coordination, we trained individual fish to expect food at one of two alternative, exposed locations and paired individuals with different expectations. Compared with isolated individuals, members of a pair showed a significantly increased tendency to alternate between foraging sites, together taking turns to visit first one individual's favored site and then the other individual's. Using a Markov-chain model to infer the individual rules underlying their joint behavior, we found that fish respond to a partner that breaks the pattern of alternation by themselves reverting to less regular behavior. Our results confirm theoretical predictions that conflict over group coordination can be resolved by taking turns [7-10] and show that, in this system, the pattern of alternation is actively monitored and maintained.

Anxiety is a cluster of responses that can involve both operant and respondent behavior, which can be both public and/or private in nature, and occurs when an upcoming aversive stimulus is signaled. Despite the reported high comorbidity of autism and anxiety, there has been very limited research on how to directly assess and treat anxiety, especially with individuals who have limited communication skills. In Study 1, anxiety was assessed in five individuals with autism, ranging in age from 10 to 19 years old. Anxiety was assessed by measuring behavior during (1) a baseline (with no putative anxiety-provoking stimuli present), (2) signals for an upcoming aversive event, and (3) exposure to that aversive event. Anxiety presented in several different ways, as both conditioned activation and suppression, and both with and without problem behavior during the aversive event. In Study 2, individualized treatments involving differential reinforcement of alternative responses and stimulus fading were used to successfully reduce anxious responding in all four participants who displayed anxiety. These studies demonstrated a potentially useful means of assessing anxiety in individuals with autism which may not only help to measure anxious behavior and identify anxiety-provoking events, but may also lead to effective treatment.