Risk sensitivity in starlings: variability in food amount and food delay

Abstract

Starlings' preferences for constant versus variable food sources were studied in the laboratory. The constant alternative gave a fixed amount of food after a fixed delay. The variable alternative offered either a varying amount of food after a fixed delay (treatment A) or a fixed amount of food after a variable delay (treatment B). In both treatments the ratio of amount of food over trial length (the sum of intertrial interval plus delay and handling times) of the constant alternative equaled the average of the two ratios of the variable alternative. The variable ratios were 30% higher and 30% smaller than the fixed ratio. In free-choice trials (both options available in each trial), the subjects were risk-averse or indifferent in treatment A and indifferent or riskprone in treatment B. In no-choice trials (only one source available per trial), the latency to respond was longer in the variable than in the constant source in treatment A and the opposite in treatment B. The greater preference for variability in time than for variability in reward amount is not consistent with either maximizing the ratio of expected energy over expected time or the expected ratio of energy over time for individual trials. There was a negative correlation between individual intake rate and degree of risk proneness for both kinds of variability. We present a model of choice based on an information-processing theory for temporal memory that accounts for the different effects of variability in delay and in amount but cannot explain the effects of intake rate.

Neuro-economics in chicks: Foraging choices based on amount, delay and cost

“…A similar separation has also been documented in details using starlings as subjects [7,34,36], although the brain mechanisms have not yet been studied. In rats, on the other hand, lesions to the nucleus accumbens have been reported to enhance risk aversion for the food amount [9]; therefore, the risk sensitivity might be functionally linked to the impulsiveness [8].…”

Computational mate choice: Theory and empirical evidence

“…In contrast, the latency-ratio test performed much better than the recognition and discrimination tests, emphasizing the importance of time responses in assessing female mating preferences Leonard and Hedrick, 2010;Rebar et al, 2011;Shackleton et al, 2005), as already shown in other contexts of choice (i.e. food preferences Bateson and Kacelnik, 1995;Reboreda and Kacelnik, 1991).…”

Subjective value of risky foods for individual domestic chicks: a hierarchical Bayesian model

“…According to the SET-based hypothesis, both factors (amount and delay) are sampled with the Gaussian noise that follows Weber's law, so that the variable factor will inevitably be undervalued at recall. An undervalued amount will lead to risk aversion, whereas an undervalued delay (or overvalued proximity) will lead to risk-prone choices (Reboreda and Kacelnik 1991).…”

“…Based on the scalar expectancy theory of Gibbon (1977), they argued that the pattern of risk sensitivity could arise by constraint of information processing in which the recalled amount and time are inevitably accompanied by errors (Reboreda andKacelnik 1991, Kacelnik and. Basically, a similar pattern was found in domestic chicks and this similarity could be due to common constraints on cognitive processes rather than convergent evolution due to common selection pressure.…”