Subjective value of risky foods for individual domestic chicks: a hierarchical Bayesian model
Subjective value of risky foods for individual domestic chicks: a hierarchical Bayesian model
复制标题
DOI：
10.1007/s10071-009-0293-1
复制
发表时间：
2010-05-01
期刊：
影响因子：
2.7
通讯作者：
Matsushima, Toshiya
中科院分区：
文献类型：
--
作者：
Kawamori, Ai;Matsushima, Toshiya
对于动物来说，要决定攻击哪一个猎物，食物的获得和延迟必须整合在一个价值函数中。然而，当风险伴随时，主观价值不是通过预期盈利能力获得的。为了估计主观价值，我们研究了选择在一个十字形的迷宫与两个彩色馈线在国内小鸡。当通过食物量或延迟的逆转进行测试时，小鸡在两种条件下（实验1）改变了类似的选择。因此，我们研究了风险敏感性的数量和延迟（实验2），通过提供一个馈线与固定盈利能力的食物和替代馈线与高或低盈利能力的食物在相同的概率。盈利能力在数量上（组1和2在高和低方差）或延迟（组3）。为了找到平衡，在总共18个区块中调整固定喂食器中食物的量（组1和2）或延迟（组3）。将马尔可夫链蒙特卡罗方法应用于多层贝叶斯模型，估计主观价值。在群体水平上，鸡低估了第1组的可变饲养员，第2组无差异，但高估了第3组的可变饲养员。没有滴定程序的重新检查（实验3）表明，每个选项的主观值不是绝对的。当延迟是变化的，可变选项往往是一个矛盾的高价值取决于固定的选择。因此，唯一确定的价值函数的基本假设可能会受到质疑。
For animals to decide which prey to attack, the gain and delay of the food item must be integrated in a value function. However, the subjective value is not obtained by expected profitability when it is accompanied by risk. To estimate the subjective value, we examined choices in a cross-shaped maze with two colored feeders in domestic chicks. When tested by a reversal in food amount or delay, chicks changed choices similarly in both conditions (experiment 1). We therefore examined risk sensitivity for amount and delay (experiment 2) by supplying one feeder with food of fixed profitability and the alternative feeder with high- or low-profitability food at equal probability. Profitability varied in amount (groups 1 and 2 at high and low variance) or in delay (group 3). To find the equilibrium, the amount (groups 1 and 2) or delay (group 3) of the food in the fixed feeder was adjusted in a total of 18 blocks. The Markov chain Monte Carlo method was applied to a hierarchical Bayesian model to estimate the subjective value. Chicks undervalued the variable feeder in group 1 and were indifferent in group 2 but overvalued the variable feeder in group 3 at a population level. Re-examination without the titration procedure (experiment 3) suggested that the subjective value was not absolute for each option. When the delay was varied, the variable option was often given a paradoxically high value depending on fixed alternative. Therefore, the basic assumption of the uniquely determined value function might be questioned.