Division of Psychiatry, University of Nottingham, UK.
Behav Processes. 2020 Dec;181:104275. doi: 10.1016/j.beproc.2020.104275. Epub 2020 Oct 16.
According to the Multiplicative Hyperbolic Model (MHM), an extension of Mazur's (1987) model of delay discounting, the value of a reinforcer is an increasing hyperbolic function of its size (q). This experiment tested a prediction based on a revised version of MHM (MHM-R), according to which the maximum value may differ between different reinforcers, a possibility not envisaged in MHM. Rats were trained under a discrete-trials adjusting-magnitude schedule in which a response on lever B delivered a dilute sucrose reinforcer (0.2 M) of fixed volume, q, while a response on lever A delivered a more concentrated sucrose solution (0.4 M) the volume of which, q, was adjusted according to the rats' choices. When B was preferred in a given block of trials, q was increased in the following block; when A was preferred, q was reduced in the following block. q was varied across four phases of the experiment and the corresponding indifference magnitudes of A (q) were measured. In keeping with expectations based on both MHM and MHM-R, q increased approximately linearly with q, and 1/q with 1/q. The ratio q/q increased linearly with q with a positive intercept, a finding that is consistent with MHM-R but not with MHM.
根据多重双曲模型（MHM），这是马祖尔（1987）延迟折扣模型的扩展，强化物的价值是其大小（q）的递增双曲函数。本实验根据 MHM 的修订版（MHM-R）测试了一个预测，根据该预测，不同强化物之间的最大值可能不同，这是 MHM 中未设想的可能性。老鼠在离散试验调整幅度方案下接受训练，其中对杠杆 B 的反应会产生固定体积 q 的稀释蔗糖强化物（0.2 M），而对杠杆 A 的反应会产生更浓缩的蔗糖溶液（0.4 M），其体积 q 根据老鼠的选择进行调整。在给定的试验块中，当 B 被优先选择时，q 在下一个块中增加；当 A 被优先选择时，q 在接下来的块中减少。q 在实验的四个阶段中变化，测量相应的 A（q）的无差异量。与基于 MHM 和 MHM-R 的预期一致，q 与 q 大致呈线性关系，与 1/q 呈线性关系。q/q 与 q 呈线性关系，截距为正，这一发现与 MHM-R 一致，但与 MHM 不一致。
