We used a midsession reversal task to investigate how temporal and situational cues may combine to determine choice in frequently changing environments. Pigeons learned a simultaneous discrimination with 2 stimuli: S1 and S2. Choices of S1 were reinforced only during the first trials, and choices of S2 were reinforced only during the last trials of the session, that is, the reinforcement contingencies reversed once during the session. To weaken the temporal cue (time into the session) that signaled the reversal trial, we varied the location of reversal trial randomly across sessions; to weaken the situational cue (the outcome of the previous trials that might support a win-stay/loose shift strategy), we varied the payoff probabilities associated with S1 and S2. Performance was consistent neither with the exclusive use of a timing strategy, nor with the exclusive use of a situational, win-stay/lose-shift strategy. Instead, choice seemed to be under joint control of both cues. The relative influence of these cues was dynamic: When payoff was higher for S1 than S2, behavior was less time-controlled than when the payoff was higher for S2 that S1, or when they were equal. We advance a descriptive mixture model of joint control for the midsession reversal task. (PsycInfo Database Record (c) 2021 APA, all rights reserved).

中文翻译：

---

不断计时，但并不总是受时间控制：来自会话中期逆转任务的证据。

我们使用会话中逆转任务来研究时间和情境线索如何结合以确定在频繁变化的环境中的选择。鸽子学会了同时区分两种刺激：S1 和 S2。S1 的选择只在第一次试验中得到强化，S2 的选择只在会话的最后一次试验中得到强化，即强化偶然性在会话期间逆转一次。为了削弱表示反转试验的时间线索（进入会话的时间），我们在各个会话中随机改变反转试验的位置；为了削弱情境线索（可能支持双赢/松散转移策略的先前试验的结果），我们改变了与 S1 和 S2 相关的回报概率。性能与单独使用计时策略不一致，也不完全使用情境化、双赢/输赢策略。相反，选择似乎在两个线索的共同控制之下。这些线索的相对影响是动态的：当 S1 的收益高于 S2 时，与 S2 的收益高于 S1 或两者相等时，行为受时间控制的程度较低。我们为会话中逆转任务提出了联合控制的描述性混合模型。（PsycInfo 数据库记录 (c) 2021 APA，保留所有权利）。我们为会话中逆转任务提出了联合控制的描述性混合模型。（PsycInfo 数据库记录 (c) 2021 APA，保留所有权利）。我们为会话中逆转任务提出了联合控制的描述性混合模型。（PsycInfo 数据库记录 (c) 2021 APA，保留所有权利）。