Saslow C A
J Exp Anal Behav. 1968 Mar;11(2):89-98. doi: 10.1901/jeab.1968.11-89.
Two monkeys (M. mulatta) were trained to press a telegraph key after onset of a tone and release it quickly in response to a subsequent light or click stimulus occurring after a variable interval. After training first with a fixed time limit on response latency for key release and then with a continuously adjusting limit, reaction time to click was 160 msec and to light, about 200 msec. Temporal contingencies or "payoff bands" were then introduced which reinforced only responses with latencies which fell between two limits 50 msec apart. Feedback was given as to whether each latency was too slow, within the band, or too fast. A trained monkey could precisely center its latency distribution on any 50 msec-wide payoff band located from 200 to 600 msec after the stimulus, with from 60 to 80% of its responses achieving reinforcement. Distribution statistics were comparable to those of trained human subjects. Because such precise timing might be accomplished by a peripheral adjustment, such as changing the manner of holding the key, latency of electromyographic activation was measured in participating arm muscles in one monkey. Electromyographic activation preceded key release by a constant interval, regardless of response latency, indicating a more central mechanism for timing of brief intervals.
两只恒河猴经过训练,要在听到音调后按下电报键,并在随后经过可变间隔出现的光或咔嗒声刺激时迅速松开按键。首先在按键松开的反应潜伏期设置固定时间限制进行训练,然后采用不断调整的限制进行训练,对咔嗒声的反应时间为160毫秒,对光的反应时间约为200毫秒。随后引入了时间偶然性或“回报区间”,仅对潜伏期落在相隔50毫秒的两个限制之间的反应给予强化。会反馈每次潜伏期是太慢、在区间内还是太快。一只经过训练的猴子能够将其潜伏期分布精确地集中在刺激后200至600毫秒之间的任何一个50毫秒宽的回报区间上,其60%至80%的反应能够得到强化。分布统计数据与经过训练的人类受试者的统计数据相当。由于这种精确的计时可能是通过外周调整来完成的,比如改变按键的方式,因此在一只猴子的参与实验的手臂肌肉中测量了肌电图激活的潜伏期。无论反应潜伏期如何,肌电图激活都在按键松开之前有一个恒定的间隔,这表明存在一种更中枢性的机制来对短暂间隔进行计时。