Delayed signal detection, differential reinforcement, and short-term memory in the pigeon.
Accuracy in delayed conditional discriminations falls fast, but steady reinforcement and mid-delay cues can keep choice on track.
01Research in Context
What this study did
Rider et al. (1984) worked with three pigeons in a three-key chamber.
Each trial lit the center key red or green.
After a delay of 0–10 s, the side keys turned red and green.
Pecking the side key that matched the sample earned food on a variable-interval schedule.
The researchers measured how often the birds picked the correct color after each delay.
What they found
Accuracy dropped fast, then leveled off.
The curve fit a simple hyperbola: sharp fall in the first 2 s, small change after 6 s.
Even when the sample was almost forgotten, the birds still followed the richer VI schedule.
Reinforcement rate, not fading stimulus control, drove final choice.
How this fits with other research
Cameron et al. (1996) used the same task and showed the curve stays the same, but adding unequal payoff ratios tips the whole curve up or down.
Bird et al. (2011) went further: flashing a big-reward cue mid-delay flips accuracy right away, proving the delay curve is not locked to time.
Locurto et al. (1980) looks like a contradiction—they paid for wrong pecks and saw accuracy crash.
The studies differ only in what gets reinforced: rewarding errors hurts, rewarding correct choices helps, both obey the matching law.
Why it matters
When you run delayed matching with learners, expect a quick drop in correct answers after two seconds.
Keep the payoff for right answers steady; it still guides choice even when the sample is weak.
Avoid accidental reinforcement for wrong picks—Locurto et al. (1980) show even rare error payoffs erase gains.
If you must stretch the delay, flash high-value praise or tokens just before the choice to pull accuracy back up, as Bird et al. (2011) did with pigeons.
Want CEUs on This Topic?
The ABA Clubhouse has 60+ free CEUs — live every Wednesday. Ethics, supervision & clinical topics.
Join Free →Keep the delay under 2 s for new learners; if you must go longer, deliver extra praise or tokens right before the choice.
02At a glance
03Original abstract
In two discrete-trial delayed-detection experiments, six pigeons were trained on dependent concurrent variable-interval schedules. Pecking a red side key was reinforced when the brighter of two white lights (S1) had been presented on the center key, and pecking a green side key was reinforced when the duller of two white lights (S2) had been presented on the center key. Incorrect responses were red side-key pecks following S2 presentations and green side-key pecks following S1 presentations; these resulted in three-second blackouts. In Experiment 1, the time between presentation of S1 or S2 on the center key and the onset of the red and green side keys was varied nonsystematically from 0.06 seconds to 19.69 seconds across experimental conditions. Stimulus discriminability decreased as the stimulus-choice delay increased. A rectangular-hyperbolic function better described this decrease in discriminability over time than did a negative-exponential function. In Experiment 2, at each of three stimulus-choice delays (0.06, 3.85, and 10.36 seconds), relative reinforcer frequency for correct responses to the red and green side keys was varied by changing the values of the dependent concurrent variable-interval schedules. The sensitivity of choice to relative reinforcer frequency was independent of the decrease in stimulus discriminability with increasing stimulus-choice delay.
Journal of the experimental analysis of behavior, 1984 · doi:10.1901/jeab.1984.42-87