Generalization gradients of inhibition following auditory discrimination learning.

The team taught rats to press a lever when they heard one tone. They then tested how much the rats slowed down when nearby tones played.

They wanted to map the 'inhibitory gradient' — the drop-off in responding as tones moved away from the trained pitch.

The inhibition curve was flat and wide. The rats still pressed a lot even when the tone was clearly different.

The expected sharp peak of inhibition never showed up. Discrimination did not create the tidy mirror-image of excitation the model predicted.

Schwartz et al. (1971) later got clean, steep latency gradients in monkeys. They measured how fast, not how much, the animals responded. The sharper curves show the gradient shape depends on the measure you pick.

RISLEY (1964) added a second response key and saw no drop in rate at all. Two-response methods can split a stimulus line without the broad inhibition M et al. saw.

AZRIN et al. (1963) ran the same tone test with blind, severely intellectually disabled children and got usable gradients. The basic effect holds outside the lab rat, even when the curve is lopsided.

Do not assume a single-response drop equals strong stimulus control. Flat inhibition curves warn us that slow responding can look like discrimination when it is really weak control. Try latency or two-choice probes if you need a sharper picture of auditory boundaries in speech or listening programs.