Generalization peak shift in rats under conditions of positive reinforcement and avoidance.

Scientists trained rats to tell two click speeds apart. One speed meant food. The other speed meant mild shock.

After training, the team played many click speeds. They watched how the rats responded to each new speed.

The goal was to see if the well-known peak-shift effect shows up under avoidance just like it does under food reward.

Peak shift appeared in both setups. The strongest responding moved away from the trained cue in the same direction.

Food or shock made no difference. The gradient slid the same way, showing the shift is about learned associations, not about seeking or escaping.

Hendry et al. (1969) first saw asymmetrical fear gradients in rats. The new study adds the clear peak-shift form under avoidance, closing a gap.

WINOGRAD (1965) gave the method that makes clean suppression gradients. Grosch et al. (1981) used that method to prove peak shift under aversive control.

Neuringer (1973) found flat gradients after simple presence-absence training with pigeons. The rat study shows intradimensional avoidance training gives the sharp peak shift that presence-absence training missed.

You can expect peak shift any time a learner tells two cues apart, whether the outcome is goodies or removal of something bad. When you see responding drift away from the trained cue, do not assume the contingency type is the cause. Check the discrimination history instead.