Visual attention in retarded adults: combining stimuli which control incompatible behavior.

Researchers worked with adults who have intellectual disabilities. They wanted to see how past rewards shape what people notice when two cues are shown together.

First, they taught each person two separate tasks. One shape earned points when it appeared with a red light. A different shape earned points when it appeared with a green light. Later, both shapes appeared at the same time. The team watched which shape the adults chose.

When the two shapes appeared together, adults picked the one that still led to points. They ignored the shape whose reward rule had changed. The old reward history guided their eyes and hands.

If both shapes still matched the current rule, choices were even. No single shape took over. Past rewards only won when they clashed with new rules.

Nevin (1969) first showed this effect with pigeons. The 1980 study proves the same rule holds for humans with ID. Same lab setup, new species—result holds.

Bennett et al. (1973) saw compound cues hurt accuracy. That sounds opposite, but they never gave conflicting reward histories. Without conflict, the brain averages the cues and accuracy drops. H et al. added conflict, so the brain picked a side instead of averaging.

Thrailkill et al. (2025) later showed pigeons also favor the element that paid off more often. Together, the four papers form a line: reward history → selective attention, across species and setups.

When you stack prompts or use compound visuals, check each element’s reward past. If one part used to earn breaks or snacks and now you want attention elsewhere, expect a tug-of-war. Either remove the old cue or re-train it first. Clean stimulus control from the start saves you from later errors that look like ‘non-compliance’ but are really history at work.