In order to speed up reward-oriented behaviors and thereby maximizing rewards per unit time, humans and animals mainly rely on two mechanisms.
Their ability of finding good objects therefore depends on this kind of value-based learning,[3][4] where consistency plays an important role.
As a result of this comparatively consistent object-value mapping across people's lifespan, their internal value assignment to objects tends to be stable.
[7] This automatic gaze or attention shift is pronounced even when no reward is expected, potentially increasing the subject's chance of survival in some critical circumstances.
[2] Studies have shown that when humans have extensively learned to find some particular objects, they start to recruit much less cognitive resource in order to finish the task.
[8][9] A study on monkeys with free-viewing procedure for testing gaze shift bias has also provided evidence to support the involuntary nature of object skill.
The core issue then would be to figure out an optimal balance between relying on object skill and referring to local changes.