Overmier and his students are generally concerned with how hedonic events not currently present can regulate and modulate later recognition and choice responding. These are fundamentally questions of the nature and functional properties of reinforcers and representations of them. They study this in animals and humans--both normal and patient populations. For example, after a stimulus has been presented only briefly and removed, animals and people often act (respond) to produce some outcome (reward); the stimulus is said to guide or cue behavior while the incentive value of the outcome is said to motivate the behavior. However, neither the stimulus nor the outcome is present at the time of the act, raising the question of how these events regulate behavior. It is not sufficient to say that "representations" do it. Overmier and his students are asking experimentally what features of events are represented (as "memories" and "expectancies") and what are the functional properties of these representations in the control of behavior. One way they do this is through discriminated conditional choice behaviors wherein each different stimulus response sequence eventuates in a different reward. With appropriate variations, this paradigm allows one to assess, for example, whether an expectancy has only general motivational properties, whether it has cue properties, or both.
Overmier and his associates have found that both animals and normal human children do form expectancies of outcome events and that these expectancies can guide or cue choice behavior. Their work with animals suggests that in fact expectancies are temporally more persistent than memories at bridging long time delays and more powerful in cuing behavior. Additionally, expectancies have been shown to be modulated by both general features of the reward and by the hedonic features of the reward, with the hedonic features apparently more powerful in producing distinctive expectancy-based cuing. Other research has shown that the anticipation of aversive events increases the vulnerability of an organism to other challenges and is likely a significant factor in psychosomatic dysfunctions such as gastric ulcer.
Because the use of differential outcomes is so powerful a technique for guiding choice behavior in the absence of an external stimulus-cue, Overmier and his associates are now beginning to ask whether this technique can form the basis of training techniques useful with learning and memory disordered subjects--animal and human. Additional work in animals is exploring whether the memory and expectancy types of representations are mediated by different neurotransmitter systems. In animals this will be done by blocking specific systems neurochemically and assessing performance, while in humans it will be done by selecting patient populations with neurochemical transmitter dysfunctions (e.g., Korsakoff patients who have cholinergic dysfunction) and assessing their performances utilizing these representations to guide their conditional choice behaviors.