From Basic to Applied Science: The Case of the Differential Outcomes Procedure [Review]

In this article the authors set out “to illustrate how the differential outcomes procedure has evolved to address important questions related to discriminative learning and memory in healthy humans and in clinical populations”. I like the article and the work made by the authors; I find it very clear and well-written, and of interest for a wide community of researchers and practitioners. Although a caveat is probably necessary: I am an expert in experimental psychology, but I have little to none experience with special populations; I think that it would be very interesting if a clinical expert/researcher could review the paper too. I have some comments for the authors, mostly related with the basic experimental psychology part of the review; maybe they find them helpful or interesting in some way. Please, do not feel that I am forcing you to include these ideas in the final article. Associative model. I wonder if the OR theory for goal-directed action, aka the associative-cybernetic model, could explain the DOE. Among other things, this theory proposes that during reinforcement learning the animals can learn outcome-response links, which might guide goal-directed behaviour. Hence, having only one O for each R should facilitate learning and performance according to this theory (see de Wit & Dickinson, 2009). Differences in motivation. Also, it occurs to me that the advantage of the DOE group over its control conditions might be explained by differences in motivation. Let me use an example. When you are buying an item in a vending machine, you get a different item (outcome) depending on your response (pressing the option A1, A2, etc in the keypad). Each response has one outcome, so you can select what you want in that moment. This “normal vending machine” (vending machine #1) would be like the DOP procedure. Following the example, the NOP would be a machine (vending machine #2) in which no matter your response, you get an item randomly selected. The common-outcome procedure would be like a vending machine (vending machine #3) in which no matter your response you can only get one item (eg., water). It is easy to see that the vending machine #1 will be more popular than the other two, just because you can select what you want in that moment. If I am not missing something, these differences in motivation (you are more motivated when you can select what you want) could explain the DOE. Is my analysis correct? Task difficulty. The authors claim that the DOE is especially beneficial when the task is not easy. My first thought is that these results are produced by a ceiling effect: When the task is very easy, the performance is almost perfect in all conditions so you cannot distinguish between them. This explanation in terms of a ceiling effect does not need the assumption of a compensatory mechanism in those participants with brain damage or/and cognitive impairment. I have the impression that the authors have overlooked this possibility. Again, maybe I am missing something important, if that is the case, I am sorry! Minor p 3. The Pavlovian example. I think it would be better to use an operant conditioning example (S-R-O) and not a Pavolovian one (S-O), because the tasks and experiments reviewed in this article are mostly about operant conditioning. p 5. I think that the use of the term “handicapped” is not recommended. References: de Wit, S., & Dickinson, A. (2009). Associative theories of goal-directed behaviour: a case for animal–human translational models. Psychological Research, 73, 463-476.