Skip to main content
Advertisement
Browse Subject Areas
?

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

< Back to Article

Neural Correlates of Effective Learning in Experienced Medical Decision-Makers

Figure 1

Task Design and Learning Model.

A: The task consisted of a training phase followed by a testing phase. During the training phase, subjects proceeded through a fixed sequence of 64 patient encounters. They had 10 s to select one of two fictional medications. A red frame appeared around the selected medication at the time of the key press. At the end of the 10 s, they saw the outcome of their selection: either ‘SUCCESS: MI aborted’ or ‘FAILURE: No response’. The outcome remained on the screen for 6 s. A fixation cross then appeared during an intertrial interval of 4–8 s before the next encounter. The testing phase used a different permutation of encounters, and the presented outcome was always the neutral statement ‘Selection recorded’. B: In the modified Rescorla-Wagner learning model, subjects predict the efficacy of the two treatments using association rules, modeled as weighting vectors for each of the 6 patient features (plus a constant term). Following treatment choice and outcome presentation, the weighting vector for the selected treatment is adjusted according to Δβdrug = αoutcome·PE·X, where PE is the prediction error. One can therefore derive separate learning weight constants for successes and failures, using each individual's choices during testing, in combination with each individual's choices and associated outcomes during training.

Figure 1

doi: https://doi.org/10.1371/journal.pone.0027768.g001