We compare methods of aggregating repeated decisions from a single decision maker to improve medical image decision making using metacognitive processes (confidence judgments) and representational similarity from artificial neural networks for both novice and expert participants.