The Repetition Decrement Effect: A Direct Measure of Encoding Costs Attributable to Prior Experience

Abstract

The brain is the single most expensive organ in the human body (Berg, Tymoczko, & Stryer, 2002). Given that energy is scarce, evolutionary pressures ought to promote the development of cognitive systems that efficiently attend to and learn our environment (Christie & Schrater, 2015). One way of achieving efficiency involves reducing the amount of resources we devote to information that is already well-learned. Although the idea that attention is biased against redundancy is well supported (Posner & Cohen, 1984; Tipper, 1985), evidence for a similar bias in learning and memory is less clear. The classic spacing effect (Ebbinghaus, 1885) does imply that immediate repetitions triggers ‘deficient processing’ and poor memory relative to spaced repetitions (Hintzman, 1976). However, the link between the spacing effect and deficient processing relies on indirect inference. In this thesis, I propose that the repetition decrement effect (Rosner, López-Benítez, D’Angelo, Thomson, & Milliken, 2018) is a direct measure of deficient processing. The repetition decrement effect is a recognition memory deficit for words presented twice at study relative to words presented only once. In this thesis, this effect occurred when: (1) the first presentation of two identical words was poorly processed, and (2) the second presentation of two identical words followed immediately after the first. When repetitions were spaced, repetition always improved recognition. The interaction between repetition and spacing provides evidence that the repetition decrement effect is driven by the same ‘deficient processing’ mechanism that underlies the spacing effect. An instance model of memory (based on Minerva-AL; Jamieson, Crump, and Hannah, 2012) that mathematically formalises this deficient processing mechanism successfully predicted both the repetition decrement and spacing effects. The repetition decrement effect represents the strongest evidence to date that, like attention, learning mechanisms are mediated by an adaptive system that biases against the processing of redundant information.