DRYAD and ADH: Further comments on explaining age-related differences in memory.

Abstract

Recently, Smyth and Naveh-Benjamin (2016) questioned some of the main assumptions/hypotheses of DRYAD (or density of representations yields age-related deficits), a global-deficit model of aging and memory judgments (Benjamin, 2010; Benjamin et al., 2012). Smyth and Naveh-Benjamin (2016) provided empirical evidence that seems incompatible with DRYAD, but that fits the associative deficit hypothesis (ADH; Naveh-Benjamin, 2000), 1 specific-deficit theoretical view. In response, Aaron Benjamin (2016) offered a discussion of the complementary strengths and weaknesses of the DRYAD and the ADH, and the potential ways they might work together. We agree with many of his comments, but are not convinced that DRYAD is able to explain basic replicable empirical evidence of the type mentioned in Smyth and Naveh-Benjamin (2016). We discuss the reasons why we are not fully convinced by the demonstration of DRYAD's simulation of results presented in Benjamin (2016) and then present an implementation of ADH in a computationally based age-related impaired neuromodulation approach that was shown to simulate the basic empirical results of age-related associative memory deficits. We also discuss the issues of parsimony of theories and the appropriate type of representation, in the context of global versus specific deficits theoretical views. Finally, we show that the ADH's take on the distinction between items and associations has been adopted by some global computational models of memory. We believe that considerations of the above issues and others raised by Benjamin (2016) can lead to fruitful discussions that will benefit both theory development and existing knowledge of aging and memory.