Modeling age-related differences in immediate memory using SIMPLE

Abstract

In the SIMPLE model (Scale Invariant Memory and Perceptual Learning), performance on memory tasks is determined by the locations of items in multidimensional space, and better performance is associated with having fewer close neighbors. Unlike most previous simulations with SIMPLE, the ones reported here used measured, rather than assumed, dimensional values. The data to be modeled come from an experiment in which younger and older adults recalled lists of acoustically confusable and non-confusable items. A multidimensional scaling solution based on the memory confusions was obtained. SIMPLE accounted for the overall difference in performance both between the two age groups and, within each age group, the overall difference between acoustically confusable and non-confusable items in terms of the MDS coordinates. Moreover, the model accounted for the serial position functions and error gradients. Finally, the generality of the model's account was examined by fitting data from an already published study. The data and the modeling support the hypothesis that older adults' memory may be worse, in part, because of altered representations due to age-related auditory perceptual deficits. KeywordsMemory; Memory models; Aging and memory Brown, Neath, and Chater (2002; see also Neath and Brown 2006) proposed a local distinctiveness model of memory called SIMPLE (Scale Invariant Memory and Perceptual Learning) in which memory retrieval, regardless of the paradigm, is viewed as discrimination of items in terms of their location along one or more dimensions. Items with fewer close neighbors on the relevant dimensions will be better remembered than items with more close neighbors. In a typical simulation, the researcher makes an assumption about what the relevant underlying dimensions should be. For many simulations, the choice of dimensions is relatively uncontroversial; for example, for free recall, the primary dimension is typically assumed to be time (Brown et al., 2002) and for absolute identification of tones, it is assumed to be frequency

A Remember-Know Analysis of the Semantic Serial Position Function

“…In contrast, these results are consistent with the predictions of the relative distinctiveness principle (e.g., Kelley et al, 2013;Neath et al, 2006;Neath & Saint-Aubin, 2011;Surprenant, Neath & Brown, 2006). According to this view, serial position functions arise when items are ordered along one or more dimensions regardless of the hypothetical underlying memory system.…”

“…The particular dimension(s) can vary: for example, in episodic tasks, the dimension is usually temporal (i.e., relative time), but need not be; items can be ordered along perceptual dimensions or a position dimension (Surprenant et al, 2006) or any other dimension that is useful. In semantic tasks, the ordering is less likely to be temporally-based and more likely to be a nominal or logical ordering (e.g., the second verse follows the first verse); time per se is not a factor.…”

Serial position functions in general knowledge.

“…SIMPLE represents items as points along a dimension (or points in multi-dimensional space) where the dimension depends on the task. In the typical episodic task, relative time is frequently the dimension but in other settings, the dimension may be serial position (see Surprenant, Neath, & Brown, 2006) or some perceptual dimension (Neath, Brown, McCormack, Chater, & Freeman, 2006). The values on this dimension are log transformed (see Brown et al, 2007, for a detailed discussion) and performance depends on the relative distinctiveness of these log-transformed values.…”

From Brown-Peterson to continual distractor via operation span: A SIMPLE account of complex span.

“…It is also possible, in principle, to get more objective information about the dimensions. For example, Surprenant, Neath, and Brown (2006) used multidimensional scaling to obtain the main dimensions, and then used those derived dimensions in the model. Whereas their MDS solution yielded only 2 dimensions, such procedures can be extended to include more than 2 dimensions.…”