Tests of the DRYAD theory of the age-related deficit in memory for context: not about context, and not about aging

Abstract

Older adults exhibit a disproportionate deficit in their ability to recover contextual elements or source information about prior encounters with stimuli. A recent theoretical account, DRYAD, attributes this selective deficit to a global decrease in memory fidelity with age, moderated by weak representation of contextual information. The predictions of DRYAD are tested here in three experiments. We show that an age-related deficit obtains for whichever aspect of the stimulus subjects' attention is directed away from during encoding (Experiment 1), suggesting a central role for attention in producing the age-related deficit in context. We also show that an analogous deficit can be elicited within young subjects with a manipulation of study time (Experiment 2), suggesting that any means of reducing memory fidelity yields an interaction of the same form as the age-related effect. Experiment 3 evaluates the critical prediction of DRYAD that endorsement probability in an exclusion task should vary nonmonotonically with memory strength. This prediction was confirmed by assessing the shape of the forgetting function in a continuous exclusion task. The results are consistent with the DRYAD account of aging and memory judgments and do not support the widely held view that aging entails the selective disruption of processes involved in encoding, storing, or retrieving contextual information.

Figure 1

Different theoretical views of the age-related deficit in context memory. Under the selective-deficit view (top portion of figure), older adults encode, store, or retrieve the contextual aspects of the stimulus less faithfully than the item (the car). Under the global-deficit view (DRYAD), the entire stimulus is maintained less faithfully in the elderly, but attention during encoding moderates the resulting deficit. Thus, greater attention to the car (indicated by its greater size and brightness) and lesser attention to the background leads to a disproportionate deficit in memory for the background.

Figure 2

Memory performance by younger and older adults on tests of item and context memory in the chair-room experiment (Experiment 1). Top graphs plot performance as percent correct and bottom graphs plot performance as d′. Left graphs shows performance from both conditions combined; middle graphs shows the condition in which memory for chairs was emphasized; right graphs shows the condition in which memory for rooms was emphasized. Error bars represent standard error pooled over age groups for each type of test.

Figure 3

The effects of age group and memory test type on discriminability (d_a) as a function of study time (Experiment 2). Error bars represent standard error pooled over age groups for each type of test.

Figure 4

Discriminability (d_a) for recognition and exclusion as a function of study time for younger subjects (Experiment 2). Error bars represent standard error pooled over all cells in the 2×3 within-subjects design.

Figure 5

A plot of the theoretical relationship between memory strength and endorsement probability for to-be-endorsed old items (HR), to-be-excluded old items (FAR-TBX), and new items (FAR-new). FAR-TBX can either rise or fall with additional memory strength depending on the position within the function.

Figure 6

Empirical confirmation of the theoretical relationship shown in Figure 5. Hit rates (HR), false-alarm rates to to-be-excluded items (FAR-TBX), and false-alarm rates to new items (FAR-new) as a function of the study-test retention interval in the continuous exclusion task (Experiment 3).