Age differences in retrieval-related reinstatement reflect age-related dedifferentiation at encoding

Abstract

Age-related reductions in neural specificity have been linked to cognitive decline. We examined whether age differences in specificity of retrieval-related cortical reinstatement could be explained by analogous differences at encoding, and whether reinstatement was associated with memory performance in an age-dependent or age-independent manner. Young and older adults underwent fMRI as they encoded words paired with images of faces or scenes. During a subsequent scanned memory test participants judged whether test words were studied or unstudied and, for words judged studied, also made a source memory judgment about the associated image category. Using multi-voxel pattern analyses, we identified a robust age-related decline in scene reinstatement. This decline was fully explained by age differences in neural differentiation at encoding. These results suggest that, regardless of age, the specificity with which events are neurally processed at the time of encoding determines the fidelity of cortical reinstatement at retrieval.

Age-related differences in encoding-retrieval similarity and their relationship to false memory

“…Third, due to the of current study design, every item viewed during encoding is its own perceptual category. We are therefore unable to calculate neural distinctiveness during encoding to use as a baseline neural similarity measure, as recent studies have with face and house categories of visual stimuli (Hill et al, 2021;Kobelt et al, 2021). However, we contend that global ERS is an appropriate and theoretically meaningful baseline assessment of recapitulation between younger and older adults when assessing single-item ERS.…”

“…Either possibility would result in overall reduce similarity between neural patterns within occipital cortices across encoding and retrieval. Recent work from Hill et al (2021) provides evidence for the former explanation, as age deficits in neural reinstatement became nonsignificant after controlling for the distinctiveness of neural patterns in visual regions during encoding. Similarly, age deficits in single-item ERS within the current study became nonsignificant after controlling for global ERS capturing the neural pattern similarity of all items viewed at encoding (see below for more discussion on global ERS deficits and relation to behavior).…”

“…Reduced neural recapitulation in sensory cortices is likely a reflection of impaired reactivation of perceptual details, as predicted by the sensory reactivation hypothesis (Slotnick & Schacter, 2004). For example, it is possible that older adults do not store, and thus do not have available as many or as high-quality, encoding-related details for retrieval compared to their younger counterparts (Dennis, Bowman, et al, 2014;Hill et al, 2021). It is also possible that older adults have such details stored in their memory, but fail to reactivate these memory traces at the time of retrieval (Trelle et al, 2020;Naspi et al, 2021).…”

“…Recent research suggests that age differences in neural recapitulation can be accounted for by baseline differences in neural similarity between younger and older adults, as older adults have reduced pattern similarity compared to younger adults (Hill et al, 2021). To account for the possibility of such baseline age differences in the current study, we calculated a global ERS value within each region of interest to estimate the general degree of recapitulation between items viewed during encoding and retrieval (i.e., the similarity of all items at retrieval and all items at encoding).…”

“…Since it is possible that age effects in ERS can be accounted for by a general reduction in neural similarity in older compared to younger adults (Hill et al, 2021), we investigated the effect of global ERS on age differences in single-item level ERS. To do so we repeated the above 2(hemisphere: left, right) x 3(trial type: target, item lure, theme lure) x 2(age group: young, old) within-between repeated measures ANOVA while including global ERS as a covariate.…”