Cognitive aging and verbal labeling in continuous visual memory

Abstract

The decline of working memory (WM) is a common feature of general cognitive decline, and visual and verbal WM capacity appear to decline at different rates with age. Visual material may be remembered via verbal codes or visual traces, or both. Souza and Skóra, Cognition, 166, 277–297 (2017) found that labeling boosted memory in younger adults by activating categorical visual long-term memory (LTM) knowledge. Here, we replicated this and tested whether it held in healthy older adults. We compared performance in silence, under instructed overt labeling (participants were asked to say color names out loud), and articulatory suppression (repeating irrelevant syllables to prevent labeling) in the delayed estimation paradigm. Overt labeling improved memory performance in both age groups. However, comparing the effect of overt labeling and suppression on the number of coarse, categorical representations in the two age groups suggested that older adults used verbal labels subvocally more than younger adults, when performing the task in silence. Older adults also appeared to benefit from labels differently than younger adults. In younger adults labeling appeared to improve visual, continuous memory, suggesting that labels activated visual LTM representations. However, for older adults, labels did not appear to enhance visual, continuous representations, but instead boosted memory via additional verbal (categorical) memory traces. These results challenged the assumption that visual memory paradigms measure the same cognitive ability in younger and older adults, and highlighted the importance of controlling differences in age-related strategic preferences in visual memory tasks.

Verbal descriptions improve visual working memory but have limited impact on visual long-term memory.

“…When both the color and object's identity were labeled concurrently though (Experiment 3), there was only a labeling benefit in visual working memory. These findings extend previous results by confirming that labeling affects the storage of visual information in visual working memory (Souza & Skóra, 2017; see also Forsberg et al, 2020;Souza et al, 2021). Our results show that labeling adds information to the visual features stored in visual working memory, and this can lead to augmented retention of the labeled feature, even if this may come at the expense of the nonlabeled features.…”

“…When both the color and object's identity were labelled concurrently though (Experiment 3), there was only a labeling benefit in visual working memory. These findings extend previous results by confirming that labeling affects the storage of visual information in visual working memory (Souza & Skóra, 2017; see also Forsberg et al, 2020;).…”

You can’t “count” how many items people remember in working memory: The importance of signal detection-based measures for understanding change detection performance

“…There are potentially broad implications for the fact that K values confound response bias with memory performance, as K values underlie many critical conclusions about visual working memory (Alvarez & Cavanagh, 2004;Brady & Alvarez, 2015;Endress & Potter, 2014;Eriksson, Vogel, Lansner, Bergstrom, & Nyberg, 2015;Forsberg, Johnson & Logie, 2020;Fukuda, Woodman, & Vogel, 2015;Fukuda, Kang & Woodman, 2016;Irwin, 2014;Luria & Vogel, 2011;. For example, one major research domain for which our results have profound implications, is the study of how visual working memory capacity relates to global indices of cognitive function (Luck & Vogel, 2013;Vogel & Awh, 2008).…”

“…Throughout the manuscript, we focus on K values because they are incredibly common in visual working memory experiments (e.g., Alvarez & Cavanagh, 2004;Brady & Alvarez, 2015;Endress & Potter, 2014;Forsberg, Johnson & Logie, 2020;Irwin, 2014;Luria & Vogel, 2011;. However, percent correct and corrected hit rate (i.e., hits minus false alarms) also predict linear ROC curves (e.g., Swets, 1986) and thus are also invalid measures of memory performance according to our data.…”

What affects the magnitude of age-related dual-task costs in working memory? The role of stimulus domain and access to semantic representations

“…The multiple-component framework predicted that verbal processing would involve the phonological loop, and hence this would undermine rehearsal of the verbal memory items. As noted in the Introduction, this model assumes that verbal processing and verbal memory abilities are largely unaffected in healthy ageing, whereas executive functions and visuospatial function decline from early adulthood, and older people appear to use verbal strategies to perform visuospatial tasks (e.g., Forsberg et al, 2020 ; Johnson et al, 2010 ; Myerson et al, 2003 ). This could account for the smaller difference in older participants than in younger participants between the impact on memory of semantic judgements with verbal items compared with the other judgements required.…”