Age-related differences in the neural bases of phonological and semantic processes in the context of task-irrelevant information

Cognitive, Affective, & Behavioral Neuroscience (2019) 19:829–844 https://doi.org/10.3758/s13415-018-00671-2 Age-related differences in the neural bases of phonological and semantic processes in the context of task-irrelevant information Michele T. Diaz 1 & Micah A. Johnson 2 & Deborah M. Burke 3 & Trong-Kha Truong 4 & David J. Madden 4,5 Published online: 28 November 2018 # The Author(s) 2018 Abstract As we age we have increasing difficulty with phonological aspects of language production. Yet semantic processes are largely stable across the life span. This suggests a fundamental difference in the cognitive and potentially neural architecture supporting these systems. Moreover, language processes such as these interact with other cognitive processes that also show age-related decline, such as executive function and inhibition. The present study examined phonological and semantic processes in the presence of task-irrelevant information to examine the influence of such material on language production. Older and younger adults made phonological and semantic decisions about pictures in the presence of either phonologically or semantically related words, which were unrelated to the task. FMRI activation during the semantic condition showed that all adults engaged typical left-hemisphere language regions, and that this activation was positively correlated with efficiency across all adults. In contrast, the phonological condition elicited activation in bilateral precuneus and cingulate, with no clear brain–behavior relationship. Similarly, older adults exhibited greater activation than younger adults in several regions that were unrelated to behavioral performance. Our results suggest that as we age, brain–behavior relations decline, and there is an increased reliance on both language-specific and domain-general brain regions that are seen most prominently during phonological processing. In contrast, the core semantic system continues to be engaged throughout the life span, even in the presence of task-irrelevant information. Keywords Aging . fMRI . Semantics . Phonology . Language production In language production, one of the most commonly reported challenges for older adults is word retrieval (for a review, see Electronic supplementary material The online version of this article (https://doi.org/10.3758/s13415-018-00671-2) contains supplementary material, which is available to authorized users. * Michele T. Diaz Micah A. Johnson 1 Department of Psychology, Pennsylvania State University, University Park, PA 16803, USA 2 Department of Psychology, University of California, Los Angeles, Los Angeles, CA, USA 3 Brain Imaging and Analysis Center, Duke University School of Medicine, Durham, NC, USA 4 Department of Linguistics and Cognitive Science, Pomona College, Claremont, CA, USA 5 Department of Psychiatry and Behavioral Sciences, Duke University School of Medicine, Durham, NC, USA Burke & Shafto, 2008). These retrieval deficits are hypothesized to arise, at least in part, from impaired phonological retrieval (Burke, Locantore, Austin, & Chae, 2004; James & Burke, 2000; MacKay & James, 2004; Maylor, 1990; Vitevitch & Sommers, 2003). Consistent with this hypothesis, older adults report an increased number of tip of the tongue experiences, where one knows the meaning of a word but is unable to produce the phonology (Brown & McNeill, 1966; Burke, Mackay, Worthley, & Wade, 1991). In contrast, older adults tend to show comparable performance, relative to younger adults, in most forms of semantic retrieval. Relative to younger adults, older adults have similar levels of semantic priming (Bowles, Williams, & Poon, 1983; Burke, White, & Diaz, 1987; Madden, Pierce, & Allen, 1993), stable or increased vocabulary scores (Kemper & Sumner, 2001; Singer, Verhaeghen, Ghisletta, Lindenberger, & Baltes, 2003; Verhaeghen, 2003), and more lexically diverse language production (Kemper & Sumner, 2001), at least until their 70s (Barresi, Nicholas, Connor, Obler, & Albert, 2000; Verhaegen & Poncelet, 2013). These observations point to a larger and more elaborate semantic system for older adults, although there is considerable individual variability in 830 semantic processing, particularly at older ages, that may be influenced by educational quality (Paolieri, Marful, Morales, & Bajo, 2018), cognitive ability (Federmeier, McLennan, De Ochoa, & Kutas, 2002; Singer et al., 2003), and habits (Payne, Gao, Noh, Anderson, & Stine-Morrow, 2012). Moreover, these contrasting patterns of decline and retention in phonology and semantics suggest that these aspects of language have different cognitive and neural bases. Indeed, the transmission deficit theory proposes that while all connections decline with age, the phonological system is particularly vulnerable to decline because it has fewer converging connections (Burke, MacKay, & James, 2000; Burke et al., 1991). On the other hand, because the semantic system is more heavily interconnected, weaker links are less likely to produce a behavioral deficit. Phonological and semantic processes, however, interact with other aspects of cognition, such as working memory and executive function, which tend to show age-related decline. Inhibition, one component of executive function, refers to the ability to ignore or suppress prepotent responses or irrelevant information (Miyake et al., 2000). Indeed, during language production in healthy younger adults, there is strong evidence that overt production involves inhibition (Shao, Meyer, & Roelofs, 2013; Shao, Roelofs, Martin, & Meyer, 2015; Shao, Roelofs, & Meyer, 2012), likely at the level of response selection (Piai, Roelofs, & Schriefers, 2014). Hasher and Zacks (1988) proposed that age-related declines in the ability to inhibit irrelevant information leads to age-related slowing. For example, in language comprehension, older adults are slowed more than younger adults are when text is interspersed with distracting text, particularly when the distracting text is related to the target text (Connelly, Hasher, & Zacks, 1991), or when it is presented in an unpredictable location (Carlson, Hasher, Zacks, & Connelly, 1995). Other behavioral studies examining interactions between executive function and language production have shown that under dual task conditions, older adults are able to modify their rate of speech without compromising other factors such as fluency and grammatical complexity (Kemper, Herman, & Lian, 2003; Kemper, Schmalzried, Herman, Leedhal, & Mohankumar, 2009), at least until task demands become very high (Kemper, Schmalzried, Hoffman, & Herman, 2010). Others have shown that variability in verbal fluency, a combined measure of both production and executive function, is related to the use of prediction during language comprehension (Federmeier et al., 2002), again demonstrating the interaction of domain-general resources with both language production and comprehension. Beyond age-related decl (...truncated)