Anxiety disorders have the highest lifetime prevalence of any psychiatric disorder (Kessler et al., 2005), even in late-life (Beekman et al., 1998, Le Roux et al., 2005). Anxiety - either categorized as any one of several disorders or as a severe symptom in late life - is associated with poor quality of life and ill-being, (de Beurs et al., 1999) increased morbidities, including cardiovascular, autoimmune, neurodegenerative diseases, and increased mortality (Carroll et al., 2009, Lambiase et al., 2014, Martens et al., 2010, Roy-Byrne et al., 2008, Tully et al., 2008, Tully et al., 2016, van Hout et al., 2004).

Under the Research Domain Criteria framework (Thomas Insel et al., 2010), increased emphasis is placed on characterizing the transdiagnostic constructs underlying mental health disorders. Worry, rumination, somatic anxiety, and emotion dysregulation are such components of anxiety that present across a range of psychiatric disorders (Everaert and Joormann, 2019). Our group and others have previously reported distinct neural signatures of these phenotypes (Andreescu et al., 2015a, Feurer et al., 2021). Emotion dysregulation is a key component of anxiety (Amstadter, 2008, Cisler et al., 2010) and may contribute to its maintenance (Gross, 2013). Emotion dysregulation may be particularly important in late-life since the mechanisms of emotion regulation appear to be dynamic across the lifespan, with age-related reduction in cognitive function balanced by increased efficiency (Morgan and Scheibe, 2014).

Moreover, emerging evidence suggests anxiety may be related to cognitive impairment and subsequent development of dementia (Santabarbara et al., 2020b, Santabarbara et al., 2019). At this point the direction of causality is unclear, with evidence supporting anxiety either as a prodromal symptom (Gulpers et al., 2016, Palmer et al., 2007) or a risk factor (Gimson et al., 2018, Mah et al., 2015, Petkus et al., 2016). Atrophy of the hippocampus may provide a pathway in the pathophysiological chain from severe anxiety to cognitive decline (Mah et al., 2016). Late-life generalized anxiety is associated with poor neuropsychological function, including impairment in processing speed, working memory, inhibition, and immediate and delayed memory (Butters et al., 2011, Mantella et al., 2007). The presence of anxiety symptoms accelerates cognitive decline in individuals with amyloid-beta deposition (Pietrzak et al., 2015, Pietrzak et al., 2014). Recent prospective cohort studies have confirmed the independent effect of anxiety symptoms on increased risk of Alzheimer’s disease (AD) as well as all-cause dementia (Santabarbara et al., 2020a, Santabarbara et al., 2020b, Santabarbara et al., 2019). Our group has recently reported on the specific effect of worry and rumination on brain aging in a cohort of older adults (Karim et al., 2021).

Hippocampal atrophy is one of the standard neuroimaging hallmarks of Alzheimer's disease and related dementias (ADRD), in addition to the brain cortical volume loss and amyloid-beta deposition (Gordon et al., 2016, Jack et al., 2011, Sperling et al., 2011). While various patterns of atrophy in hippocampal subfields are described over the course of AD(Bartsch and Wulff, 2015; Burke and Barnes, 2006; de Flores et al., 2015; Van Petten, 2004), in the early phases of the disease the volume reduction are most noticeable in subiculum/presubiculum, CA1/CA3 and dentate gyrus (Blanken et al., 2017, Carlesimo et al., 2015, de Flores et al., 2015, Jack et al., 1999, Zhao et al., 2019). Such region-specific patterns of hippocampal atrophy have also been shown to differ across mood disorders and stages of cognitive decline (Hansen et al., 2021). In cognitively intact older adults, hippocampal volume is associated with a wide range of cognitive measures, including processing speed, visuospatial abilities, working memory and episodic memory (O'Shea et al., 2016).

In rodents, anxiety may negatively affect neuroplasticity (Vance et al., 2010). Rodent models of anxiety have shown the hippocampus to play a critical role (Jimenez et al., 2018, Schumacher et al., 2018), which have identified that ventral CA1 were enriched for cells that are activated in anxiogenic environments and which were required for both avoidance behavior and learning of avoidance behavior. The hippocampus has been explicitly related to mechanisms of emotion regulation (Anderson and Floresco, 2022, Engen and Anderson, 2018, Gagnepain et al., 2017, Phillips et al., 2008, Taylor, 2014). Given the evidence that worry and rumination are associated with faster rates of cognitive decline in anxiety (Beaudreau et al., 2017, Mah et al., 2021, Schoen and Holtzer, 2017), we posit that the hippocampus may provide an important link between anxiety and cognitive function. Overall, there are few studies that investigate the association of phenotypes of anxiety, in particular worry and rumination, with hippocampal volume. Most studies that investigate rumination focus on treatment resistant depression groups and not rumination itself. Only one study has shown that worry is associated with lower hippocampal volume in midlife adults (Markova et al., 2020).

In this study, we investigate the association between late-life anxiety and hippocampal volume, as well as neuropsychological measures related to cognitive decline. To our knowledge, there is no literature exploring the association between late-life anxiety, worry, or rumination and hippocampal subfields. In addition, much of this literature has focused on anxiety or fear, but not other components like worry or rumination, thus we sought to understand these associations. This present study aims to examine a) the correlations between clinical phenotypes of anxiety – global anxiety, worry, rumination, and emotion regulation strategies – and the volume of the whole hippocampus and distinct hippocampal subfields, and b) whether clinical phenotypes of anxiety and/or the volume of the hippocampus (including subfields) are associated with neuropsychological performance. We hypothesized that anxiety phenotypes in late life would be negatively associated with global hippocampal and subfields volumes, while emotion regulation strategies designed to reduce anxiety severity (e.g., cognitive reappraisal) would have a positive association with hippocampal and subfields volume. We include certain features that are often correlated to anxiety to ensure that these effects relate to anxiety phenotypes and not for instance neuroticism, general stress, or cumulative illness. Given the limited data on cognitive function and hippocampal subfields, we conducted an exploratory analysis to investigate whether any of these subregions were correlated with cognitive function.