Theory of Mind Associations With Other Cognitive Functions and Brain Imaging in Normal Aging Rebecca A. Charlton, Thomas R. Barrick, and Hugh S. Markus St. George���s University of London Robin G. Morris Institute of Psychiatry, King���s College, University of London The study investigated age-related differences in theory of mind and explored the relationship between this ability, other cognitive abilities, and structural brain measures. A cohort of 106 adults (ages 50���90 years) was recruited. Participants completed tests of theory of mind, verbal and performance intelligence, executive function, and information processing speed and underwent structural magnetic resonance imaging (measurement of whole brain volume, volume of white matter hyperintensities, and diffusion tensor imaging of white matter integrity). Theory of mind ability declined with increasing age, and the relationship between theory of mind and age was fully mediated by performance intelligence, executive function, and information processing speed and was partially mediated by verbal intelligence. Theory of mind performance correlated significantly with diffusion tensor imaging measures of white matter integrity but not with volume of white matter hyperintensities or whole-brain volume. Theory of mind age-related decline may not be independent of other cognitive functions it may also be particularly susceptible to changes in white matter integrity. Keywords: normal aging, theory of mind, cognitive function, diffusion tensor imaging Supplemental materials: http://dx.doi.org/10.1037/a0015225.supp The extent to which social cognitive function declines in normal aging is still a matter for debate. This is true in particular for theory of mind, the ability to understand the mental states of other people and use this knowledge in social interaction (as has been widely discussed in the developmental literature Happe ��, 1994a). The first studies to investigate this ability in normal aging employed a series of short stories (the Strange Stories test). In this test, inferences could be made about the thoughts or feelings of protagonists (theory of mind). Older adults performed better than younger adults, suggesting at least no decline and possibly enhanced ability (Happe ��, Winner, & Brownell, 1998). A number of more recent studies, however, have found reduced theory of mind performance in older adults (Maylor, Moulson, Muncer, & Taylor, 2002 Saltz- man, Strauss, Hunter, & Archibald, 2000 Slessor, Phillips, & Bull, 2007 Sullivan & Ruffman, 2004). One study (McKinnon & Moscovitch, 2006) demonstrated age differences in response to multiple (second-order) but not single (first-order) viewpoints. Some studies have explored whether age-related differences in theory of mind might be related to changes in other aspects of cognition rather than being specific in nature (see the developmen- tal literature Gopnik & Wellman, 1992 Stone, Baron-Cohen, & Knight, 1998). This possibility was raised initially in Happe �� et al. (1998), which suggested that the older adults in the study may have performed better because they had higher intellectual functioning. Other studies have investigated this directly with tests relating to intelligence. For example, Saltzman et al. (2000) matched younger and older adults on the Vocabulary subtest from the Wechsler Adult Intelligence Scale���Revised (Wechsler, 1981). Their older adults performed worse on a composite theory of mind measure but on only one individual measure (an adaptation of the Knower��� Guesser test Povinelli, Nelson, & Boysen, 1990). Maylor et al. (2002) identified significant correlations between theory of mind and vocabulary, information processing speed, and executive func- tion when controlling for these cognitive functions, they found that the association between theory of mind and age remained. However, Sullivan and Ruffman (2004) found in an analysis of covariance that use of fluid intelligence accounted for group dif- ferences in performance between older and younger participants. In the developmental literature, a specific association between theory of mind and executive function has been proposed execu- tive abilities have been hypothesized either to contribute to per- formance on social���cognitive abilities (Leslie, Friedman, & Ger- man, 2004 Leslie, German, & Polizzi, 2005) or to be involved in their development (Carlson & Moses, 2001 Moses, 2001). A recent study (McKinnon & Moscovitch, 2006) has demonstrated Rebecca A. Charlton, Thomas R. Barrick, and Hugh S. Markus, Clinical Neuroscience, St. George���s University of London, London, United King- dom Robin G. Morris, Department of Psychology, King���s College, Insti- tute of Psychiatry, University of London, London, United Kingdom. Additional materials are available on the Web at http://dx.doi.org/ 10.1037/a0015225.supp Research was funded by Research Into Ageing Grant 227. We thank the staff and volunteers from Church Lane Surgery, Merton Park, London, for help with recruitment and Rhonda Booth (King���s Institute of Psychiatry) for interrater reliability coding. Correspondence concerning this article should be addressed to Rebecca A. Charlton, Clinical Neuroscience, St. George���s University of London, Cranmer Terrace, London SW17 0RE, United Kingdom. E-mail: rcharlton@sgul.ac.uk Psychology and Aging �� 2009 American Psychological Association 2009, Vol. 24, No. 2, 338���348 0882-7974/09/$12.00 DOI: 10.1037/a0015225 338

that reduced efficiency of executive abilities has a significant impact on theory of mind performance both in older adults and in young adults under dual-task conditions this finding supports the notion that theory of mind relies on contributions from domain- general cognitive abilities. Maylor et al. (2002) explored the effect of memory ability using the Strange Stories test. When the task was administered in a manner that placed demands on memory, young���old adults (mean age 67) and old���old adults (mean age 81) showed a decline relative to young adults (mean age 19). When memory load was removed, only the old���old adults showed a decline relative to both other groups. The reasons for discrepancies between studies of theory of mind abilities in normal aging remain unclear. Most of these studies have used the same study design, comparing older and younger adults, and participants typically are within the same age range. It is unlikely, therefore, that discrep- ancies are a result of study design. Although there have been some differences in the tests used to measure performance, there is also commonality, in particular, in use of the Strange Stories test (Happe �� et al., 1998). It may be that differences between these cohorts (e.g., in their familiarity with testing environments, edu- cational levels, or intellectual or other cognitive abilities) are leading to different findings, but this question is currently unre- solved. A further consideration is the degree to which variations in age-related brain imaging measures might affect theory of mind. A range of brain changes occurs in the course of normal aging, such as decreases in brain volume (Matsumae et al., 1996 Resnick et al., 2000 Scahill et al., 2003) increases in the number and volume of white matter hyperintensities (WMH Breteler et al., 1994 Schmidt et al., 2005 van den Heuvel et al., 2006 Ylikoski et al., 1993) and, as demonstrated by more recent diffusion tensor im- aging (DTI) studies, a decrease in white matter integrity (Charlton et al., 2006 O���Sullivan et al., 2001 Pfefferbaum & Sullivan, 2003). We have not identified, to date, any studies on the relation- ship between structural or functional brain measures and theory of mind in normal aging. In young adults, research with functional magnetic resonance imaging (fMRI) has suggested that theory of mind abilities are related to activation of specific brain regions, such as the anterior cingulate (Gallagher & Frith, 2003 Gallagher et al., 2000 Sommer et al., 2007), the medial prefrontal cortex (Amodio & Frith, 2006 Frith & Frith, 2006), the orbitofrontal cortex (Baron-Cohen & Ring, 1994 Vo ��llm et al., 2006), the temporal poles (Gallagher & Frith, 2003 Vo ��llm et al., 2006), temporoparietal junction (Kobayashi, Glover, & Temple, 2007 Saxe, 2006 Sommer et al., 2007 Vo ��llm et al., 2006), and the inferior parietal lobule (Kobayashi et al., 2007). If theory of mind performance is supported by a network of brain regions (Abu- Akel, 2003 Vo ��llm et al., 2006), the functional connectivity be- tween the networks across different brain regions may be at least as important as the integrity of the localized neuronal structures. To explore these issues further, we measured theory of mind as part of our GENIE (St. George���s Neuropsychology and Imaging in the Elderly) study, which comprised a cohort of 106 people rang- ing in age from 50 to 90 years. Our first aim was to explore age-related differences in theory of mind within this age span and then to relate theory of mind ability to other aspects of cognitive function and to measures of brain integrity. Performance on theory of mind was compared with other aspects of cognitive function measured with a battery of tests that included tests of verbal and performance intelligence, information processing speed, and exec- utive function. In addition, we scanned all participants with MRI to provide information on brain tissue. Two main indictors were taken: normalized whole brain volume and white matter integrity as measured with a volumetric analysis of WMH and DTI. Method Participants The GENIE study is a community-based population sample that comprises 106 healthy adults (55 men, 51 women) between 50 and 90 years of age (mean 69 years the ages of participants were spread approximately evenly across 4 decades). Participants were recruited via a primary care local family doctor practice by random sampling (in the United Kingdom, all individuals are registered with a doctor���s practice as part of the National Health Service) individuals were randomly selected from those who met the study criteria, and study invitation letters were sent by post. Individuals who responded and indicated they were interested in participating were screened for suitability according to inclusion criteria by a telephone interview. Participants were required to have English as their first language, to be from a Caucasian ethnic background (due to the cultural and language demands of the neuropsychological assessment), and to have no contraindications for magnetic reso- nance scanning. Those participants with any prior psychiatric disorders (e.g., any current depression or any past episode of major or recurrent depression) or neurological disorders (e.g., stroke) were excluded. Of the 663 individuals contacted, 408 responded and 158 agreed to participate. Fifty-two individuals were excluded, in accordance with the criteria above. This left 106 individuals enrolled in the study, all of whom completed at least part of a neuropsychological assessment. All participants had data available for the verbal and performance intelligence, information process- ing speed, and executive function, but 1 participant did not com- plete the theory of mind assessment. For the MRI protocol, 1 individual declined the MRI and 1 was unable to be positioned in the MRI head coil due to severe kyphosis of the spine. Of the remaining 104 participants, all completed brain volume and fluid attenuation inverse recovery (FLAIR) scans, but 3 individuals did not complete the DTI protocol (2 due to anxiety, 1 due to a technical problem). All images were checked for quality, and 2 individuals were excluded due to a poor-quality scan resulting from artifacts in the image (from movement or metallic bridges on teeth). Scans deemed good quality for analysis were as follows: volume scan 104, FLAIR 103, DTI 99. Cognitive Testing The participants underwent a battery of standardized neuropsy- chological tests. Theory of mind performance was assessed with a subset of the Happe �� Strange Stories (Happe �� et al., 1998). This measure was chosen to facilitate comparison between the current study and previous studies (Happe �� et al., 1998 Maylor et al., 2002 Slessor et al., 2007 Sullivan & Ruffman, 2004) none of the participants had difficulties with reading, and individuals used their own spectacles to correct any eyesight deficits. Stimuli were short passages presented on one page and were followed by a question on the subsequent page four passages required under- 339 THEORY OF MIND, COGNITION, AND BRAIN IN AGING