The present study replicated previous findings that daily functioning is notably comprised in a large, well-defined group of MCI patients. Regardless of cognitive subtype, MCI patients had more difficulties in daily functioning, as rated by knowledgeable informants, than cognitively normal elderly. More than one third of MCI patients have difficulty keeping appointments, finding their belongings, remembering current events, and following TV programs. About 20% reported difficulties driving and using transportation, managing their finances, organizing and completing activities, and even taking medications. Amnestic and non-amnestic MCI patients had similar levels of functional impairment. However, patients with cognitive impairments in more than one domain had more difficulties in daily activities than those with impairment in a single domain, possibly due to more widespread brain pathology.
These findings have important implications. The number of MCI informants reporting difficulties in multiple domains of daily functioning highlights the importance of a detailed assessment of functional abilities among these patients. Although the functional impairments in MCI are not as severe as in dementia, MCI patients still require assistance with more cognitively demanding daily activities. Second, since functional decline might be a harbinger of a dementing condition, we might predict that multiple-domain MCI patients are at higher risk of developing dementia than single-domain patients. This conclusion is consistent with several recent studies (Alexopoulos et al., 2006
; Tabert et al., 2006
Our second finding is that among three empirically validated executive function components (planning/problem solving, working memory, and judgment), only working memory was associated with ratings of daily functioning. Even then, the association was quite modest. When we controlled for other demographic, health-related and cognitive factors, working memory contributed unique variance on only one of the two measures, the IQCODE. We found that measures of global functioning and constructional praxis, the MMSE and the clock drawing test, were better predictors of ADL-PI, adding more unique variance than any other cognitive measure.
Most prior studies investigating the relationship of executive functions and performance on daily activities have not controlled for overall cognitive abilities (Cahn-Weiner et al., 2000
; Bell-McGinty et al., 2002
; Lewis and Miller, 2007
; Mitchell and Miller, 2008
). It might be argued that global cognitive measures, such as the MMSE, rely on several cognitive abilities and can therefore obscure the true influence of a specific cognitive domain (Peters and Pinto, 2008
). Similarly, the clock drawing test, although primarily a measure of constructional praxis and visuospatial skills, relies on a broader range of cognitive abilities, including semantic memory and executive functions (Lowery et al., 2003
). Thus, one reason why clock drawing contributes to everyday functioning may be that it requires planning, seriation, and other ostensibly executive skills (Lewis and Miller, 2007
Our results are not directly comparable with previous studies, since we used empirically derived executive function component scores rather than individual tests or composite scores with only face validity. We believe this approach to be a significant improvement over prior studies as it emphasizes the role of underlying neurocognitive deficits rather than test scores. Nevertheless, we suggest that the present findings are in general agreement with studies showing that Trail Making Test (part B) and D-KEFS Color-Word Interference are significant predictors of IADLs (Bell-McGinty et al., 2002
; Cahn-Weiner et al., 2002
; Schmitter-Edgecombe et al., 2009
). Both tasks load on our working memory component (Brandt et al., 2009
), which was found to predict functional outcome in this study. This component is, of course, not an unambiguous measure of working memory, but it captures the contributions of tests requiring multiple tracking, divided attention, and inhibitory control (see and also Brandt et al., 2009
The specific mechanisms whereby working memory affects everyday functioning remain incompletely specified. Multi-step tasks in everyday life clearly depend on temporary active maintenance of specific elements in working memory while other elements are being performed (Humphrey et al., 2001
). These elements include representation of goals, stimuli in the environment, response states, and response contingencies and production rules (e.g., if condition X, then perform action Y) (Baddeley, 1986
; Kimberg and Farah, 1993
). Kimberg and Farah (1993)
suggested that selective damage or ‘weakening of associations’ (p. 114) between these elements in working memory can disrupt the successful completion of a task. Moreover, three other functions necessary for everyday functioning depend on working memory: (a) the ability to ‘maintain the finer temporal details of the structure of a script’ (Sirigu et al., 1996
, p. 297), (b) the ability to monitor conflicts between the actual and the required sequence of actions (Shallice and Burgess, 1996
), and (c) the inhibition of environmental stimuli distraction and the ‘rejection of action alternatives presented by the stimulus situation’ (Zanini et al., 2002
, p. 88). Further studies on the pattern of difficulties and errors in everyday tasks are necessary to clarify how working memory affects everyday functioning in MCI patients.
Regarding the magnitude of the contribution of executive cognition to functional status, our results stand in contrast to those of Bell-McGinty et al. (2002)
who found up to 54% of variance in functioning explained by executive cognition. However, a recent meta-analysis of 68 studies in demented and non-demented individuals supports our findings; Royall et al. (2007)
found that the unique variance in functional outcome explained by executive function and other neurocognitive measures is small to modest (no more than 12%).
Our study has several limitations. First, our MCI sample is a non-random sample of convenience, with a higher proportion of men. Second, our selection of criteria for defining MCI may be questioned, since there is no universally accepted prescription for how the Petersen/Mayo criteria should be operationalized. However, we applied both clinical criteria (interview with an informant, yielding a CDR = 0.5), as well as psychometric criteria based on well-recognized neuropsychological procedures. Third, one might argue that proxy reports used to assess functional status are less valid than performance-based tests. The latter are, of course, much more labor-intensive, and performance of an activity while being examined is likely to be different than performance in everyday life. Finally, the differential contribution of the executive function domains to everyday functioning was examined to the MCI patients as a group due to small subgroup sample sizes. This warrants further investigation.
- Functional abilities are impaired in all MCI subtypes. Working memory impairment contributes significantly to IQCODE. General cognitive impairment accounts for more unique variance than executive impairment.