In this study, we examined the influence of executive function on memory performance in amnestic MCI based on the assumption that amnestic MCI is a heterogeneous population and that multiple factors may contribute to the observed memory impairment in these individuals. To that end, we examined the relationship between morphometric measures and cognition (i.e., memory and executive function). Results showed that executive function modulates some aspects of memory ability and that MCI individuals with poorer executive function demonstrate greater thinning in frontal and cingulate cortices relative to those with HEF. Further, thinning in these areas was related to memory performance beyond effects attributable to hippocampal atrophy.
We hypothesized that amnestic MCI individuals with HEF would show better performance on memory relative to amnestic MCI individuals with LEF, particularly during the learning phase when executive support is likely to be critical for the successful organization of the material to be encoded. As predicted, and consistent with previous studies in normal aging (Crawford et al. 2000
; Salthouse et al. 2003
), the MCI HEF group outperformed the MCI LEF group on the LM immediate recall, AVLT total learning, and LD recall. However, on other memory variables such as LM retention or SD recall of AVLT, the 2 MCI groups demonstrated comparable performance. The finding that greater differences occurred between the 2 MCI groups on learning components than on memory retention is consistent with literature regarding the differential roles of frontal and medial temporal lobes in memory, with the frontal lobe more associated with the acquisition of information, and the medial temporal lobe more associated with memory consolidation (Ferbinteanu et al. 2006
). Our findings also support the evidence that AVLT performance relies on executive function to a greater degree than does the LM. It has been suggested that differential performance between list learning and story recall measures reflects differences in the respective task demands on the encoding process (Tremont et al. 2000
). As opposed to the unstructured word list, the LM test is presented in a structured story format that provides contextual cues, thus potentially reducing the need to actively organize the information for learning and recall.
Regarding the relationship between morphometric variables and executive function in MCI, we hypothesized that MCI individuals with LEF would show reduced thickness in frontal and cingulate areas relative to MCI individuals with HEF. Results showed that the MCI group as a whole demonstrated significant gray matter thinning in bilateral medial and lateral temporal lobe as well as frontal lobe regions compared with the HC group, consistent with previous studies (Wolf et al. 2003
; Fennema-Notestine et al. 2009
). In support of our hypothesis, the results indicated that relative to the HC group, the MCI LEF group showed greater and more widespread atrophy than the HEF group. Although the 2 MCI groups showed comparable morphometry in bilateral mesial temporal lobe regions, the LEF group demonstrated greater cortical thinning in left orbitofrontal region, left frontal pole, right rostral middle frontal lobe, bilateral middle temporal lobe, and left inferior temporal lobe regions relative to the MCI HEF. Cortical thinning observed in the MCI LEF group in the frontal regions is consistent with previous studies that have highlighted the role of prefrontal regions in executive functioning (for a review, see Stuss and Alexander 2000
). However, the results regarding the difference between the 2 MCI groups in middle and inferior temporal lobe regions and their significant correlations with executive function were unexpected because middle and inferior temporal cortices have not traditionally been associated with executive function. These significant relationships, however, may reflect language or visual abilities required in the executive function tasks used in the current study.
Significantly decreased thickness was also found in bilateral PCC but not ACC in the MCI LEF group compared with the MCI HEF group, which was somewhat unexpected given that ACC is considered to be involved in executive function and is part of the frontal circuitry (Kobayashi and Amaral 2003
). Cortical thinning in PCC in the MCI LEF group compared with the MCI HEF group was not unexpected given that the PCC is considered part of the limbic system and has reciprocal connections with the medial temporal lobe, including entorhinal cortex and hippocampal formation as well as frontal areas, particularly BA 46, 9, 10, and 11 (Kobayashi and Amaral 2003
). Hypometabolism and volumetric reduction in PCC has been identified as a feature of early AD (Chua et al. 2008
; Choo et al. 2008
; Pengas et al. 2008
), and several recent studies have reported PCC hypometabolism or/and volume reduction in individuals with MCI (Choo et al 2008; Chua et al 2008; Fennema-Notestine et al. 2009; Pengas et al. 2008). Our findings support the notion that PCC abnormality can be detected in a preclinical stage of AD and suggest that atrophy in the PCC is associated with executive dysfunction. Overall, the results of cortical thinning in the frontal lobe and PCC region observed in the MCI LEF group compared with the MCI HEF group are consistent with prior studies (Bell-McGinty et al. 2005
; Whitwell et al. 2007
; Fennema-Notestine et al. 2009)
that examined volumetric differences between different MCI subtypes (i.e., MCI-amnestic vs. MCI-multiple cognitive domain subtype) using criteria proposed by Petersen et al. (2001)
In support of our final hypothesis, the results showed that thickness in numerous regions of frontal cortex and bilateral PCC were significantly associated with memory performance above and beyond the contribution of mesial temporal regions known to be associated with episodic memory. These relations remained significant when the analysis was restricted to the MCI groups, suggesting that they did not arise principally from the large difference between controls and MCI participants. These results correspond with the finding that MCI individuals with LEF showed lower memory performance and greater cortical thinning in frontal and PCC regions relative to those with HEF; despite equivalent hippocampal volumes and thickness of mesial temporal lobe regions (Supplementary Table 1
There are a number of interindividual differences that may account for the current executive function/memory findings including premorbid cognitive function, genetic factors, and diagnostic heterogeneity. It is unlikely that the relationship observed here between executive function and memory is an artifact of differences in general cognitive functioning because the HEF and LEF groups evidenced comparable CDR-SB scores as well as similar performance on many other neuropsychological measures, including MMSE, memory retention, TMT-A, animal fluency, and the visuoconstruction subscale of the ADAS-cog. Furthermore, the group effects on learning and recall measures remained significant after controlling for estimated VIQ and CDR-SB. Genetic factors, however, may have contributed to the observed effects. APOE ϵ4 has been documented as a genetic risk factor for late-onset AD (Bennett et al. 2003
) with some studies suggesting that the APOE ϵ4 genotype is associated with subtle impairments in executive functions such as working memory, inhibition, and attentional switching in healthy adults and prodromal dementia (Greenwood et al. 2000
; Small et al. 2004
; Jacobson et al. 2005
; Wetter et al. 2005
; McQueen et al. 2007
). Consistent with this, we found that both MCI groups had a higher frequency of APOE ϵ4 carriers than the HC group and that the LEF group had a higher frequency of APOE ϵ4 carriers than the HEF group. Diagnostic heterogeneity may also have contributed to the relationship between executive function and memory. MCI groups with more widespread gray matter loss at baseline have been shown to progress more rapidly to AD relative to those with focal gray matter loss (Whitwell et al. 2008
; McEvoy et al. 2009
). Evidence from neuropsychological data shows a parallel indication: Executive dysfunction in addition to memory decline improves prediction of conversion to AD in individuals with MCI (Albert et al. 2001
; Chen et al. 2001
; Backman et al. 2005
; Rapp and Reischies 2005
; Belleville et al. 2007
; Blacker et al. 2007
). Consistent with this, we found that the 1-year conversion rate to a clinical diagnosis of probable AD was higher in the MCI LEF group than the MCI HEF group. Thus, the 2 MCI groups in the present study may represent a continuum of the MCI-AD course with the MCI LEF group having progressed farther toward AD than the MCI HEF group, or they may represent different underlying etiologies. For example, the MCI LEF group may show more vascular or mixed pathologies than the MCI HEF group. Future analysis of longitudinal change in brain structure and neuropsychological task performance over ADNI's 3-year follow-up period, once sufficient data have been collected and processed, may help elucidate this issue.
Potential limitations of the present study include the restricted number of executive measures and nonverbal memory tasks available for analyses. Executive function is a broad cognitive domain that includes many subcomponents such as working memory, planning, inhibition, problem solving, and mental set shifting capacities (Stuss and Alexander 2000
). With the limited number of executive function tests available in the ADNI, it is not possible to determine the extent to which each of these components relates to memory performance. Additionally, the present study is unable to provide information about the relation between nonverbal memory and executive function measures. Finally, the relationships observed between cognition and morphometric measures in frontal and temporal lobe areas were low (r
0.12–0.22), suggesting that much of the variance in cognitive scores is not explained by brain morphometry. It is likely that the above-mentioned factors related to interindividual differences, such as premorbid cognitive function, genetic factors, and diagnostic heterogeneity, may contribute more to the variance in executive functioning.
Despite these limitations, the present study suggests that executive function influences verbal memory ability in amnestic MCI, particularly with regard to learning components of memory. This influence is not related to variation in hippocampal volume but reflects cortical thinning in bilateral prefrontal lobe areas and the PCC. These results demonstrate the importance of evaluating executive function in individuals with amnestic MCI to predict involvement of brain areas beyond the mesial temporal lobe. Future follow-up analyses will provide information on the ability of these measures to predict decline in MCI.