Cortical Thickness and Microstructural White Matter Changes Detect Amnestic Mild Cognitive Impairment

Abstract

Both the apolipoprotein E (APOE) ɛ4 allele and amnestic mild cognitive impairment (aMCI) are considered to be risk factors for Alzheimer’s disease (AD). The primary aim of this study was to determine whether the aMCI-related abnormality in gray matter (GM) cortical thickness and white matter (WM) tracts integrity would be modified by the APOE genotype. A total of 146 older adults, including 64 aMCI patients (28 ɛ4 carriers and 36 non-carriers) and 82 healthy controls (39 ɛ4 carriers and 43 non-carriers), underwent a standardized clinical interview, neuropsychological battery assessment, and multi-modal brain magnetic resonance imaging scans. Compared with control subjects, the patients with aMCI showed significantly reduced cortical thickness bilaterally in the parahippocampal gyrus and disrupted WM integrity in the limbic tracts (e.g., increased mean diffusivity in the right parahippocampal cingulum and bilateral uncinate fasciculus). However, no significant main effects of the APOE genotype and diagnosis-by-genotype interaction on GM thickness and WM integrity were observed. Further, diffusivity measures of the limbic WM tracts were significantly correlated with the parahippocampal atrophy in aMCI. Importantly, the parahippocampal thickness and diffusivity measures of the limbic WM tracts were significantly correlated with the cognitive performance (i.e., episodic memory Z score) in patients with aMCI. These results demonstrate that WM microstructural disruptions in the limbic tracts are present at the early stage of AD in an APOE-independent manner; and this degeneration may occur progressively, in parallel with parahippocampal atrophy, and may specifically contribute to early initial impairment in episodic memory.

Keywords

Alzheimer’s disease amnestic mild cognitive impairment apolipoprotein E cortical thickness white matter integrity

INTRODUCTION

Alzheimer’s disease (AD) is a neurodegenerative disorder characterized by a progressive impairment of cognitive function and behavior. Amnestic mild cognitive impairment (aMCI) is the most well-characterized cognitive risk factor for AD and is considered to be a transitional state between normal aging and dementia with rates of progression to AD reported to range between 16% and 40% [1, 2]. Thus, aMCI provides an important model to study the mechanisms of AD. Nonetheless, the outcome for any subjects with aMCI is highly variable because many subjects may remain stable or even revert to a normal state, while others progress to dementia, possibly due to the interplay of physiologic, environmental, and genetic factors [3, 4].

The apolipoprotein E (APOE) ɛ4 allele is the most common susceptibility gene for AD [5] and influences the course of disease by increasing the risk for developing AD and lowering the age at disease onset [6, 7]. Case-control studies have shown that the ɛ4 allele is overrepresented in aMCI, and increased frequency of the allele was determined to be the strongest predictor of clinical progression from aMCI to AD [8 –10]. The role of APOE-ɛ4 allele in aMCI and conversion to AD has also been the subject of longitudinal studies. However, these studies have been inconsistent. Although some studies in this field have shown that APOE-ɛ4 allele leads to greater longitudinal metabolic decline in aMCI subjects converting to AD [11, 12], other studies have observed that cognitive decline and progression to AD is not strongly related to APOE-ɛ4 allele [13, 14]. Therefore, there remain large discrepancies in the literature concerning the role of APOE-ɛ4 allele in the progression from aMCI to AD.

It is becoming increasingly clearer that to achieve a comprehensive picture of AD evolution, investigations need to be based on both neuropsychological and anatomical information from subjects at high-risk of AD. Previous neuroimaging studies on AD/aMCI typically focused on investigating cortical structural changes in medial temporal grey matter (GM) such as volume reduction in the entorhinal cortex and hippocampus [15 –17]. Dickerson et al. [18] also identified several cortical regions in which thickness is a potential biomarker for AD, referred to as “AD-signature” cortical thickness. This measure has predicted symptom severity, amyloid binding in asymptomatic older adults, and conversion to AD among aMCI subjects [19]. Another technique, diffusion tensor imaging (DTI), is capable of analyzing white matter (WM) integrity by using measures such as factional anisotropy (FA) and mean diffusivity (MD). Previous DTI studies on AD/aMCI have demonstrated WM changes, especially in the fiber tracts connecting to the medial temporal cortex, such as cingulum bundle and uncinate fasciculus [20 –22]. Furthermore, there is also evidence that GM volume or WM integrity might be affected by the APOE genotype. For example, some [23 –26], but not all [27 –29], structural imaging studies of cognitively intact ɛ4-carrying adults have shown reduced brain volume (e.g., hippocampal volume) in the ɛ4 carriers. Similar DTI studies in healthy subjects also showed ɛ4-related WM integrity changes (e.g., decreased FA value) in the corpus callosum, fronto-occipital fasciculus, and parahippocampal cingulum [26 , 30–32]. Overall, these above studies suggested the association of the presence of aMCI or APOE-ɛ4 status with the alterations of GM and WM integrity. However, the independent contributions of cognitive and genetic risk factors (i.e., aMCI and APOE-ɛ4 allele) for AD have received considerable attention, but less is known about the possible interactive effects.

The primary aim of this study was to determine whether the aMCI-related abnormality in GM cortical thickness and WM tract integrity would be modified by the APOE genotype. Second, several previous studies showed a significant association between hippocampal atrophy and damage to the cingulum bundle (e.g., decreased FA in the cingulum) [33] as well as between corpus callosum FA and atrophy of connected cortical areas in patients with AD [34]. Therefore, in an attempt to further explore the mechanisms associated with WM damage in the early phase of AD (i.e., aMCI), we sought to examine whether WM damage is related to GM atrophy (e.g., cortical thickness reduction). Third, we tried to clarify whether the GM and WM damage is able to explain the impairment of specific cognitive domains (e.g., episodic memory) in patients with aMCI. Finally, we conducted an exploratory classification analysis to explore the feasibility of integrating neuropsychological and neuroimaging measures to accurately distinguish patients with aMCI from healthy controls.

MATERIALS AND METHODS

Participants

Patients with aMCI and healthy controls (HCs) were recruited to establish a registry at the Affiliated Zhongda Hospital, Southeast University [35, 36]. All participants were recruited through a normal community health screening and newspaper advertisements, and they underwent a standardized clinical interview, neuropsychological battery assessment, genetic screening, and multi-modal brain MRI examinations (for details, see the below). The detailed inclusion and exclusion criteria are described in the Supplementary Material. The study was approved by the Research Ethics Committee of ZhongDa Hospital Affiliated to Southeast University, and written informed consent was obtained from all participants.

In the original registry, the frequencies of the ɛ2, ɛ3, and ɛ4 allele in the HCs (8.3%, 80.9%, and 10.8%) and the aMCI patients (6.9%, 70.7%, and 22.4%) were similar to results of previous in Chinese population [37, 38]. In this present study, we sought to investigate the main effects of diagnosis and APOE genotype and diagnosis-by-genotype interactions on GM cortical thickness and WM tracts integrity. Therefore, we selected comparable number of participants in each group to the MRI study. This experimental design has been widely used in the previous APOE ɛ4-related studies [36, 39]. 146 elderly participants (age range 60–80 years), including 64 patients with aMCI and 82 HCs, were Chinese Han and right-handed. Of these aMCI patients, there were 28 ɛ4 carriers (21 ɛ3/ɛ4 and 7 ɛ4/ɛ4) and 36 non-carriers (ɛ3/ɛ3); of these controls, there were 39 ɛ4 carriers (37 ɛ3/ɛ4 and 2 ɛ4/ɛ4) and 43 non-carriers (ɛ3/ɛ3). The participants with one or more ɛ2 allele(s) were excluded from this study due to the allele’s possible protective effect [40, 41]. Table 1 presents the demographic information, APOE status, and cognitive scores of the participants included in this study.

Neuropsychological assessment

For all participants, we assessed their general cognitive function using the Mini-Mental State Examination (MMSE) and Mattis Dementia Rating Scale-2 (MDRS-2), and performed a neuropsychological battery to evaluate their specific functions in episodic memory, visuospatial skills, information processing speed, and executive function, respectively. This battery consisted of the auditory verbal learning test with a 20-min delayed recall, the logical memory test with a 20-min delayed recall, the Rey-Osterrieth complex figure test with a 20-min delayed recall, the clock drawing test, the digital symbol substitution test, trail-making test A and B, the Stroop color-word test A, B, and C, the verbal fluency test, the digital span test, and the semantic similarity test.

In this study, a composite score analysis [36] of these neurocognitive measures was further conducted to increase statistical power by reducing random variability and floor and ceiling effects. For each subject, the raw scores from each test were first transformed to z scores with reference to the means and standard deviations of the test across all subjects. Then, the composite scores were calculated by averaging the z scores within the neuropsychological domains listed below: episodic memory (three tests, including the auditory verbal learning test with a 20-min delayed recall, the logical memory test with a 20-min delayed recall, and the Rey-Osterrieth complex figure test with a 20-min delayed recall), visuospatial function (two tests, including the Rey-Osterrieth complex figure test and the clock drawing test), information processing speed (four tests, including the digital symbol substitution test, the trail making test-A, and Stroop A and B), and executive function (five tests, including the verbal fluency test, the digital span test-backward, the trail making test-B, Stroop C, and the semantic similarity test). Notably, MMSE and MDRS-2 were used for descriptive and diagnostic classifications, but not for the composite measures.

APOE genotyping

Genomic DNA of each subject was extracted from 250 μl EDTA-anticoagulated blood using a DNA direct kit (Tiangen, China). A polymerase chain reaction-based restriction fragment length polymorphism (PCR-RFLP) assay was applied to detect the alleles of rs7412 and rs429358, the haplotype of which ultimately determined the APOE genotype. The specific process is described in the Supplementary Material.

Data acquisition

MR images were acquired in a 3.0 T Siemens Verio scanner (Siemens, Erlangen, Germany) with a 12-channel head coil at the ZhongDa Hospital affiliated to Southeast University. All participants lay supine with the head snugly fixed by a belt and foam pads to minimize head movement. High-resolution T1-weighted axial images covering the whole-brain were acquired using 3D magnetization prepared rapid gradient echo sequence as below: repetition time (TR) = 1900 ms; echo time (TE) = 2.48 ms; flip angle (FA) = 9°; acquired matrix = 256×256; field of view (FOV) = 250 mm×250 mm; thickness = 1.0 mm; gap = 0 mm; number of slices = 176. Diffusion-weighted imaging was acquired with single-shot echo-planar imaging sequence in alignment with the anterior-posterior commissural plane. The diffusion sensitizing gradients were applied along 30 non-collinear directions (b = 1000 s/mm²), together with an acquisition without diffusion weighting (b = 0). Seventy contiguous axial slices were acquired with a slice thickness of 2 mm and no gap. The acquisition parameters were as follows: TR = 10,000 ms; TE = 90 ms; FA = 90°; acquisition matrix = 128×128; FOV = 256 mm×256 mm; number of excitations = 2.

Data preprocessing

Cortical thickness mapping

We used the CIVET pipeline to measure thickness on the cortical surface as previously described [42]. Briefly, the native T1-weighted MR images were first linearly aligned into the stereotaxic space and corrected for nonuniformity artifacts using the N3 algorithm [43]. The registered and corrected images were then automatically segmented into GM, WM, cerebrospinal fluid, and background by using a partial volume classification algorithm [44]. Next, the inner and outer GM surfaces were automatically extracted from each hemisphere using the Constrained Laplacian Automated Segmentation using Proximities (CLASP) algorithm [45, 46]. Cortical thickness was further measured as the distance between corresponding vertices of inner and outer surfaces of GM across 40,962 vertices in each hemisphere [45]. Finally, the thickness data was blurred using a surface-based diffusion smoothing kernel of 30 mm FWHM that preserves corticaltopology [47].

DTI imaging analysis

The DTI data were preprocessed and analyzed using the FSL package (http://www.fmrib.ox.ac.uk/fsl/) [48] and the PANDA software (http://www.nitrc.org/projects/panda/) [49]. Briefly, preprocessing included correction for motion and eddy current effects in DTI images. Subsequently, FMRIB’s Diffusion Toolbox was used to fit the tensor model and to compute the fractional anisotropy (FA), mean diffusivity (MD), axial diffusivity (DA), and radial diffusivity (DR) maps. After that, the TBSS analysis [50] was utilized. First, all subjects’ FA images were skull-stripped and spatially registered to a target image (FMRIB58_FA) in standard space using a nonlinear transformation tool FNIRT. The mean FA image was thresholded at 0.2 and skeletonized to contain only the centers of major WM tracts in standard space. Then, the spatially normalized FA, MD, DA, and DR images for each subject were projected onto the skeleton using the derived registration and projection parameters. Finally, in order to investigate the diffusion changes in several specific tracts, the digital WM atlas JHU ICBM-DTI-81 (see http://cmrm.med.jhmi.edu/) was used to parcel the entire WM into 20 tracks of interest (TOIs) (Supplementary Figure 1). Then, the mean values of each DTI measures (i.e., FA, MD, DA, and DR) for each TOI were extracted for all subjects.

Statistical analysis

Demographic and neuropsychological variables

Statistical analyses of demographics and cognitive scores were performed using two-way analysis of covariance (ANCOVA) for continuous variables and using chi-square tests for categorical variables. Specifically, the main effects of diagnosis (i.e., aMCI versus HC) and APOE genotype (i.e., ɛ4 carriers versus non-carriers), and diagnosis-by-genotype interactions were assessed. For the ANCOVA analyses in the cognitive variables, age, gender and education were considered as covariate variables. These analyses were implemented in SPSS 17.0 (SPSS, Inc., Chicago, IL).

Group differences in cortical thickness and DTI indices

We first assessed the main effects of diagnosis and APOE genotype, and diagnosis-by-genotype interactions on cortical thickness. Specifically, a general linear model with “diagnosis”, “APOE genotype”, and “diagnosis-by-genotype” as predictor variables was applied across the entire cerebral cortex, wherein age, gender, and years of education were included as covariates and cortical clusters with an FWE-corrected p < 0.05 were considered as significant. This analysis was implemented using the SurfStat (http://www.math.mcgill.ca/keith/surfstat/). Then, to investigate the group-differences in DTI measures on WM TOIs, tract-specific ANCOVAs were used to examine the main effects of diagnosis and APOE genotype and diagnosis-by-genotype interactions, after adjusting for age, gender, andeducation.

Relationships of cortical thickness, DTI indices, and cognitive performances

When the bilateral parahippocampal gyrus (PHG) showed significant main effects of diagnosis on the thickness (see results), we further tested the associations between bilateral PHG thickness and the DTI indices of the identified WM TOIs (as determined from the ANCOVA, see results) using the Pearson’s correlation test among the aMCI patients. These analyses were applied to the aMCI patients to understand the relationship between the damage classically attributed in AD (e.g., hippocampal/parahippocampal gyrus atrophy), the deterioration of the WM tracts with the disease, and to show whether these two phenomena can be potential attributed to a linked disease pathway. Finally, multiple linear regression analyses were conducted to calculate the correlations between the neuroimaging measures (i.e., cortical thickness and DTI indices) and cognitive performance in the aMCI and HC groups separately, after controlling for age, gender, education, and APOE genotype. Bonferroni correction for multiple comparisons was used with the significant level considered at p < 0.0125 (p = 0.05/4 composite scores). To statistically compute the differences in magnitudes of the correlations between the neuroimaging measures and cognitive performance in the aMCI and HC groups, correlation coefficients obtained above were further converted into z values by using Fisher’s r-to-z transform. This transformation generated values that were approximately normally distributed. A Z statistic was then used to compare these transformed z values to determine the significance of the between-group differences in correlations, and Cohen’s q was used to quantify the magnitude of difference between correlations; according to |Cohen’s q| <0.20 is considered a small effect, 0.30 a moderate effect, and 0.50 a largeeffect [51].

Exploratory classification analysis

To measure the potential of neuropsychological and neuroimaging measures for future use in diagnosis, the combination of targeted cognitive performance, cortical thickness and DTI indices were tested to see if they could be used as a feature that could separate aMCI patients from normal controls. We took the cognitive measures, mean cortical thickness of the identified regions, and DTI indices of identified WM TOIs for each subject in the aMCI and HC groups as features in the stepwise discrimination analysis. To test the robustness of the results, we also validated the results by using the leave-one-out cross-validation method. This analysis was implemented in SPSS 17.0 (SPSS, Inc., Chicago, IL).

RESULTS

Demographic and neuropsychological variables

Table 1 illustrates the demographics and neuropsychological measures for aMCI and HC subjects stratified by APOE-ɛ4 status. The four subgroups did not differ in age, gender and education (all ps> 0.05). Two-way ANCOVA analyses revealed the main effects of diagnosis and APOE genotype and the diagnosis-by-genotype interactions on cognitive measures. A significant main effect of diagnosis on each cognitive domain was observed, with the aMCI group showing worse cognitive performance than the HC group. There was no significant main effect of APOE genotype on any cognitive measure. In addition, we found a significant diagnosis-by-genotype interaction only on the visuospatial function, with ɛ4 carriers showing worse performance than non-carriers in the aMCI group but no genotype difference in the HC group.

Group-based differences in cortical thickness and DTI indices

Cortical thickness

The main effects of diagnosis on GM cortical thickness are illustrated in Fig. 1. There were two significant cortical clusters (Fig. 1A), which were located around the bilateral parahippocampal cortex (FWE-corrected p < 0.05). The bar chart revealed a thinner thickness of the aMCI patients compared with the control subjects (Fig. 1B). No regions showed significant main effect of APOE genotype or diagnosis-by-genotype interaction.

DTI measures

The main effects of diagnosis on DTI indices are showed in Fig. 2. TOI analyses revealed increased MD in the WM tracts connecting to the medial temporal lobe in the aMCI patients comparted with the control subjects, including the right parahippocampal cingulum (PHC) and uncinate fasciculus (UF) tracts (p < 0.05). Meanwhile, increased DR and DA were found only in the right UF in the aMCI patients compared with the controls (p < 0.01). Further, DA of the right PHC and left UF were increased in the aMCI patients relative to the control subjects (p < 0.05). However, there were no significant tract-specific FA differences between the aMCI patients and control subjects. In addition, no significant main effect of APOE genotype or diagnosis-by-genotype interaction was observed on the mean diffusion metrics of each WM TOI (all ps> 0.05).

Relationship between WM abnormalities and GM atrophy

We further examined the correlation between DTI-measured WM changes and cortical thickness analysis of GM changes in the patients with aMCI. Significant correlations between changes in DTI measures and GM atrophy are shown in Table 2. In the identified WM TOIs, such as right PHC and bilateral UF, we observed that the mean diffusion metrics (i.e., MD, DA, and DR indices) correlated negatively with the bilateral PHG thickness in the aMCI group (all ps < 0.005).

Relationship between cognitive performance and neuroimaging measures

We finally examined the relationships between the cognitive composite Z scores (i.e., episodic memory, visuospatial function, information processing speed, and executive function) and neuroimaging measures (i.e., cortical thickness and DTI indices) in the aMCI and HC groups. In the aMCI group, the bilateral PHG thickness positively correlated with the episodic memory Z scores (Table 3, Supplementary Figure 2, all ps< 0.001). In addition, MD, DA, and DR indices were negatively correlated with the episodic memory and executive function Z scores in the right PHC and UF tracts for the patients with aMCI (Table 3, Supplementary Figure 3, all ps< 0.005). However, in the control subjects, the neuroimaging measures (i.e., cortical thickness and DTI indices) were not correlated significantly with any cognitive performance. Importantly, statistical comparison of the correlations using Z statistic further indicated that the associations were significantly different between the aMCI and HC groups (Table 3, all ps< 0.05, |Cohen’s q| >0.30).

Exploratory classification analysis

We selected cognitive composite scores and neuroimaging measures as independent variables for stepwise discrimination analysis in our model selection. Considering the relatively small number of subjects included in this study, we used the cognitive and neuroimaging measures with significant aMCI versus HC group differences as candidate variables. Therefore, four cognitive measures (episodic memory, visuospatial function, information processing speed, and executive function Z scores) and eight neuroimaging measures (bilateral PHG thickness, right PHC, MD, and DA indices, right UF MD, DA, and DR indices, and left UF DA index) entered in the model. Using this composite model, we could correctly distinguish the aMCI patients from the control subjects in 86.3% of the cases (sensitivity, 82.8%; specificity, 89.0%) (Table 4). The leave-one-out cross-validation results also showed that 85.6% of subjects could be correctly classified between the two groups (Table 4). Taking each subject’s discriminative score as a threshold, the performance in terms of the receiver operating characteristic (ROC) curve was shown in Fig. 3. The area under the ROC curve (AUC) of the proposed method was 0.95, indicating an excellent diagnostic power. Further analysis revealed that the most discriminative feature for classification included the episodic memory and information processing speed Z score, bilateral PHG thickness, and right UF MD index.

DISCUSSION

Neuropathological investigations suggest that the degenerative progress occurring in AD is initially localized in the medial temporal lobe (e.g., entorhinal and hippocampal cortex), and then extend to other neocortical regions [52]. This succession of pathophysiological events is thought to underlie the corresponding progressive accumulation of cognitive deficits [53]. In this perspective, we investigated the abnormalities of GM and WM in patients with aMCI, which represent the earliest stage of AD pathology. This study yielded four principal findings, which we will describe in turn. First, we found that patients with aMCI displayed significantly reduced GM thickness bilaterally in the PHG and abnormal WM integrity in the limbic tracts (i.e., right PHC and bilateral UF) in an APOE-independent manner. Second, significant and anatomically congruent correlations between parahippocampal atrophy and diffusivity measures of limbic WM tracts were observed in aMCI patients, suggesting the WM disruption of limbic tract is secondary to cortical neuronal damage. Third, the PHG thickness and diffusivity measures of limbic WM tracts (i.e., right PHC and UF) were significantly related to the cognitive performance (i.e., episodic memory Z score) in patients with aMCI. Finally, the combination of neuropsychological and neuroimaging measures could accurately distinguish patients with aMCI from healthy controls in 86.3% of the cases; the most discriminative features for classification involved the episodic memory and information processing speed Z score, bilateral PHG thickness, and right UF MD index.

aMCI-related decline in GM thickness and WM integrity

Cortical thickness of the medial temporal lobe was most severely reduced in patients with aMCI, and within the medial temporal lobe, the parahippocampal gyrus was most affected. This is consistent with previous MRI and histopathological studies showing that this region of the brain is affected early and profoundly in the course of AD [54 –56]. In addition, the present study further demonstrated the microstructural changes of the limbic tracts (including PHC and UF) as manifested in higher MD, DA, and DR among the patients with aMCI. According to previous animal model studies, these changes are likely related to demyelination and axonal loss [57 –59]. Interestingly, the limbic WM degeneration has been previously documented in AD/aMCI, indicating that decreased WM integrity in the limbic tracts is a highly replicable finding across DTI studies of AD/aMCI [22 , 60–62]. More importantly, further correlation analyses revealed that PHG thickness and DTI indices in the WM limbic tracts were significantly correlated with the composite episodic memory Z scores across patients with aMCI. Therefore, these findings suggest that alterations within the medial temporal cortex and limbic tracts may serve as a potential imaging biomarker of early AD-related brain changes.

In the present study, the ANCOVA analyses revealed no significant main effect of APOE genotype and diagnosis-by-genotype interaction on GM thickness or WM integrity. Many previous neuroimaging studies have suggested an age-dependent effect of APOE ɛ4 on GM and WM integrity. For example, using region-of-interest or voxel-based morphometric analysis, some previous studies have shown an association between cerebral morphometry and APOE genotype in healthy young and middle-aged individuals (<65 years) [63 –66] but not in healthy elderly subjects (≥65 years) [28, 29]. Likewise, using whole-brain DTI parameters, Heise et al. [67] compared the WM integrity for younger (mean age 28.6 years±4.20) and older (mean age 64.9±7.19) APOE ɛ4 carriers and non-carriers. These authors observed widespread decreases in FA values for younger but not older APOE ɛ4 carriers relative to non-carriers. Consistent with these GM/WM studies, we also revealed no association between APOE genotype and GM/WM integrity in the healthy elderly subjects (mean age 70.1 years±6.03). One plausible explanation for the lack of APOE effect in the later life is the different demographics of the nondemented subjects. It is widely accepted that inter-subjects anatomic variability increases with advancing age, especially in older subjects [29]. Therefore, it is possible that an ɛ4 effect on GM/WM integrity was present in our sample but was small enough to have been overwhelmed by the much greater effect of inter-subjects’ anatomic variability.

Relationship between WM abnormality and GM atrophy

The current study allowed us to directly correlate changes between GM and WM in patients with aMCI. DTI measures in the right PHC and bilateral UF were found to correlate with the PHG thickness across aMCI patients. In agreement, correlations between hippocampal volume and WM limbic tract (e.g., cingulum and UF) FA have been previously reported in AD patients [20 , 68] or a combination of aMCI and AD patients [69]. Further studies in aMCI showed that the atrophy rate of the cingulum and UF was found to correlate with baseline hippocampal GM atrophy [70], suggesting that both WM and GM degenerative processes are not independent from each other over the course of AD development. Our correlation results in patients with aMCI fully support these findings. It is demonstrated that the UF connects the anterior part of medial temporal lobe (i.e., entorhinal and perirhinal cortices) with the orbital and polar frontal cortex [71, 72]; importantly, the entorhinal cortex is the first site of neuronal death in AD. Therefore damage in the UF could be tightly related with medial temporal GM atrophy and may result in memory impairment [70, 73]. Regarding the PHC, it connects the hippocampus with the posterior cingulate cortex (PCC) [74]. Hypometabolism in the PCC has been consistently reported in patients with aMCI [75]. Further studies indicated that PCC activity during memory tasks is correlated with hippocampal volume [76, 77] and that de-afferentation of the PCC may contribute to its dysfunction from the early-AD stage (e.g., aMCI) [70, 78]. Taken together, our present finding is consistent with the classic secondary degeneration model of AD [33, 68], and further suggests that the observed WM tract changes in patients with aMCI may be a consequence of neuronal loss in the medial temporal cortex (e.g., hippocampal and parahippocampal gyrus). However, a direct effect of amyloid-β on WM tissue cannot be excluded [79].

Exploratory classification analysis

Finally, in this present study, we selected cognitive composite scores and neuroimaging measures as independent variables for stepwise discrimination analysis in our model selection. Interestingly, this composite model could accurately distinguish patients with aMCI from healthy controls in 86.3% of the cases (sensitivity, 82.8%; specificity, 89.0%); and the most discriminative features for classification involved the episodic memory and information processing speed Z score, bilateral PHG thickness, and right UF MD index. Thus, this finding indicates that cortical thinning in the PHG, damage in the UF, a combination of episodic memory and information processing speed deficits may provide complementary information for the diagnostic classification. More generally, these finding reinforce the view that the detection of prodromal AD may benefit from the combination of targeted neuropsychological and neuroimaging markers [80, 81].

Limitations

There were several limitations in this study. First, it lacked amyloid imaging and histopathological data suggestive of AD pathology in patients with aMCI, thereby relying on clinical diagnosis only. Therefore, a considerable amount of clinical and biological heterogeneity existed in the present sample of MCI subjects. Notably, the present study only recruited aMCI subjects, who are considered to have a high risk to convert AD dementia. Importantly, as shown in the Results section, the aMCI group in our study showed significantly reduced cortical thickness in the bilateral parahippocampal gyrus compared with healthy controls. As the new guidelines for the AD-related diagnosis [82, 83] suggested the hippocampal or parahippocampal gyrus atrophy as a neurodegeneration biomarker, the aMCI subjects recruited in this study may likely for a relatively homogeneous group. Second, further validation on a larger sample is required to understand the present results, and longitudinal studies should be conducted to clarify the progression of brain changes over time. Finally, regarding the exploratory classification analysis, one might argue that the use of neuropsychological markers in this model is a tautology because they were used to diagnose the aMCI. However, this may not be correct because the diagnosis of aMCI already requires episodic memory impairment at a level consistent with AD itself [80]. Notably, use of cognitive markers has certain advantages: clear and significant effect on odds ratios, objectivity in scoring, comparative economy in terms of expense and time, and reliability[80, 84].

CONCLUSIONS

In summary, this present study reveals that WM integrity in the limbic tract (i.e., right PHC and bilateral UF) is disrupted from the early stage of AD (i.e., aMCI) in an APOE-independent manner. This degeneration seems to be closely related with, and possibly a consequence of, medial temporal GM atrophy. Additionally, this study further supports the view that the diagnosis of aMCI might benefit from the combined use of cognitive testing and brain imaging.

Footnotes

ACKNOWLEDGMENTS

We thank all the patients and volunteers for participating in this study. We would like to thank Dr. Alan Evans for kindly providing the CIVET software. This study was supported by the Projects of International Cooperation and Exchanges NSFC (grant number 81420108012), the National Key Basic Research Program of China (grant number 2014CB846102), the Natural Science Foundation of China (grant number 81601559), the Guangdong Provincial Natural Science Foundation of China (grant number 2016A030310233), and the Key Program for Clinical Medicine and Science and Technology: Jiangsu Province Clinical Medical Research Center (grant numbers BL2013025, BL2014077).

Authors’ disclosures available online ().

Appendix

References

Marra

, Ferraccioli

, Vita

, Quaranta

, Gainotti

(2011) Patterns of cognitive decline and rates of conversion to dementia in patients with degenerative and vascular forms of MCI. Curr Alzheimer Res 8, 24–31.

Wierenga

, Dev

, Shin

, Clark

, Bangen

, Jak

, Rissman

, Liu

, Salmon

, Bondi

(2012) Effect of mild cognitive impairment and APOE genotype on resting cerebral blood flow and its association with cognition. J Cereb Blood Flow Metab 32, 1589–1599.

Collie

, Maruff

(2000) The neuropsychology of preclinical Alzheimer’s disease and mild cognitive impairment. Neurosci Biobehav Rev 24, 365–374.

Petersen

(2000) Mild cognitive impairment: Transition between aging and Alzheimer’s disease. Neurologia 15, 93–101.

Bertram

, Lill

, Tanzi

(2010) The genetics of Alzheimer disease: Back to the future. Neuron 68, 270–281.

Corder

, Saunders

, Strittmatter

, Schmechel

, Gaskell

, Small

, Roses

, Haines

, Pericak-Vance

(1993) Gene dose of apolipoprotein E type 4 allele and the risk of Alzheimer’s disease in late onset families. Science 261, 921–923.

Strittmatter

, Roses

(1995) Apolipoprotein E and Alzheimer disease. Proc Natl Acad Sci U S A 92, 4725–4727.

Petersen

, Smith

, Ivnik

, Tangalos

, Schaid

, Thibodeau

, Kokmen

, Waring

, Kurland

(1995) Apolipoprotein E status as a predictor of the development of Alzheimer’s disease in memory-impaired individuals. JAMA 273, 1274–1278.

Tierney

, Szalai

, Snow

, Fisher

, Nores

, Nadon

, Dunn

, St George-Hyslop

(1996) Prediction of probable Alzheimer’s disease in memory-impaired patients: A prospective longitudinal study. Neurology 46, 661–665.

10.

DeCarli

, Miller

, Swan

, Reed

, Wolf

, Carmelli

(2001) Cerebrovascular and brain morphologic correlates of mild cognitive impairment in the National Heart, Lung, and Blood Institute Twin Study. Arch Neurol 58, 643–647.

11.

Hsiung

G-YR

, Sadovnick

, Feldman

(2004) Apolipoprotein E ɛ4 genotype as a risk factor for cognitive decline and dementia: Data from the Canadian Study of Health and Aging. CMAJ 171, 863–867.

12.

Mosconi

, Perani

, Sorbi

, Herholz

, Nacmias

, Holthoff

, Salmon

, Baron

, De Cristofaro

, Padovani

, Borroni

, Franceschi

, Bracco

, Pupi

(2004) MCI conversion to dementia and the APOE genotype: A prediction study with FDG-PET. Neurology 63, 2332–2340.

13.

Barabash

, Marcos

, Ancin

, Vazquez-Alvarez

, de Ugarte

, Gil

, Fernández

, Encinas

, López-Ibor

, Cabranes

(2009) APOE, ACT and CHRNA7 genes in the conversion from amnestic mild cognitive impairment to Alzheimer’s disease. Neurobiol Aging 30, 1254–1264.

14.

Fleisher

, Sun

, Taylor

, Ward

, Gamst

, Petersen

, Jack

, Aisen

, Thal

(2008) Volumetric MRI vs clinical predictors of Alzheimer disease in mild cognitive impairment. Neurology 70, 191–199.

15.

Wisse

, Reijmer

, ter Telgte

, Kuijf

, Leemans

, Luijten

, Koek

, Geerlings

, Biessels

, Utrecht Vascular Cognitive Impairment (VCI) Study Group (2015) Hippocampal disconnection in early Alzheimer’s disease: A 7 tesla MRI study. J Alzheimers Dis 45, 1247–1256.

16.

Stoub

, Bulgakova

, Wilson

, Bennett

, Leurgans

, Wuu

, Turner

(2004) MRI-derived entorhinal volume is a good predictor of conversion from MCI to AD. Neurobiol Aging 25, 1197–1203.

17.

Pennanen

, Kivipelto

, Tuomainen

, Hartikainen

, Hänninen

, Laakso

, Hallikainen

, Vanhanen

, Nissinen

, Helkala

E-L

(2004) Hippocampus and entorhinal cortex in mild cognitive impairment and early AD. Neurobiol Aging 25, 303–310.

18.

Dickerson

, Bakkour

, Salat

, Feczko

, Pacheco

, Greve

, Grodstein

, Wright

, Blacker

, Rosas

(2009) The cortical signature of Alzheimer’s disease: Regionally specific cortical thinning relates to symptom severity in very mild to mild AD dementia and is detectable in asymptomatic amyloid-positive individuals. Cereb Cortex 19, 497–510.

19.

Bakkour

, Morris

, Dickerson

(2009) The cortical signature of prodromal AD: Regional thinning predicts mild AD dementia. Neurology 72, 1048–1055.

20.

Remy

, Vayssiere

, Saint-Aubert

, Barbeau

, Pariente

(2015) White matter disruption at the prodromal stage of Alzheimer’s disease: Relationships with hippocampal atrophy and episodic memory performance. Neuroimage Clin 7, 482–492.

21.

, Muftuler

, Chen

, Ward

, Budde

, Jones

, Franczak

, Antuono

, Li

, Goveas

(2014) Effects of the coexistence of late-life depression and mild cognitive impairment on white matter microstructure. J Neurol Sci 338, 46–56.

22.

Zhuang

, Sachdev

, Trollor

, Reppermund

, Kochan

, Brodaty

, Wen

(2013) Microstructural white matter changes, not hippocampal atrophy, detect early amnestic mild cognitive impairment. PLoS One 8, e58887.

23.

Reiman

, Caselli

, Yun

, Chen

, Bandy

, Minoshima

, Thibodeau

, Osborne

(1996) Preclinical evidence of Alzheimer’s disease in persons homozygous for the ɛ4 allele for apolipoprotein E. New Engl J Med 334, 752–758.

24.

Cohen

, Small

, Lalonde

, Friz

, Sunderland

(2001) Effect of apolipoprotein E genotype on hippocampal volume loss in aging healthy women. Neurology 57, 2223–2228.

25.

Liu

, Paajanen

, Westman

, Wahlund

, Simmons

, Tunnard

, Sobow

, Proitsi

, Powell

, Mecocci

, Tsolaki

, Vellas

, Muehlboeck

, Evans

, Spenger

, Lovestone

, Soininen

, AddNeuroMed Consortium (2010) Effect of APOE epsilon4 allele on cortical thicknesses and volumes: The AddNeuroMed study. J Alzheimers Dis 21, 947–966.

26.

Honea

, Vidoni

, Harsha

, Burns

(2009) Impact of APOE on the healthy aging brain: A voxel-based MRI and DTI study. J Alzheimers Dis 18, 553–564.

27.

Cherbuin

, Anstey

, Sachdev

, Maller

, Meslin

, Mack

, Wen

, Easteal

(2008) Total and regional gray matter volume is not related to APOE*E4 status in a community sample of middle-aged individuals. J Gerontol A Biol Sci Med Sci 63, 501–504.

28.

Lemaitre

, Crivello

, Dufouil

, Grassiot

, Tzourio

, Alperovitch

, Mazoyer

(2005) No epsilon4 gene dose effect on hippocampal atrophy in a large MRI database of healthy elderly subjects. Neuroimage 24, 1205–1213.

29.

Jack

Jr , Petersen

, Xu

, O’Brien

, Waring

, Tangalos

, Smith

, Ivnik

, Thibodeau

, Kokmen

(1998) Hippocampal atrophy and apolipoprotein E genotype are independently associated with Alzheimer’s disease. Ann Neurol 43, 303–310.

30.

Bagepally

, Halahalli

, John

, Kota

, Purushottam

, Mukherjee

, Sivakumar

, Bharath

, Jain

, Varghese

(2012) Apolipoprotein E4 and brain white matter integrity in Alzheimer’s disease: Tract-based spatial statistics study under 3-Tesla MRI. Neurodegener Dis 10, 145–148.

31.

Persson

, Lind

, Larsson

, Ingvar

, Cruts

, Van Broeckhoven

, Adolfsson

, Nilsson

L-G

, Nyberg

(2006) Altered brain white matter integrity in healthy carriers of the APOE ɛ4 allele A risk for AD? Neurology 66, 1029–1033.

32.

Nierenberg

, Pomara

, Hoptman

, Sidtis

, Ardekani

, Lim

(2005) Abnormal white matter integrity in healthy apolipoprotein E epsilon4 carriers. Neuroreport 16, 1369–1372.

33.

Xie

, Xiao

, Wang

, Wu

, Gong

, Jiang

(2005) Evaluation of bilateral cingulum with tractography in patients with Alzheimer’s disease. Neuroreport 16, 1275–1278.

34.

Sydykova

, Stahl

, Dietrich

, Ewers

, Reiser

, Schoenberg

, Möller

H-J

, Hampel

, Teipel

(2007) Fiber connections between the cerebral cortex and the corpus callosum in Alzheimer’s disease: A diffusion tensor imaging and voxel-based morphometry study. Cereb Cortex 17, 2276–2282.

35.

Chen

, Shu

, Wang

, Liu

, Shi

, Zhang

(2015) The interaction of APOE genotype by age in amnestic mild cognitive impairment: A voxel-based morphometric study. J Alzheimers Dis 43, 657–668.

36.

Shu

, Shi

, Chen

, Wang

, Liu

, Yue

, Ward

, Li

, Xu

, Chen

, Guo

, Xu

, Li

, Zhang

(2016) Opposite neural trajectories of apolipoprotein E 4 and 2 alleles with aging associated with different risks of Alzheimer’s disease. Cereb Cortex 26, 1421–1429.

37.

Ding

, Zhao

, Guo

, Meng

, Wang

, Yu

, Luo

, Zhou

, Yu

, Zheng

, Chu

, Mortimer

, Borenstein

, Hong

(2014) The Shanghai Aging Study: Study design, baseline characteristics, and prevalence of dementia. Neuroepidemiology 43, 114–122.

38.

Wang

, Wang

, Li

, Yu

(2014) Frequency of the apolipoprotein E epsilon4 allele in a memory clinic cohort in Beijing: A naturalistic descriptive study. PLoS One 9, e99130.

39.

Taylor

, Scanlon

, Farrell

, Hernandez

, Adamson

, Ashford

, Noda

, Murphy

Jr , Weiner

(2014) APOE-epsilon4 and aging of medial temporal lobe gray matter in healthy adults older than 50 years. Neurobiol Aging 35, 2479–2485.

40.

Serrano-Pozo

, Qian

, Monsell

, Betensky

, Hyman

(2015) APOEepsilon2 is associated with milder clinical and pathological Alzheimer’s disease. Ann Neurol 77, 917–929.

41.

Conejero-Goldberg

, Gomar

, Bobes-Bascaran

, Hyde

, Kleinman

, Herman

, Chen

, Davies

, Goldberg

(2014) APOE2 enhances neuroprotection against Alzheimer’s disease through multiple molecular mechanisms. Mol Psychiatry 19, 1243–1250.

42.

Gong

, He

, Chen

, Evans

(2012) Convergence and divergence of thickness correlations with diffusion connections across the human cerebral cortex. Neuroimage 59, 1239–1248.

43.

Sled

, Zijdenbos

, Evans

(1998) A nonparametric method for automatic correction of intensity nonuniformity in MRI data. IEEE Trans Med Imaging 17, 87–97.

44.

Tohka

, Zijdenbos

, Evans

(2004) Fast and robust parameter estimation for statistical partial volume models in brain MRI. Neuroimage 23, 84–97.

45.

Kim

, Singh

, Lee

, Lerch

, Ad-Dab’bagh

, MacDonald

, Lee

, Kim

, Evans

(2005) Automated 3-D extraction and evaluation of the inner and outer cortical surfaces using a Laplacian map and partial volume effect classification. Neuroimage 27, 210–221.

46.

MacDonald

, Kabani

, Avis

, Evans

(2000) Automated 3-D extraction of inner and outer surfaces of cerebral cortex from MRI. Neuroimage 12, 340–356.

47.

Chung

, Worsley

, Robbins

, Paus

, Taylor

, Giedd

, Rapoport

, Evans

(2003) Deformation-based surface morphometry applied to gray matter deformation. Neuroimage 18, 198–213.

48.

Smith

, Jenkinson

, Woolrich

, Beckmann

, Behrens

, Johansen-Berg

, Bannister

, De Luca

, Drobnjak

, Flitney

(2004) Advances in functional and structural MR image analysis and implementation as FSL. Neuroimage 23, S208–S219.

49.

Cui

, Zhong

, Xu

, He

, Gong

(2013) PANDA: A pipeline toolbox for analyzing brain diffusion images. Front Hum Neurosci 7, 42.

50.

Smith

, Jenkinson

, Johansen-Berg

, Rueckert

, Nichols

, Mackay

, Watkins

, Ciccarelli

, Cader

, Matthews

(2006) Tract-based spatial statistics: Voxelwise analysis of multi-subject diffusion data. Neuroimage 31, 1487–1505.

51.

Lim

, Ellis

, Ames

, Darby

, Harrington

, Martins

, Masters

, Rowe

, Savage

, Szoeke

, Villemagne

, Maruff

, Group

(2013) Aβ amyloid, cognition, and APOE genotype in healthy older adults. Alzheimers Dement 9, 538–545.

52.

Braak

, Braak

(1995) Staging of Alzheimer’s disease-related neurofibrillary changes. Neurobiol Aging 16, 271–278.

53.

Serra

, Cercignani

, Lenzi

, Perri

, Fadda

, Caltagirone

, Macaluso

, Bozzali

(2010) Grey and white matter changes at different stages of Alzheimer’s disease. J Alzheimers Dis 19, 147–159.

54.

Braak

, Braak

(1998) Evolution of neuronal changes in the course of Alzheimer’s disease. J Neural Transm Suppl 53, 127–140.

55.

Lerch

, Pruessner

, Zijdenbos

, Hampel

, Teipel

, Evans

(2005) Focal decline of cortical thickness in Alzheimer’s disease identified by computational neuroanatomy. Cereb Cortex 15, 995–1001.

56.

Singh

, Chertkow

, Lerch

, Evans

, Dorr

, Kabani

(2006) Spatial patterns of cortical thinning in mild cognitive impairment and Alzheimer’s disease. Brain 129, 2885–2893.

57.

Song

, Sun

, Ramsbottom

, Chang

, Russell

, Cross

(2002) Dysmyelination revealed through MRI as increased radial (but unchanged axial) diffusion of water. Neuroimage 17, 1429–1436.

58.

Sun

, Liang

, Cross

, Song

(2008) Evolving Wallerian degeneration after transient retinal ischemia in mice characterized by diffusion tensor imaging. Neuroimage 40, 1–10.

59.

Sun

, Liang

, Trinkaus

, Cross

, Armstrong

, Song

(2006) Noninvasive detection of cuprizone induced axonal damage and demyelination in the mouse corpus callosum. Magn Reson Med 55, 302–308.

60.

Taoka

, Iwasaki

, Sakamoto

, Nakagawa

, Fukusumi

, Myochin

, Hirohashi

, Hoshida

, Kichikawa

(2006) Diffusion anisotropy and diffusivity of white matter tracts within the temporal stem in Alzheimer disease: Evaluation of the tract of interest by diffusion tensor tractography. AJNR Am J Neuroradiol 27, 1040–1045.

61.

Serra

, Cercignani

, Basile

, Spano

, Perri

, Fadda

, Marra

, Giubilei

, Caltagirone

, Bozzali

(2012) White matter damage along the uncinate fasciculus contributes to cognitive decline in AD and DLB. Curr Alzheimer Res 9, 326–333.

62.

Agosta

, Pievani

, Sala

, Geroldi

, Galluzzi

, Frisoni

, Filippi

(2011) White matter damage in Alzheimer disease and its relationship to gray matter atrophy. Radiology 258, 853–863.

63.

Plassman

, Welsh-Bohmer

, Bigler

, Johnson

, Anderson

, Helms

, Saunders

, Breitner

(1997) Apolipoprotein E epsilon 4 allele and hippocampal volume in twins with normal cognition. Neurology 48, 985–989.

64.

Burggren

, Zeineh

, Ekstrom

, Braskie

, Thompson

, Small

, Bookheimer

(2008) Reduced cortical thickness in hippocampal subregions among cognitively normal apolipoprotein E e4 carriers. Neuroimage 41, 1177–1183.

65.

O’Dwyer

, Lamberton

, Matura

, Tanner

, Scheibe

, Miller

, Rujescu

, Prvulovic

, Hampel

(2012) Reduced hippocampal volume in healthy young ApoE4 carriers: An MRI study. PLoS One 7, e48895.

66.

Cherbuin

, Anstey

, Sachdev

, Maller

, Meslin

, Mack

, Wen

, Easteal

(2008) Total and regional gray matter volume is not related to APOE*E4 status in a community sample of middle-aged individuals. J Gerontol A Biol Sci Med Sci 63, 501–504.

67.

Heise

, Filippini

, Ebmeier

, Mackay

(2011) The APOE ɛ4 allele modulates brain white matter integrity in healthy adults. Mol Psychiatry 16, 908–916.

68.

Villain

, Desgranges

, Viader

, de la Sayette

, Mezenge

, Landeau

, Baron

, Eustache

, Chetelat

(2008) Relationships between hippocampal atrophy, white matter disruption, and gray matter hypometabolism in Alzheimer’s disease. J Neurosci 28, 6174–6181.

69.

Choo

, Lee

, Oh

, Lee

, Song

, Youn

, Kim

, Jhoo

, Woo

(2010) Posterior cingulate cortex atrophy and regional cingulum disruption in mild cognitive impairment and Alzheimer’s disease. Neurobiol Aging 31, 772–779.

70.

Villain

, Fouquet

, Baron

, Mezenge

, Landeau

, de La Sayette

, Viader

, Eustache

, Desgranges

, Chetelat

(2010) Sequential relationships between grey matter and white matter atrophy and brain metabolic abnormalities in early Alzheimer’s disease. Brain 133, 3301–3314.

71.

Ebeling

, von Cramon

(1992) Topography of the uncinate fascicle and adjacent temporal fiber tracts. Acta Neurochir (Wien) 115, 143–148.

72.

Kier

, Staib

, Davis

, Bronen

(2004) MR imaging of the temporal stem: Anatomic dissection tractography of the uncinate fasciculus, inferior occipitofrontal fasciculus, and Meyer’s loop of the optic radiation. AJNR Am J Neuroradiol 25, 677–691.

73.

Damoiseaux

, Smith

, Witter

, Sanz-Arigita

, Barkhof

, Scheltens

, Stam

, Zarei

, Rombouts

(2009) White matter tract integrity in aging and Alzheimer’s disease. Hum Brain Mapp 30, 1051–1059.

74.

Vann

, Aggleton

, Maguire

(2009) What does the retrosplenial cortex do? Nat Rev Neurosci 10, 792–802.

75.

Ewers

, Sperling

, Klunk

, Weiner

, Hampel

(2011) Neuroimaging markers for the prediction and early diagnosis of Alzheimer’s disease dementia. Trends Neurosci 34, 430–442.

76.

Garrido

, Furuie

, Buchpiguel

, Bottino

, Almeida

, Cid

, Camargo

, Castro

, Glabus

, Busatto

(2002) Relation between medial temporal atrophy and functional brain activity during memory processing in Alzheimer’s disease: A combined MRI and SPECT study. J Neurol Neurosurg Psychiatry 73, 508–516.

77.

Rémy

, Mirrashed

, Campbell

, Richter

(2005) Verbal episodic memory impairment in Alzheimer’s disease: A combined structural and functional MRI study. Neuroimage 25, 253–266.

78.

Nestor

, Fryer

, Smielewski

, Hodges

(2003) Limbic hypometabolism in Alzheimer’s disease and mild cognitive impairment. Ann Neurol 54, 343–351.

79.

, Chen

, Ahmed

, Chen

, Ku

, Goldberg

, Hsu

(2001) Amyloid-ß peptides are cytotoxic to oligodendrocytes. J Neurosci 21, 1–5.

80.

Gomar

, Bobes-Bascaran

, Conejero-Goldberg

, Davies

, Goldberg

(2011) Alzheimer’s Disease Neuroimaging I, Utility of combinations of biomarkers, cognitive markers, and risk factors to predict conversion from mild cognitive impairment to Alzheimer disease in patients in the Alzheimer’s disease neuroimaging initiative. Arch Gen Psychiatry 68, 961–969.

81.

Peters

, Villeneuve

, Belleville

(2014) Predicting progression to dementia in elderly subjects with mild cognitive impairment using both cognitive and neuroimaging predictors. J Alzheimers Dis 38, 307–318.

82.

Albert

, DeKosky

, Dickson

, Dubois

, Feldman

, Fox

, Gamst

, Holtzman

, Jagust

, Petersen

(2011) The diagnosis of mild cognitive impairment due to Alzheimer’s disease: Recommendations from the National Institute on Aging-Alzheimer’s Association workgroups on diagnostic guidelines for Alzheimer’s disease. Alzheimers Dement 7, 270–279.

83.

Dubois

, Feldman

, Jacova

, Hampel

, Molinuevo

, Blennow

, DeKosky

, Gauthier

, Selkoe

, Bateman

, Cappa

, Crutch

, Engelborghs

, Frisoni

, Fox

, Galasko

, Habert

, Jicha

, Nordberg

, Pasquier

, Rabinovici

, Robert

, Rowe

, Salloway

, Sarazin

, Epelbaum

, de Souza

, Vellas

, Visser

, Schneider

, Stern

, Scheltens

, Cummings

(2014) Advancing research diagnostic criteria for Alzheimer’s disease: The IWG-2 criteria. Lancet Neurol 13, 614–629.

84.

Gomar

, Conejero-Goldberg

, Davies

, Goldberg

, Alzheimer’s Disease Neuroimaging Initiative (2014) Extension and refinement of the predictive value of different classes of markers in ADNI: Four-year follow-up data. Alzheimers Dement 10, 704–712.