The Disconnection Hypothesis in Alzheimer’s Disease Studied Through Multimodal Magnetic Resonance Imaging: Structural,Perfusion,and Diffusion Tensor Imaging

Abstract

According to the so-called disconnection hypothesis, the loss of synaptic inputs from the medial temporal lobes (MTL) in Alzheimer’s disease (AD) may lead to reduced activity of target neurons in cortical areas and, consequently, to decreased cerebral blood flow (CBF) in those areas. The aim of this study was to assess whether hypoperfusion in parietotemporal and frontal cortices of patients with mild cognitive impairment who converted to AD (MCI-c) and patients with mild AD is associated with atrophy in the MTL and/or microstructural changes in the white matter (WM) tracts connecting these areas. We assessed these relationships by investigating correlations between CBF in hypoperfused areas, mean cortical thickness in atrophied regions of the MTL, and fractional anisotropy (FA) in WM tracts. In the MCI-c group, a strong correlation was observed between CBF of the superior parietal gyri and FA in the parahippocampal tracts (left: r = 0.90, p < 0.0001; right: r = 0.597, p = 0.024), and between FA in the right parahippocampal tract and the right precuneus (r = 0.551, p = 0.041). No significant correlations between CBF in hypoperfused regions and FA in the WM tract were observed in the AD group. These results suggest an association between perfusion deficits and altered WM tracts in prodromal AD, while microvasculature impairments may have a greater influence in more advanced stages. We did not find correlations between cortical thinning in the medial temporal lobes and decreased FA in the WM tracts of the limbic system in either group.

Keywords

Alzheimer’s disease diffusion tensor imaging disconnection hypothesis magnetic resonance imaging mild cognitive impairment perfusion weighted imaging

INTRODUCTION

Cognitive functions seem to be the result of temporally coordinated neuronal activity in more than one brain region, integrated into neural networks, rather than activation of an isolated group of cells [1, 2]. Disruption of these functional networks may lead to cognitive impairment. Several neuroimaging studies point to the existence of cortical disconnection as a possible cause of cognitive impairment in Alzheimer’s disease (AD) [3 –6]. According to this hypothesis, AD-related symptoms would be the consequence— at least partially— of abnormal interactions between neuronal systems, rather than the effect of localized pathophysiology in one or more brain regions [7, 8]. In fact, a strong correlation has been reported between synapse loss and cognitive impairment, whereas a weak correlation exists between clinical symptoms and the level of deposition of amyloid plaques and intracellular formation of tangles [9]. The loss of synaptic inputs in cortical neurons may lead to a reduction in the activity of target neurons and, consequently, to a decrease in cerebral blood flow (CBF) in the corresponding cortical target area [10]. According to this hypothesis, hypoperfusion in the parietotemporal and frontal cortices observed in the pre-dementia and mild dementia stages of AD in our previous work [11] might be related to a disconnection of these regions from the medial temporal lobe.

The combination of MRI techniques (structural, perfusion, and diffusion) could elucidate the structural and functional features of the disconnection hypothesis. Structural MRI is the most widely used neuroimaging technique for diagnosing and evaluating AD progression in clinical and research settings [12 –15]. Post mortem studies have confirmed the anatomical validity of structural MRI volume measurements by establishing correlations between neuronal loss in histological slices and atrophy level observed with MRI [16, 17]. Perfusion MRI is a good alternative to nuclear medicine for the evaluation of cerebral blood perfusion. It has the advantages of shorter acquisition time, lower cost, and no need for ionizing radiation, thus facilitating clinical routine and patient comfort [18, 19]. Diffusion tensor imaging (DTI) provides information about water diffusion properties in brain tissues at the cellular level. White matter (WM) tracts present higher diffusion anisotropy than other tissues. This anisotropy could be altered even in preclinical stages of AD, therefore making DTI a useful tool in the study of microstructural changes in WM fibers in the progression of AD [20, 21]. Fractional anisotropy (FA) is a DTI measurement that indirectly quantifies the degree of geometric organization of axonal bundles across the brain [22, 23]. Microstructural changes underlying FA decrease in WM tracts remain unclear. Reduced fiber density or decreased myelination in fiber tracts manifests as decreased FA, but also gliosis, inflammation and ischemic processes may reduce anisotropy in WM [24 –26].

All three MRI techniques have been combined for the characterization of patients with mild cognitive impairment (MCI) and AD [27, 28]. However, these studies focused on the use of the three techniques in the same regions, namely, the posterior cingulate cortex and the hippocampus. To our knowledge, there are no previous reports on the relationships between atrophy of the cortical gray matter (GM) of the medial temporal lobe, perfusion deficits in parietotemporal and frontal cortices, and microstructural changes in the WM tracts connecting such structures. The aim of our work was to study possible correlations between CBF values in hypoperfused areas, mean cortical thickness in the medial temporal lobe, and FA in the WM tracts. We also studied the correlation between cortical atrophy in the medial temporal lobes and microstructural changes in the WM tracts of the limbic system. Linear correlations were computed separately in three groups: patients with MCI who converted to dementia due to AD, patients with mild AD dementia and healthy controls. Finally, we computed connectivity matrices to study possible disconnections of hypoperfused parietotemporal and frontal areas from atrophied regions of the medial temporal lobes.

MATERIAL AND METHODS

Subjects

The study population comprised 48 individuals who were classified into three groups: patients with MCI who had converted to dementia due to AD after 2 years of clinical follow-up (MCI-c) (n = 16), patients with AD and mild dementia (n = 12), and volunteers with normal cognition (n = 20).

Participants were prospectively recruited during routine clinical practice in the behavioral neurology clinic of our teaching hospital by a single senior neurologist with expertise in behavioral neurology. Control subjects were chosen from volunteers attending the behavioral neurology clinic as patient caregivers and from among researchers’ acquaintances. Written informed consent was obtained from all participants, and all examinations were performed according to a protocol approved by the Hospital General Universitario Gregorio Marañón Ethics Committee. All the participants underwent an interview and a physical and neurological examination, including a formal battery of neuropsychological tests, the details of which can be found elsewhere [11]. The specific inclusion criteria for each study group were:

MCI-c group: MCI was diagnosed using the criteria of Winblad et al. [29]. The selection criteria were as follows: 1) cognitive impairment had to be supported by an abnormal performance (1-1.5 SD below the expected performance for age and education) in 1 or more tests from the neuropsychological battery; 2) patients had to be at stage 0.5 in the Clinical Dementia Rating (CDR); and 3) after 2 years of clinical follow-up, the patients had to have progressed to dementia and fulfilled the same criteria as the mild AD group.

Mild AD group: Probable AD was diagnosed using the criteria established by the National Institute on Aging and the Alzheimer’s Association workgroup [30]. Hence, deterioration in memory and other cognitive functions had to be documented, and instrumental activities of daily living had to be affected. The diagnosis of mild AD was always supported by a score of 1 (i.e., mild dementia) in the CDR.

Control group: Subjects were cognitively normal, and performed within the normal limits in the cognitive tests, according to age and education. They showed a score of 0 in the CDR and a score above 24 in the Mini-Mental State Examination (MMSE).

MRI-based structural and perfusion data were obtained from all participants. Diffusion data could only be obtained from 16/20 controls, 14/16 MCI-c patients, and 10/12 AD patients owing to body movements during image acquisition. Table 1 summarizes the demographic and clinical data of the participants who had a DTI scan. No significant between-group differences were found for these demographic variables, except for the MMSE score, which was lower in the AD and MCI-c groups than in the control group.

Image acquisition

MRI data were acquired using a Philips Intera 1.5 T scanner (Philips Medical Systems, Best, The Netherlands). The imaging protocol included a volumetric T1-weighted 3D gradient echo, which was used for tissue segmentation (flip angle = 30°; repetition time [TR] = 16 ms; echo time [TE] = 4.6 ms; matrix size = 256×256; FOV = 256×256 mm and 100 slices with a thickness of 1.5 mm).

Perfusion-weighted images (PWI) were obtained using an echo-planar imaging sequence (EPI factor = 61; flip angle = 40°; TR = 1439 ms; TE = 30 ms; matrix size = 128×128; FOV = 230×230 mm; section thickness = 5 mm) after injecting a bolus of gadolinium chelate (10 ml of gadobutrol; Gadovist^® Bayer-Schering AG, Berlin, Germany), which was followed by administration of 30 ml of saline solution (4 ml/s). Forty volumes (30 slices each) per subject were obtained during administration of the gadolinium contrast (automatic injector, 4 ml/s).

Diffusion-weighted images (DWI) were acquired using a single-shot echo-planar imaging sequence with the following parameters: imaging plane, axial; phase encoding direction, A–P; TE = 68 ms, TR = 11 886 ms; flip angle = 90°; EPI factor = 77; number of slices = 60; interslice gap = 0 mm; voxel size = 2×2×2 mm; and matrix size = 128×128; 15 diffusion weighting directions with b = 800 s/mm² and one additional image without diffusion weighting (b = 0).

Image processing

Structural and PWI processing

The anatomical T1-weighted image was used to estimate regional mean cortical thickness using the FreeSurfer package (version 4.5.1, http://surfer.nmr.mgh.harvard.edu). First, brain masks were obtained from skull-stripped baseline images generated using the VBM8 toolbox from the SPM8 package (Wellcome Trust Centre for Neuroimaging, London, UK). The brain mask images were then introduced into the default FreeSurfer processing pipeline because they ensure more accurate skull stripping [11]. The white and gray cortical surfaces were reconstructed from the raw unaligned images in native space using the methods described by Fischl and Dale [31]. Mean cortical thickness for different regions of interest (ROIs) was obtained by applying the Desikan-Killiany atlas [32]. T1-weighted images were rigidly co-registered with the CBF parametric maps of PWI [11]. This registration procedure was done using an in-house algorithm [33] that uses normalized mutual information as the cost function and Powell as the optimization algorithm. A trilinear interpolation was used, with two subsampling steps: 442 (XYZ) and 221. Gyrus-based ROI masks, which were obtained by applying the Desikan-Killiany atlas, were then applied to perfusion maps and the average CBF per ROI was computed. Further information about the specific methods used to calculate the CBF parametric maps can be found in Gordalizaet al. [34].

DWI processing

ROI analysis of FA

Diffusion-weighted images were processed using version 4.1 of the software package FSL (FMRIB Software Library, FMRIB, Oxford, UK) [35].

For each subject, the 15 diffusion-weighted volumes acquired were co-registered to the b0 image using FLIRT (FMRIB’s Linear Image Registration Tool) to correct for possible eddy current–induced distortion and subject motion. A brain mask was created from the first b0 image using BET (Brain Extraction Tool) [36], and FDT (FMRIB’s Diffusion Toolbox) [37] was used to fit the tensor model and to compute the FA maps. Parcellation of the individual FA maps into different tracts was based on the ICBM-DTI-81 white matter labels atlas, which is one of the standard atlases available in FSL [38, 39]. We extracted the mean FA values of each tract-based ROI following the methodology described in Navas-S $\overset{´}{a}$ nchez et al. [40].

WM tractography and connectivity matrices computation

The T1 image for each subject was registered to the diffusion space using the b0 volume as the target image. A two–step registration procedure (affine followed by non–linear) was done using ANTS [41] using mutual information as the cost function. This additional registration step was required in order to obtain the transformations between MNI and diffusion spaces. Diffusion images were processed using a fully automated fiber tractography algorithm [42] and connectivity values (anatomical connection density, ACD) were obtained for each pair of ROIs in the Desikan-Killiany atlas.

Quantification

The first step in our study was to detect which WM tracts of the whole brain presented microstructural changes in each patient group. We then performed a correlation analysis focusing on the altered WM tracts that, based on the literature, presumably connect atrophied areas of the medial temporal lobe with hypoperfused regions in the parietotemporal and frontal cortices. Gyrus-based ROIs which showed cortical thinning and hypoperfusion were taken from our previous work that studied the same sample [11]. These ROIs are detailed in Correlation analysis of CBF, mean cortical thickness, and FA data section. We also selected significantly altered WM tracts of the limbic system to study correlations between microstructural changes in WM tracts and atrophy of the medial temporal lobe. All WM ROIs selection methodology is detailed in the next section, Tract-based ROI analysis and selection.

Finally, we calculated the ACD between every hypoperfused and atrophied ROI observed within each patient group (see ROIs included in the ACD study section), as well as the correlation between ACD and FA in the tracts connecting hypoperfused to atrophiedregions.

Tract-based ROI analysis and selection

We carried out a between-group comparative analysis of FA values in WM tracts of the whole brain in order to detect possible microstructural changes in WM fibers between the study groups. We compared mean FA values of tract-based ROIs in the major association fibers and limbic system fibers from both hemispheres and callosal tracts [39]. In the case of the association fibers, we focused on i) the superior longitudinal fasciculus and the superior fronto-occipital fasciculus (which connect the frontal, temporal, parietal, and occipital lobes); and ii) the uncinate fasciculus (which connects the temporal and frontal lobes). In the case of the limbic system tracts, we focused on i) the cingulum and the parahippocampal tract (the part of the cingulum situated in the medial temporal lobe) which interconnect the medial temporal lobe and parietotemporal regions; and ii) the fornix and fornix-stria terminalis (ST) (which connect the medial temporal lobe with the mammillary bodies). Callosal tracts included the genu, the body, and the splenium.

Correlation analysis of CBF, mean cortical thickness, and FA

Figure 1 illustrates the set of ROIs included in the correlation analysis. Correlations were computed between cortical thickness of every atrophied region of the medial temporal lobes (bottom boxes in Fig. 1) and CBF of every hypoperfused region in parietotemporal or frontal cortices (superior boxes in Fig. 1), separately for each patient group. Correlations were also computed between FA of the altered WM tracts (wide arrows in Fig. 1) and CBF of hypoperfused regions and cortical thickness of atrophied regions. In the control group, correlations were computed for the same ROIs selected in both patient groups. In order to unveil possible associations between perfusion deficits in the parietotemporal and frontal cortices and atrophy in the medial temporal lobes, we investigated correlations between the mean cortical thickness of the atrophied ROIs and CBF values of the hypoperfused ROIs in each hemisphere. We also investigated correlations between CBF in hypoperfused ROIs and the FA data of the cingulate and parahippocampal WM tracts that revealed significant microstructural differences in the comparative analysis of WM tracts. These two specific tracts were selected because they connect the medial temporal lobe with the parietal, frontal, and lateral temporal lobes [7].

Finally, we correlated the mean cortical thickness in atrophied ROIs of the medial temporal lobe with FA values of WM tracts of the limbic system that showed significant microstructural changes in the previous comparative analysis.

ROIs included in the ACD study

We computed the ACD values between every hypoperfused ROI in parietotemporal and frontal areas and atrophied ROIs in the medial temporal lobe, separately for each patient group (Fig. 1). In the control group, ACD values were computed for the same ROIs selected in both patient groups.

Correlation analysis of FA and ACD

In order to elucidate whether the association between hypoperfusion in the parietotemporal cortex and FA decrease in the parahippocampal tract depends on disconnection, we computed the correlation between FA in this tract and the ACD between parahippocampal gyrus and hypoperfused ROIs. This analysis was applied to those groups that showed a significant correlation between parietotemporal hypoperfusion and FA decrease in the parahippocampal tract (Fig. 1).

Statistical analysis

We tested for between-group differences in FA in tract-based ROIs using an ANCOVA model including age and gender as covariates. However, gender did not show any significant effect on the model and we removed it from the final ANCOVA in order to optimize the degrees of freedom, in the light of our small sample size. When the ANCOVA model revealed significant differences between groups, we used a post hoc test (Dunnett’s test) to compare mean values between patient groups and controls and computed the effect size (Cohen’s coefficient, d). We used Pearson’s coefficient to study correlations. Finally, we used a t-test to compare mean ACD in MCI-c versus controls and in AD versus controls. Data were processed using Statistical Analysis System (SAS) version 9.0 (SAS Institute Inc., Cary, North Carolina, USA).

RESULTS

Comparative ROI-based analysis of FA data

The ANCOVA model revealed significant between-group differences in the genu (p = 0.025), splenium (p = 0.035), left cingulate tract (p = 0.014), parahippocampal tracts (left: p = 0.003; right: p = 0.001), and fornix-ST (left: p = 0.027; right: p = 0.003). Figure 2 shows tract-based ROIs which presented significant FA differences between each patient group and controls. Table 2 summarizes the effect size and p value (Dunnett’s test) for differences in FA between controls and each patient group. MCI-c group presented decreased FA in the genu (d = –1.01; p = 0.023), left cingulum (d = –1.13; p = 0.014), left fornix-ST (d = –1.07; d = 0.015), right fornix-ST (d = –1.35; p = 0.002), left parahippocampal tract (d = –1.28; p = 0.004), and right parahippocampal tract (d = –1.57; p = 0.001). AD group presented decreased FA in the splenium (d = –1.41; p = 0.024), left cingulum (d = –1.15; p = 0.044), left parahippocampal tract (d = –1.75; p = 0.010), and right parahippocampal tract (d = –1.37; p = 0.010).

Correlation between cortical thickness in the medial temporal lobe and CBF in parietotemporal and frontal cortices

We did not find a significant correlation between mean cortical thickness in the atrophied ROIs and CBF values in hypoperfused ROIs for any of the three groups studied.

Correlation between FA in the parahippocampal and cingulate WM tracts and CBF in the parietotemporal and frontal cortices

A strong correlation between FA in the left parahippocampal tract and CBF of the left superior parietal gyrus was observed in the MCI-c group (r = 0.90, p < 0.0001). Correlations were found in the right hemisphere between FA in the parahippocampal tract and CBF of the superior parietal gyrus (r = 0.597, p = 0.024) and precuneus (r = 0.551, p = 0.041) (Fig. 3). No significant correlations were found for AD patients and controls.

Correlation between cortical thickness in the medial temporal lobes and FA in WM tracts of the limbic system

We did not find a significant correlation between mean cortical thickness in atrophied ROIs of the medial temporal lobe (entorhinal cortex or parahippocampal gyrus) and decreased FA in the tract-based ROIs of the limbic system (parahippocampal tract, cingulum and fornix-ST) for any of the three groups studied.

Anatomical connection between hypoperfused ROIs in the parietotemporal and frontal cortices and atrophied ROIs in the medial temporal lobes

We observed a trend of reduced ACD between atrophied regions in the medial temporal lobe and hypoperfused cortical areas in parietotemporal and frontal cortices in MCI-c and AD groups when compared to controls. However it did not reach statistical significance in the MCI-c group; in the AD group, the ACD reduction was only significant between the right entorhinal cortex and the right superior temporal lobe (d = –0.84; p = 0.044) (Table 3).

Correlation between FA in the parahippocampal tract and ACD between atrophied and hypoperfused regions

In the MCI-c group, we found significant correlations in the right hemisphere between FA of the parahippocampal tract and two ACD values: a) between parahippocampal gyrus and the superior parietal gyrus (r = 0.388, p = 0.034); and b) between parahippocampal gyrus and precuneus (r = 0.435, p = 0.016).

DISCUSSION

According to the so-called disconnection hypothesis, perfusion deficit in the parietotemporal and frontal cortices of patients with AD is thought to be related to a functional disconnection from the medial temporal lobe. However, several neuropathological studies have demonstrated that cerebrovascular lesions are very frequent in AD, making mixed dementia (neuronal and vascular) the most common type of dementia, especially in the later stages of the disease [43 –45]. The study of hemodynamic alterations using multimodal MRI may help to elucidate whether perfusion deficits in AD patients are related to disconnection syndrome.

In the present study we investigated possible correlations between cortical atrophy in the medial temporal lobes, CBF deficits in the parietotemporal and frontal cortices, and decreased FA in WM tracts connecting these areas. To our knowledge, this is the first study to apply multimodal MRI (structural, perfusion, and diffusion) to test the disconnection hypothesis by comparing patients with AD, MCI-c, and aged-matched healthy controls. According to our results, in the MCI-c group, hypoperfusion in the parietal lobes is related to a decrease in FA in parahippocampal WM fibers but not to cortical thinning in the medial temporal lobe. On the contrary, in the AD group, hypoperfusion in parietotemporal and frontal cortices did not correlate either with FA decrease in WM tracts or cortical thickness in the medial temporal lobe.

Between-group comparative analysis for FA data

Our results in the comparative tract-based ROI analysis agree with previously published data. Decreased FA in the cingulate and parahippocampal WM tracts, fornix, and corpus callosum has been extensively reported in MCI and AD patients in the literature [46 –51]. Possible damage to these WM tracts could underlie the breakdown of functional networks that could in turn generate cognitive impairment. The hippocampus, cingulum, and fornix axis form part of the memory circuit, and microstructural changes in those tracts are considered potential imaging markers of brain changes in preclinical stages of AD [21]. Microstructural changes in the parahippocampal WM fibers could indicate degeneration of the connections between the medial temporal lobes and parietotemporal and frontal cortices. Changes in the fornix-ST may correspond to damaged connections between the entorhinal cortex, amygdala, and the mammillary bodies, which are implicated in memory tasks. Some authors have observed a strong correlation between FA values in the fornix and performance in episodic memory tests in patients with MCI and AD, suggesting that damage in this structure is an early marker of progression of AD [47]. Changes in cingulate WM fibers could affect connectivity between the anterior thalamus, cingulate cortex, association cortices (in the frontal, temporal, and parietal lobes), and the hippocampi [7]. The decreased FA observed in some regions of the corpus callosum in the MCI-c group (genu) and the AD group (splenium) could indicate the occurrence of interhemispheric disconnection early in AD [51].

Correlations between atrophy in the medial temporal lobes, microstructural changes in parahippocampal WM fibers, and hypoperfusion in parietotemporal and frontal cortices in early AD

The existence of atrophy of the medial temporal lobes and reduced blood perfusion in the parietotemporal lobes of patients with AD was described by Jobst et al. [52], who also reported that the asymmetrical severity of atrophy of the medial temporal lobes was matched by an asymmetrical reduction in CBF in the parietotemporal regions. These results agree with the idea that the loss of neuronal inputs from the hippocampal region may play a role in the decrease in CBF in functionally connected cortical areas. As we reported in our previous work [11], atrophy in MCI-c patients was more extensive in the right medial temporal lobe than in the left one, and larger hypoperfused areas were also observed in the right parietal lobe, consistent with the observations of Jobst et al. However, in the present study, we did not find a correlation between cortical thinning in the atrophied medial temporal lobes and CBF values in any of the hypoperfused ROIs in the parietal lobes. This result suggests that the degree of parietal hypoperfusion is not directly related to the level of atrophy in the medial temporal lobes.

On the contrary, the fact that CBF values in parietal lobes correlate with FA values in the parahippocampal WM tracts suggests that hypoperfusion in the parietal lobes is related to microstructural changes in WM fibers, which connect hypoperfused areas with the medial temporal lobes. This observation is consistent with the idea that reduced activity in the parietotemporal and frontal cortices may be associated with damage in neuronal connections instead of with macrostructural changes in the cortex of the medial temporal lobes. The parietal lobe has traditionally been associated with spatial attention and visuo-motor control [53]. However, functional MRI studies have defined a parieto-hippocampal network associated with memory tasks [54]. Our results suggest that, owing to microstructural alterations in WM fibers connecting both lobes, this network could be physically damaged in the MCI-c patients we assessed. The results observed in the MCI-c group could support the existence of a disconnection syndrome that already appears in pre-dementia stages. To our knowledge, the present study is the first to use multimodal MRI to show evidence of disconnection syndrome that is potentially unrelated to cortical atrophy of the medial temporal lobes in MCI-c patients [55].

In AD patients, hypoperfusion in the parietotemporal and frontal areas seemed to be unrelated to either cortical thinning in the medial temporal lobes or microstructural changes in the WM fibers in the parahippocampal and cingulate tracts. This suggests that vascular disease may have a greater effect on cortical hemodynamic alterations than the altered connectivity from the medial temporal lobe. These facts would be in agreement with a mixed etiology (neuronal and vascular) for AD [45 , 57]. Vascular factors such as amyloid deposition around blood vessels, perivascular astrocytosis and vascular fibrosis, due to accumulation of extracellular matrix proteins in the endothelial basement membrane, may have greater influence in blood perfusion deficits in more advanced stages of disease. These vascular alterations are associated with the accumulation of neurotoxic molecules in the cerebral cortex that may subsequently contribute to neuronal degeneration [58, 59]. These findings agree with those reported in our previous work [11], where perfusion deficits seemed to be more related to functional deficits in the MCI-c group while the AD group showed biomarkers related to structural changes in the microvasculature, and they are also consistent with the idea that abnormal vascular function may precede vascular degeneration [60]. However, since the lack of significant correlations could be due to the small sample size, results are not decisive in this group. The small sample size in the AD group could also explain the absence of significant differences for the FA in the fornix compared to controls. On the other hand, in both MCI-c and AD patients, decreases in FA in the parahippocampal and left cingulate tracts did not correlate with thinning of the cortex in the parahippocampal gyrus and the entorhinal cortices. Similarly, in MCI-c, the FA of the fornix-ST did not correlate with entorhinal cortical thickness; these results suggest that reductions in FA are not associated with cortical atrophy in the medial temporal lobes, thus supporting that microstructural changes in WM are independent from macrostructural cortical changes in the medial temporal lobes. This independence of macrostructural and microstructural changes was recently reported by Canu et al. and Palesi et al. in patients with mild to moderate AD [61, 62]. However, a study in MCI patients proposed the existence of a sequence of events in which hippocampal atrophy induces a progressive breakdown of WM fibers, which could in turn lead to hypometabolism in other cortical areas [63]. The authors demonstrated that correlation between hippocampal and WM atrophy was not present at baseline but existed between basal values in hippocampal atrophy and rate of changes in WM volume over time. Unfortunately, the authors only measured macrostructural changes in WM, since they used structural MRI instead of DTI; therefore, their findings are not comparable to ours. A recent study has also reported a lack of correlation between the entorhinal cortex atrophy and FA decrease in the fornix and parahippocampal tracts; however, the authors suggest that atrophy in hippocampal subfield volumes, namely the subiculum, correlates with the FA decrease in the fornix [64].

The disconnection hypothesis is also supported by the results of histological studies: the distribution of AD-related neuropathology (neurofibrillary tangles and neuritic plaques) across the brain follows a vulnerability pattern that makes these markers particularly abundant in specific laminae in the cerebral cortex. Those specific laminae harbor pyramidal neurons responsible for cortico-cortical connections within and between hemispheres [7]. Therefore, an association seems to exist between cortical disconnection and the spread of the neuropathology and atrophy across the brain.

FA decrease in MCI and AD has been linked to myelin breakdown processes related to amyloid and tau deposits. Myelin breakdown may lead to axonal transport disruption and therefore to brain signaling failure [65, 66]. In our study, results of WM tractography supported the existence of a disconnection between hypoperfused regions in parietotemporal areas and atrophied areas in the medial temporal lobe, although only a non-statistically significant trend in this direction was observed. Tractography has been reported to be less sensitive, compared to direct assessment of FA in the domain of AD [67], that could explain the lack of statistical significance in this study. The existence of correlation between the FA and ACD values in the MCI-c group may additionally support the idea that microstructural changes in WM fibers (as assessed by FA) may depend on disconnection.

Although our sample size is similar to that of other DTI studies recently reviewed by Liu et al. [68] (each with 10 to 20 patients), the small number of participants limits statistical power. Therefore, our results should be interpreted with caution, especially for the AD group. Another limitation is the cross-sectional design. A longitudinal study could be more informative for the characterization and assessment of the progression of a possible disconnection syndrome. The atlas used for segmentation of tract-based ROIs is based on healthy individuals and may lead to misregistration problems when used in atrophied brains. However, the same atlas has been used before in patients with AD [69, 70]. Voxel-wise approaches (e.g., track-based spatial statistics, TBSS) [71] could have been used in our work. In the case of TBSS, all subjects’ FA maps are non-linearly registered to a normalized FA template and group differences may be reduced or eliminated during this registration procedure [72]. The ROI analysis has the advantage of been more sensitive, since the average of multiple voxels within a ROI enhances the signal to noise ratio [25], and more reliable because non-linear registration is not required. Also we did a tract-based ROI analysis in order to be consistent with the ROI analysis for structural and perfusion variables. Other methodological limitations are related to acquisition parameters: the number of directions for DWI (15) is below the optimal number for accurate tensor estimation (32) and the voxel size for T1 images (1x1x1.5 mm) may not be optimal for cortical thickness measurements. However, these parameters were chosen in order to reduce the acquisition time, since this is an important factor to consider in patients with dementia, as they tend to move inside the MRI scanner. The main asset of our study is the multimodal nature of the data (T1, PWI, and DTI) used to test the hypothesis of a disconnection syndrome in early AD. In addition, our access to information on the clinical progression of the MCI group enabled us to test this hypothesis in prodromal AD (i.e., MCI-c group), according to the new lexicon for Alzheimer’s disease proposed by Dubois et al. [73].

Conclusions

We did not find a correlation between CBF deficits in the parietotemporal and frontal cortices and cortical thinning in the medial temporal lobes in the MCI-c group or the AD group. Therefore, hypoperfusion seems not to be directly associated with medial temporal atrophy. In the MCI-c group, we observed a strong correlation between microstructural changes in the parahippocampal WM fibers and perfusion deficits in the parietal lobes: this finding may support the existence of a disconnection syndrome that could begin in pre-dementia stages of the disease. Tractography results aligned with the FA findings, showing a certain reduction in anatomical connections in MCI-c and AD groups, although not reaching statistical significance. In the AD group, the lack of a correlation between CBF values in the parietotemporal and frontal cortices and FA in parahippocampal and cingulate WM fibers suggests that microvasculature impairments may have a greater effect on perfusion deficits in more advances stages of the AD. The lack of a correlation between mean cortical thickness in the medial temporal lobes and the FA in WM tracts of the limbic system in both groups suggests that microstructural changes in these tracts are not directly associated with cortical atrophy in the medial temporal lobes.

Footnotes

ACKNOWLEDGMENTS

The authors are grateful to Dr Y Iturria-Medina, from the Brain Imaging Center, Montreal Neurological Institute, McGill University, for his helpful comments during the preparation of this manuscript. Also, we would like to thank V García-Vázquez, from the Medical Imaging Laboratory, HGUGH, for her help with methodological issues and F La Calle for his help during figure preparation. This work was funded by Brain Inspired Data Engineering, BRADE (S2013/ICE-2958).

Authors’ disclosures available online ().

References

McIntosh

, Gonzalez-Lima

(1994) Network interactions among limbic cortices, basal forebrain, and cerebellum differentiate a tone conditioned as a Pavlovian excitor or inhibitor: Fluorodeoxyglucose mapping and covariance structural modeling. J Neurophysiol 72, 1717–1733.

Horwitz

, Warner

, Fitzer

, Tagamets

M-A

, Husain

, Long

(2005) Investigating the neural basis for functional and effective connectivity. Application to fMRI. Philos Trans R Soc Lond B Biol Sci 360, 1093–1108.

Wang

, Wang

, Zhang

, Mchugh

, Sun

, Li

, Yang

(2015) Interhemispheric functional and structural disconnection in Alzheimer’s disease: A combined resting-state fMRI and DTI study. PLoS One 10, e0126310.

Brier

, Thomas

, Ances

(2014) Network dysfunction in Alzheimer’s disease: Refining the disconnection hypothesis. Brain Connect 4, 299–311.

Bozzali

, Padovani

, Caltagirone

, Borroni

(2011) Regional grey matter loss and brain disconnection across Alzheimer disease evolution. Curr Med Chem 18, 2452–2458.

Drzezga

, Becker

, Van Dijk

KRA

, Sreenivasan

, Talukdar

, Sullivan

, Schultz

, Sepulcre

, Putcha

, Greve

, Johnson

, Sperling

(2011) Neuronal dysfunction and disconnection of cortical hubs in non-demented subjects with elevated amyloid burden. Brain 134, 1635–1646.

Bokde

, Ewers

, Hampel

(2009) Assessing neuronal networks: Understanding Alzheimer’s disease. Prog Neurobiol 89, 125–133.

Delbeuck

, Van der Linden

, Collette

(2003) Alzheimer’s disease as a disconnection syndrome? Neuropsychol Rev 13, 79–92.

Terry

, Masliah

, Salmon

, Butters

, DeTeresa

, Hill

, Hansen

, Katzman

(1991) Physical basis of cognitive alterations in Alzheimer’s disease: Synapse loss is the major correlate of cognitive impairment. Ann Neurol 30, 572–580.

10.

Smith

(2002) Imaging the progression of Alzheimer pathology through the brain. Proc Natl Acad Sci U S A 99, 4135–4137.

11.

Lacalle-Aurioles

, Mateos-Pérez

, Guzmán-De-Villoria

, Olazarán

, Cruz-Orduña

, Alemán-Gómez

, Martino

M-E

, Desco

(2014) Cerebral blood flow is an earlier indicator of perfusion abnormalities than cerebral blood volume in Alzheimer’s disease. J Cereb Blood Flow Metab 34, 654–659.

12.

Vemuri

, Gunter

, Senjem

, Whitwell

, Kantarci

, Knopman

, Boeve

, Petersen

, Jack

Jr (2008) Alzheimer’s disease diagnosis in individual subjects using structural MR images: Validation studies. Neuroimage 39, 1186–1197.

13.

Vemuri

, Wiste

, Weigand

, Knopman

, Trojanowski

, Shaw

, Bernstein

, Aisen

, Weiner

, Petersen

, Jack

Jr (2010) Serial MRI and CSF biomarkers in normal aging, MCI, and AD. Neurology 75, 143–151.

14.

Runtti

, Mattila

, van

Gils M

, Koikkalainen

, Soininen

, Lötjönen

(2014) Quantitative evaluation of disease progression in a longitudinal mild cognitive impairment cohort. J Alzheimers Dis 39, 49–61.

15.

Frisoni

, Fox

, Jack

, Scheltens

, Thompson

(2010) The clinical use of structural MRI in Alzheimer disease. Nat Rev Neurol 6, 67–77.

16.

Bobinski

, de Leon

, Wegiel

, Desanti

, Convit

, Saint Louis

, Rusinek

, Wisniewski

(2000) The histological validation of post mortem magnetic resonance imaging-determined hippocampal volume in Alzheimer’s disease. Neuroscience 95, 721–725.

17.

Gosche

, Mortimer

, Smith

, Markesbery

, Snowdon

(2002) Hippocampal volume as an index of Alzheimer neuropathology: Findings from the Nun Study. Neurology 58, 1476–1482.

18.

Gordon

, Robert

, Andrew

, Camper

, Tammy

, Deborah

, Yurgelun

, Perry

(1998) Dynamic susceptibility contrast MS imaging of regional cerebral blood volume in Alzheimer disease: A promising alternative to nuclear medicine. Am J Neuroradiol 19, 1727–1732.

19.

Bammer

, Skare

, Newbould

, Liu

, Thijs

, Ropele

, Clayton

, Krueger

, Moseley

, Glover

(2005) Foundations of advanced magnetic resonance imaging. NeuroRx 2, 167–196.

20.

Medina

, Gaviria

(2008) Diffusion tensor imaging investigations in Alzheimer’s disease: The resurgence of white matter compromise in the cortical dysfunction of the aging brain. Neuropsychiatr Dis Treat 4, 737–742.

21.

Zhang

, Xu

, Zhu

, Kantarci

(2014) The role of diffusion tensor imaging in detecting microstructural changes in prodromal Alzheimer’s disease. CNS Neurosci Ther 20, 3–9.

22.

Beaulieu

(2002) The basis of anisotropic water diffusion in the nervous system - A technical review. NMR Biomed 15, 435–455.

23.

Basser

, Pierpaoli

(1996) Microstructural and physiological features of tissues elucidated by quantitative-diffusion-tensor MRI. J Magn Reson B 111, 209–219.

24.

Mori

, Zhang

(2006) Principles of diffusion tensor imaging and its applications to basic neuroscience research. Neuron 51, 527–539.

25.

Xekardaki

, Giannakopoulos

, Haller

(2011) White matter changes in bipolar disorder, Alzheimer disease, and mild cognitive impairment: New insights from DTI. J Aging Res 2011, 286564.

26.

Zarei

, Damoiseaux

, Morgese

, Beckmann

, Smith

, Matthews

, Scheltens

, Rombouts

SARB

, Barkhof

(2009) Regional white matter integrity differentiates between vascular dementia and Alzheimer disease. Stroke 40, 773–779.

27.

Zimny

, Szewczyk

, Trypka

, Wojtynska

, Noga

, Leszek

, Sasiadek

(2011) Multimodal imaging in diagnosis of Alzheimer’s disease and amnestic mild cognitive impairment: Value of magnetic resonance spectroscopy, perfusion, and diffusion tensor imaging of the posterior cingulate region. J Alzheimers Dis 27, 591–601.

28.

Zimny

, Bladowska

, Neska

, Petryszyn

, Guziński

, Szewczyk

, Leszek

, Sąsiadek

(2013) Quantitative MR evaluation of atrophy, as well as perfusion and diffusion alterations within hippocampi in patients with Alzheimer’s disease and mild cognitive impairment. Med Sci Monit 19, 86–94.

29.

Winblad

, Palmer

, Kivipelto

, Jelic

, Fratiglioni

, Wahlund

, Nordberg

, Backman

, Albert

, Almkvist

, Arai

, Basun

, Blennow

, de Leon

, DeCarli

, Erkinjuntti

, Giacobini

, Graff

, Hardy

, Jack

, Jorm

, Ritchie

, van Duijn

, Visser

, Petersen

(2004) Mild cognitive impairment–beyond controversies, towards a consensus: Report of the International Working Group on Mild Cognitive Impairment. J Intern Med 256, 240–246.

30.

McKhann

, Knopman

, Chertkow

, Hyman

, Jack

, Kawas

, Klunk

, Koroshetz

, Manly

, Mayeux

, Mohs

, Morris

, Rossor

, Scheltens

, Carrillo

, Thies

, Weintraub

, Phelps

(2011) The diagnosis of dementia 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, 263–269.

31.

Fischl

, Dale

(2000) Measuring the thickness of the human cerebral cortex from magnetic resonance images. Proc Natl Acad Sci U S A 97, 11050–11055.

32.

Desikan

, Segonne

, Fischl

, Quinn

, Dickerson

, Blacker

, Buckner

, Dale

, Maguire

, Hyman

, Albert

, Killiany

(2006) An automated labeling system for subdividing the human cerebral cortex on MRI scans into gyral based regions of interest. Neuroimage 31, 968–980.

33.

Pascau

, Gispert

, Michaelides

, Thanos

, Volkow

, Vaquero

, Soto-Montenegro

, Desco

Automated method for small-animal PET image registration with intrinsic validation. Mol Imaging Biol 11, 107–113.

34.

Gordaliza

, Mateos-Pérez

, Montesinos

, Guzmán-de-Villoria

, Desco

, Vaquero

(2015) Development and validation of an open source quantification tool for DSC-MRI studies. Comput Biol Med 58, 56–62.

35.

Smith

, Jenkinson

, Woolrich

, Beckmann

, Behrens

, Johansen-Berg

, Bannister

, De Luca

, Drobnjak

, Flitney

, Niazy

, Saunders

, Vickers

, Zhang

, De Stefano

, Brady

, Matthews

(2004) Advances in functional and structural MR image analysis and implementation as FSL. Neuroimage 23(Suppl 1), S208–S219.

36.

Smith

(2002) Fast robust automated brain extraction. Hum Brain Mapp 17, 143–155.

37.

Behrens

, Woolrich

, Jenkinson

, Johansen-Berg

, Nunes

, Clare

, Matthews

, Brady

, Smith

(2003) Characterization and propagation of uncertainty in diffusion-weighted MR imaging. Magn Reson Med 50, 1077–1088.

38.

Mori

, Oishi

, Jiang

, Li

, Akhter

, Hua

, Faria

, Mahmood

, Woods

, Toga

, Pike

, Neto

, Evans

, Zhang

, Huang

, Miller

, van Zijl

, Mazziotta

(2008) Stereotaxic white matter atlas based on diffusion tensor imaging in an ICBM template. Neuroimage 40, 570–582.

39.

Wakana

, Jiang

, Nagae-Poetscher

, van Zijl

, Mori

(2004) Fiber tract-based atlas of human white matter anatomy. Radiology 230, 77–87.

40.

Navas-Sánchez

, Alemán-Gómez

, Sánchez-Gonzalez

, Guzmán-De-Villoria

, Franco

, Robles

, Arango

, Desco

(2014) White matter microstructure correlates of mathematical giftedness and intelligence quotient. Hum Brain Mapp 35, 2619–2631.

41.

Avants

, Epstein

, Grossman

, Gee

(2008) Symmetric diffeomorphic image registration with cross-correlation: Evaluating automated labeling of elderly and neurodegenerative brain. Med Image Anal 12, 26–41.

42.

Iturria-Medina

, Canales-Rodríguez

, Melie-García

, Valdés-Hernández

, Martínez-Montes

, Alemán-Gómez

, Sánchez-Bornot

(2007) Characterizing brain anatomical connections using diffusion weighted MRI and graph theory. Neuroimage 36, 645–660.

43.

Korczyn

, Vakhapova

(2007) The prevention of the dementia epidemic. J Neurol Sci 257, 2–4.

44.

Henry-Feugeas

(2009) Assessing cerebrovascular contribution to late dementia of the Alzheimer’s type: The role of combined hemodynamic and structural MR analysis. J Neurol Sci 283, 44–48.

45.

Román

, Pascual

(2012) Contribution of neuroimaging to the diagnosis of Alzheimer’s disease and vascular dementia. Arch Med Res 43, 671–676.

46.

Fellgiebel

, Schermuly

, Gerhard

, Keller

, Albrecht

, Weibrich

, Muller

, Stoeter

(2008) Functional relevant loss of long association fibre tracts integrity in early Alzheimer’s disease. Neuropsychologia 46, 1698–1706.

47.

Mielke

, Kozauer

, Chan

KCG

, George

, Toroney

, Zerrate

, Bandeen-Roche

, Wang

M-C

, Vanzijl

, Pekar

, Mori

, Lyketsos

, Albert

(2009) Regionally-specific diffusion tensor imaging in mild cognitive impairment and Alzheimer’s disease. Neuroimage 46, 47–55.

48.

Zhang

, Schuff

, Jahng

G-H

, Bayne

, Mori

, Schad

, Mueller

, Du

A-T

, Kramer

, Yaffe

, Chui

, Jagust

, Miller

, Weiner

(2007) Diffusion tensor imaging of cingulum fibers in mild cognitive impairment and Alzheimer disease. Neurology 68, 13–19.

49.

Teipel

, Stahl

, Dietrich

, Schoenberg

, Perneczky

, Bokde

, Reiser

, Moller

, Hampel

(2007) Multivariate network analysis of fiber tract integrity in Alzheimer’s disease. Neuroimage 34, 985–995.

50.

Rose

, McMahon

, Janke

, O’Dowd

, de Zubicaray

, Strudwick

, Chalk

(2006) Diffusion indices on magnetic resonance imaging and neuropsychological performance in amnestic mild cognitive impairment. J Neurol Neurosurg Psychiatry 77, 1122–1128.

51.

Rose

, Chen

, Chalk

, Zelaya

, Strugnell

, Benson

, Semple

, Doddrell

(2000) Loss of connectivity in Alzheimer’s disease: An evaluation of white matter tract integrity with colour coded MR diffusion tensor imaging. J Neurol Neurosurg Psychiatry 69, 528–530.

52.

Jobst

, Smith

, Barker

, Wear

, King

, Smith

, Anslow

, Molyneux

, Shepstone

, Soper

, et al (1992) Association of atrophy of the medial temporal lobe with reduced blood flow in the posterior parietotemporal cortex in patients with a clinical and pathological diagnosis of Alzheimer’s disease. J Neurol Neurosurg Psychiatry 55, 190–194.

53.

Malhotra

, Coulthard

, Husain

(2009) Role of right posterior parietal cortex in maintaining attention to spatial locations over time. Brain 132, 645–660.

54.

Vincent

, Snyder

, Fox

, Shannon

, Andrews

, Raichle

, Buckner

(2006) Coherent spontaneous activity identifies a hippocampal-parietal memory network. J Neurophysiol 96, 3517–3531.

55.

Fletcher

, Carmichael

, Pasternak

, Maier-Hein

, DeCarli

(2014) Early brain loss in circuits affected by Alzheimer’s disease is predicted by fornix microstructure but may be independent of gray matter. Front Aging Neurosci 6, 106.

56.

Bell

(2012) The imbalance of vascular molecules in Alzheimer’s disease. J Alzheimers Dis 32, 699–709.

57.

Altman

, Rutledge

(2010) The vascular contribution to Alzheimer’s disease. Clin Sci 119, 407–421.

58.

Zlokovic

(2011) Neurovascular pathways to neurodegeneration in Alzheimer’s disease and other disorders. Nat Rev Neurosci 12, 723–738.

59.

Lecrux

, Hamel

(2011) The neurovascular unit in brain function and disease. Acta Physiol 203, 47–59.

60.

Winkler

, Sagare

, Zlokovic

(2014) The pericyte: A forgotten cell type with important implications for Alzheimer’s disease? Brain Pathol 24, 371–386.

61.

Palesi

, Vitali

, Chiarati

, Castellazzi

, Caverzasi

, Pichiecchio

, Colli-Tibaldi

, D’Amore

, D’Errico

, Sinforiani

, Bastianello

(2012) DTI and MR volumetry of hippocampus-PC/PCC circuit: In search of early micro- and macrostructural signs of Alzheimer’s disease. Neurol Res Int 2012, 517876.

62.

Canu

, McLaren

, Fitzgerald

, Bendlin

, Zoccatelli

, Alessandrini

, Pizzini

, Ricciardi

, Beltramello

, Johnson

, Frisoni

(2010) Microstructural diffusion changes are independent of macrostructural volume loss in moderate to severe Alzheimer’s disease. J Alzheimers Dis 19, 963–976.

63.

Villain

, Fouquet

, Baron

J-C

, Mézenge

, Landeau

, de La Sayette

, Viader

, Eustache

, Desgranges

, Chételat

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

64.

Wisse

LEM

, Reijmer

, ter Telgte

, Kuijf

, Leemans

, Luijten

, Koek

, Geerlings

, Biessels

(2015) Hippocampal disconnection in early Alzheimer’s disease: A 7 tesla MRI study. J Alzheimers Dis 45, 1247–1256.

65.

Baglio

, Saresella

, Preti

, Cabinio

, Griffanti

, Marventano

, Piancone

, Calabrese

, Nemni

, Clerici

(2013) Neuroinflammation and brain functional disconnection in Alzheimer’s disease. Front Aging Neurosci 5, 81.

66.

Bartzokis

(2011) Alzheimer’s disease as homeostatic responses to age-related myelin breakdown. Neurobiol Aging 32, 1341–1371.

67.

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.

68.

Liu

, Spulber

, Lehtimaki

, Kononen

, Hallikainen

, Grohn

, Kivipelto

, Hallikainen

, Vanninen

, Soininen

(2011) Diffusion tensor imaging and tract-based spatial statistics in Alzheimer’s disease and mild cognitive impairment. Neurobiol Aging 32, 1558–1571.

69.

Zhang

, Schuff

, Du

A-T

, Rosen

, Kramer

, Gorno-Tempini

, Miller

, Weiner

(2009) White matter damage in frontotemporal dementia and Alzheimer’s disease measured by diffusion MRI. Brain 132, 2579–2592.

70.

Alves

, O’Dwyer

, Jurcoane

, Oertel-Knöchel

, Knöchel

, Prvulovic

, Sudo

, Alves

, Valente

, Moreira

, Fuβer

, Karakaya

, Pantel

, Engelhardt

, Laks

(2012) Different patterns of white matter degeneration using multiple diffusion indices and volumetric data in mild cognitive impairment and Alzheimer patients. PLoS One 7, e52859.

71.

Smith

, Jenkinson

, Johansen-Berg

, Rueckert

, Nichols

, Mackay

, Watkins

, Ciccarelli

, Cader

, Matthews

, Behrens

TEJ

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

72.

Liu

, Zhu

, Marks

, Katz

, Goodlett

, Gerig

, Styner

(2009) Voxel-wise group analysis of DTI. Proc IEEE Int Symp Biomed Imaging, pp. 807–810.

73.

Dubois

, Feldman

, Jacova

, Cummings

, Dekosky

, Barberger-Gateau

, Delacourte

, Frisoni

, Fox

, Galasko

, Gauthier

, Hampel

, Jicha

, Meguro

, O’Brien

, Pasquier

, Robert

, Rossor

, Salloway

, Sarazin

, de Souza

, Stern

, Visser

, Scheltens

(2010) Revising the definition of Alzheimer’s disease: A new lexicon. Lancet Neurol 9, 1118–1127.