Sage Journals: Discover world-class research

Abstract

To investigate putative interacting or distinct pathways for hippocampal complex substructure (HCS) atrophy and cognitive affection in early-stage Alzheimer's disease (AD) and cerebrovascular disease (CVD), we recruited healthy controls, patients with mild cognitive impairment (MCI) and poststroke patients. HCSs were segmented, and quantitative white-matter hyperintensity (WMH) load and cerebrospinal fluid (CSF) amyloid-β concentrations were determined. The WMH load was higher poststroke. All examined HCSs were smaller in amyloid-positive MCI than in controls, and the subicular regions were smaller poststroke. Memory was reduced in amyloid-positive MCI, and psychomotor speed and executive function were reduced in poststroke and amyloid-positive MCI. Size of several HCS correlated with WMH load poststroke and with CSF amyloid-β concentrations in MCI. In poststroke and amyloid-positive MCI, neuropsychological function correlated with WMH load and hippocampal volume. There are similar patterns of HCS atrophy in CVD and early-stage AD, but different HCS associations with WMH and CSF biomarkers. WMHs add to hippocampal atrophy and the archetypal AD deficit delayed recall. In line with mounting evidence of a mechanistic link between primary AD pathology and CVD, these additive effects suggest interacting pathologic processes.

Keywords

Alzheimer's atherosclerosis MRI vascular cognitive impairment white-matter disease

INTRODUCTION

Traditionally, cerebrovascular disease (CVD) has been an exclusion criterion for Alzheimer's disease (AD), but today there is strong evidence for a connection between AD and CVD,¹ and there are indications of a direct mechanistic and reciprocal link between cerebral vascular dysfunction and AD-linked neurodegeneration. Cerebral microvasculature in AD is consistently shown distorted with endothelial atrophy and pericyte degeneration,² and similar capillary affection is shown in most conditions predisposing to stroke.³ The presence of large-vessel cerebral atherosclerosis is associated with amyloid plaques postmortem,⁴ and the neurovascular unit and blood–brain barrier are disrupted in both CVD and AD, leading in turn to neurochemical imbalance.⁵

White-matter hyperintensities (WMHs) are characterized by demyelination and axonal rarefaction, and are used as surrogate markers of small vessel CVD.⁶ Though associated with lower levels of cerebrospinal fluid (CSF) amyloid β 1–42 (Aβ42) (and other amyloid precursor protein metabolites),⁷ hypoxia is known to upregulate β-secretase activity and is associated with amyloid deposition and neuritic plaque formation.⁸ Hippocampal atrophy in CVD is well described in patients with and without evidence of amyloid plaques,⁹ and up to 50% of poststroke patients will develop dementia.¹⁰

The hippocampus consists of several histologically heterogeneous subregions: the cornu ammonis (CA) comprising sectors 1 to 4 and the dentate gyrus (DG). AD typically involves atrophy of the hippocampal complex (the hippocampus and interconnected subhippocampal entorhinal (ERC) and perirhinal (PRC) cortices), and recent reports suggest selective and consecutive vulnerability of pertinent hippocampal complex substructures (HCSs) and cerebral cognitive networks.¹¹ Substantial atrophy of ERC layer II is seen in very mild AD,¹² and ensuing degeneration of the perforant pathway (originating in this layer and a major hippocampal input for memory processing) is important in hippocampal neocortical deafferentation.¹³ ERC atrophy is also a predictor of deterioration from mild cognitive impairment (MCI, a heterogeneous clinical condition associated with an increased risk of dementia¹⁴) to Alzheimer's dementia,¹⁵ but whether neurons within the HCS are similarly affected in CVD and early-stage AD remains an open question.

Early-stage neurofibrillary pathology is seen in the medial temporal lobe, first affecting the PRC, then the ERC,¹⁶ before moving on to the hippocampus (first appearing in CA1). Both CSF total microtubule-associated protein tau (T-tau) and phosphorylated microtubule-associated protein tau (P-tau) are established biomarkers of AD neurofibrillary pathology, but T-tau is an unspecific marker of neurite damage (and increased in both AD and poststroke), whereas increased P-tau is considered AD specific.¹⁷

To reveal putative interacting or distinct pathways for HCS atrophy and cognitive affection in predementia AD and CVD, we included healthy controls, poststroke cases, and patients with MCI (stratified according to CSF Aβ42 levels into amyloid-positive MCI and all-cause MCI) to answer the following research questions (RQ):

RQ 1: Are there differences in HCS size between the three groups?

RQ 2: Is WMH load related to HCS atrophy?

RQ 3: Are the CSF biomarkers (Aβ42, T-tau, and P-tau) related to HCS size?

RQ 4: Are measures of neuropsychological function related to WMH load and total hippocampal volume?

MATERIALS AND METHODS

Eligibility Criteria, Measures of Cognitive Impairment, and Ethical Conduct

Patients with MCI were recruited from a university-hospital based memory clinic. Spouses of participating patients were potentially eligible as controls provided clinically established normality with regard to memory, emotionality, and tempo. The inclusion procedure adheres to the Petersen criteria and is extensively documented elsewhere.¹⁸

Stroke patients were recruited from a university-hospital based stroke unit. Inclusion criteria for these patients were cortical and lacunar ischemic infarctions, between 40 and 79 years of age, and mini-mental state examination¹⁹ (MMSE) score ≥ 23, no severe problems of language, and visual/auditory neglect. Exclusion criteria were a history of somatic or psychiatric disorder including depression. Only patients who agreed to undergo CSF sampling were included. Thirty-seven poststroke subjects fulfilled these criteria, but ten were excluded due to pathologic CSF Aβ42 (see Cerebrospinal Fluid Biomarkers). MCI is a heterogeneous condition, and subtypes exhibit differing patterns of atrophy.²⁰ This underlines the importance of excluding subjects with non-AD MCI. Thirty MCI subjects with normal CSF Aβ42 were therefore introduced only in RQ 3 (Are the CSF biomarkers related to HCS size?), as explained under Statistical analyses RQ3 (Supplementary Tables 2 and 3).

To obtain CSF from controls, a secondary control group was recruited before lumbar puncture in conjunction with orthopedic surgery. Thirty-four of these subjects assented and successfully underwent magnetic resonance imaging (MRI) (on a different scanner). Normality with regard to memory, emotionality, and tempo (as with the primary control group) was established in a subsequent clinical interview. Three subjects were excluded due to pathologic CSF Aβ42, and five subjects were excluded due to cognitive complaints, leaving twenty-six subjects for the secondary control group.

The patient populations and the primary control group were directly comparable, but CSF was not available for the primary control group (Supplementary Tables 1 and 2).

All subjects gave their written consent, and the Regional Committee for Medical and Health Research Ethic South-East evaluated (based on the Norwegian Health and Research Act and the Helsinki Declaration of 1964; revised 2013) and approved the study. All further study conduct was based on and in line with these guidelines.

Magnetic Resonance Imaging

For the primary control group, the MCI and stroke groups, MRI scans were obtained from a Siemens Espree 1.5 T system (Siemens Medical Solutions, Erlangen, Germany), one MPRAGE 3D sequence was acquired (repetition time/echo time/inversion time (TI)/flip angle (FA) = 2,400/3.65/1,000/8°, matrix = 240 × 192), 160 sagittal slices, thickness = 1.2 mm, in-plane resolution of 1 mm × 1.2 mm. The protocol also included 2D axial FLAIR (fluid-attenuated inversion recovery) images with the following parameters: repetition time/echo time/inversion time (TI) = 13,420/121/2,500, 36 slices, spaced at 3.0 mm and 3.9 mm thick.

For the secondary control group, MRI scans were acquired on a Philips Achieva 3 T system (Philips Medical Systems, Best, The Netherlands). At 3 T, a single 3D turbo field echo sequence was acquired for morphometric analysis with the following sequence parameters: repetition time/echo time/inversion time (TI)/flip angle (FA) = 4.5 ms/2.2 ms/853 ms/8°, matrix = 256 × 213, 170 slices, thickness = 1.2 mm, in-plane resolution of 1 mm × 1.2 mm.

Magnetic resonance imaging was performed consecutively after inclusion in the control and MCI groups, and 3 months after stroke. Stroke location, type, and size were determined (Supplementary Table 3).

Table 1.

Population characteristics

Mean/range/standard deviation (SD) unless otherwise stated	Primary control group, N = 41	Amyloid-positive MCI, N = 28	Poststroke, N = 27
Age; mean (range)^a	65.3 (45–84)	66.1 (5–77)	63.7 (48–79)
Men/Women^a	19/22	11/17	21/6
WMH (mm³; median/range/SD)^c	1151/0–23259/4955	1968/70–36224/8313	4966/247–66208/13458
WMH assessment valid (missing)	41 (0)	26 (2)	27 (0)
Stroke type cortical/lacunary	Not applicable	Not applicable	12/15
Cerebrospinal fluid biomarkers
Data missing	No data available	0%	0%
Aβ42 median/ranged^d	Not applicable	540/221–678	843/720–1515
p-Tau median/range	Not applicable	88/39–238	62/41–97
t-Tau median/range	Not applicable	488/141–1407	518/181–5778
Morphometric data
Missing data	0%	0%	0%
Hippoampus; rV^e,f	5.01/3.69–6.09/0.60	4.34/3.21–5.86/0.72	4.73/3.18–5.89/0.65
Presubiculum; rV^g	0.57/0.44–0.81/0.10	0.48/0.29–0.69/0.10	0.51/0.32–0.64/0.08
Subiculum; rV^h	0.81/0.62–1.04/0.10	0.70/0.47–1.10/0.13	0.74/0.48–0.93/0.11
CA1; rVⁱ	0.44/0.36–0.51/0.04	0.41/0.29–0.49/0.05	0.42/0.33–0.54/0.05
CA2/3; rV^j	1.22/0.98–1.49/0.13	1.07/0.73–1.32/0.16	1.2/0.89–1.42/0.13
CA4/DG; rV^k	0.69/0.54–0.85/0.08	0.61/0.41–0.79/0.09	0.66/0.50–0.86/0.08
Perirhinal cortex; mT^l,m	3.32/2.68–3.72/0.25	3.0/2.02–3.68/0.39	3.21/2.68–3.86/0.30
Entorhinal cortex; mTⁿ	3.44/2.56–3.86/0.25	3.13/2.37–3.72/0.35	3.47/2.92–3.92/0.22

Abbreviations: Aβ42, amyloid β42; CA, cornu ammonis; CSF, cerebrospinal fluid; DG, dentate gyrus; MCI, mild cognitive impairment. Figures in bold are significantly different from controls, underlined figures are significantly different from MCI.

Analyses of variance do not show significant differences in age between groups; F(2,93) = 0.705, P = 0.497.

Chi-square tests do not show any significant difference in gender composition within the control (χ² = 0.220, df = 1, P = 0.639) or MCI (χ² = 1.286, df = 1, P = 0.257) groups, but there are significantly more men in the poststroke group (χ² = 8.333, df = 1, P = 0.004).

WMHs = White matter hyperintensities. There were no significant differences in WMH load between controls and MCI (Mann–Whitney U = 403, P = 0.94). There were significantly more WMH in poststroke than in controls (Mann–Whitney U = 328, P = 0.005), but not in poststroke than in MCI (Mann–Whitney U = 287, P = 0.255).

By design, there was a difference between MCI and poststroke in CSF Aβ42. CSF p-Tau was significantly higher in MCI than in poststroke (Mann–Whitney U = 190, P = 0.002), whereas t-Tau was not significantly different (Mann–Whitney U = 311, P = 0.259).

rV = relative volume; i.e., volume of relevant structure in both hemispheres in %o of intracranial volume. Morphometric group differences have been assessed using ANOVA with Dunnett one-way post hoc group differences between HC and the patient groups. Dunnett two-way post hoc was also performed to assess differences between MCI and poststroke.

F(2,93) = 12.2, P < 0.001; controls > MCI, P > 0.001; controls > poststroke, P = 0.124.

F(2,93) = 9.01, P < 0.001; controls > MCI, P < 0.001; controls > poststroke, P = 0.012.

F(2,93) = 9.83, P < 0.001; controls > MCI, P < 0.001; controls > poststroke, P = 0.043.

F(2,93) = 3.86, P = 0.025; controls > MCI, P = 0.007; controls > poststroke, P = 0.376.

F(2,93) = 11.5, P < 0.001; controls > MCI, P < 0.001; controls > poststroke, P = 0.194.

F(2,93) = 10.4, P < 0.001; controls > MCI, P < 0.001; controls > poststroke, P = 0.252.

mT = mean thickness in mm; left and right hemisphere averaged. Entorhinal and perirhinal data missing for one healthy control.

F(2,92) = 6.87, P = 0.002; controls > MCI, P < 0.001; HC > poststroke, P = 0.178.

F(2,92) = 10.8, P = < 0.001; controls > MCI, P < 0.001; HC > poststroke, P = 0.550.

Table 2.

Population neuropsychological characteristics

Variables (median/range/SD)	Controls, N = 41	Amyloid-positive MCI, N = 28	Poststroke, N = 27
General screening test
MMSE	30.0/28–30/0.6	28.0/25–29/1.1 ^a	29.0/26–30/1.2 ^b
Neuropsychology (missing)	32 (9)	27 (1)	26 (1)
Memory
Word list memory; immediate recall (RAVLT)	49.5/32–63/8.8	35.0/18–60/12.2 ^c	46.0/21–63/12.2 ^d
Word list memory; delayed recall (RAVLT)	51.0/33–75/10.8	34.0/21–62/12.9 ^e	42.5/27–75/13.7 ^f
Story memory; immediate recall (WMS-R)	Not available	45.6/31–62/9.58	57.1/34–73/8.6 ^g
Story memory; delayed recall (WMS-R)	Not available	45.6/27–68/12.0	54.7/38–73/9.4 ^h
Psychomotor speed
D-KEFS Stroop 1 (Color naming)	51.5/43–63/4.5	40.0/0–60/14.2 ⁱ	47.0/20–60/13.2 ^j
D-KEFS Stroop 2 (Word reading)	51.5/43–63/4.8	50.0/10–60/11.6^k	47.0/7–60/14.0 ^l
Digit-symbol (WAIS-III)	Not available	40/23–63/9.32	43/30–60/8.55^m
Executive functions
D-KEFS Stroop 3 (Color-Word response inhibition)	55.0/37–67/7	50.0/0–63/13.7 ⁿ	47.0/20–67/13.4 ^o
D-KEFS Stroop 4 (Color-Word Switching)	57/37–67/6.6	41.5/20–60/12.5 ^p	47.0/20–63/11.0 ^q
Letter-number sequencing (WAIS-III)	50/0–67/11.8	43.0/30–63/8.15 ^r	47.0/30–60/8.11^s

Abbreviations: D-KEFS, Delis-Kaplan Executive Function System; MCI, mild cognitive impairment; MMSE, mini-mental state examination; RAVLT, Rey Auditory Verbal Learning Test; SD, standard deviation; WAIS-III, Wechsler Adult Intelligence Scale – Third Edition; WMS-R, Wechsler Memory Scale – Revised. All neuropsychological variables (except mini-mental status examination) are T scores. Figures in bold are significantly different from controls, underlined figures are significantly different from mild cognitive impairment.

Mann–Whitney U = 52, P < 0.001 compared with controls.

Mann–Whitney U = 240, P = 0.002 compared with controls, Mann–Whitney U = 184, P = 0.004 compared with MCI.

Mann–Whitney U = 146, P < 0.001.

Mann–Whitney U = 336, P = 0.207 comparet with controls; Mann–Whitney U = 184, P = 0.003 compared with MCI.

Mann–Whitney U = 166, P < 0.001 compared with controls.

Mann–Whitney U = 307, P = 0.088 compared with controls; U = 223, P = 0.025 compared with MCI.

Mann–Whitney U = 165, P = 0.002 compared with MCI.

Mann–Whitney U = 166, P = 0.002 compared with MCI.

Mann–Whitney U = 249, P = 0.005 compared with controls.

Mann–Whitney U = 264, P = 0.016 compared with controls; Mann–Whitney U = 336, P = 0.782 compared with MCI.

Mann–Whitney U = 346, P = 0.184 compared with controls.

Mann–Whitney U = 271, P = 0.021 compared with controls; Mann–Whitney U = 299, P = 0.352 compared with MCI.

Mann–Whitney U = 333, P = 0.747 compared with MCI.

Mann–Whitney U = 264, P = 0.010 compared with controls.

Mann–Whitney U = 276, P = 0.027 compared with controls; Mann–Whitney U = 351, P = .993 compared with MCI.

Mann–Whitney U = 120, P < 0.001 compared with controls.

Mann–Whitney U = 231, P = 0.003 compared with controls; Mann–Whitney U = 244, P = 0.083 compared with MCI.

Mann–Whitney U = 260, P = 0.005 compared with controls.

Mann–Whitney U = 328, P = 0.118 compared with controls; Mann–Whitney U = 281, P = 0.206 compared with MCI.

Table 3.

Within group Spearman correlations between WMH load and hippocampal complex structure sizes

	Controls	Amyloid-positive MCI	Poststroke
Hippocampus; rV	rho = −0.269/P = 0.045	rho = −0.132/P = 0.260	rho = −0.455/P = 0.008
Presubiculum; rV	rho = −0.120/P = 0.228	rho = −0.064/P = 0.378	rho = −0.370/P = 0.029
Subiculum; rV	rho = −0.20/P = 0.105	rho = 0.013/P = 0.476	rho = −0.250/P = 0.104
CA1; rV	rho = −0.143/P = 0.186	rho = 0.242/P = 0.116	rho = −0.085/P = 0.337
CA2/3; rV	rho = −0.303/P = 0.027	rho = −0.098/P = 0.317	rho = −0.276/P = 0.082
CA4/DG; rV	rho = −0.274/P = 0.042	rho = −0.110/P = 0.297	rho = −0.225/P = 0.130
Perirhinal cortex; mT	rho = 0.187/P = 0.124	rho = 0.033/P = 0.436	rho = −0.431/P = 0.012
Entorhinal cortex; mT	rho = 0.032/P = 0.421	rho = 0.189/P = 0.177	rho = −0.368/P = 0.030

Abbreviations: CA, cornu ammonis; MCI, mild cognitive impairment; mT, mean thickness; rV, relative volume; WMH, white-matter hyperintensity. Negative correlations are denoted with a minus (−). Significant correlations are shown in bold.

Magnetic Resonance Imaging Segmentations and Analyses

Cortical reconstruction and volumetric segmentation were performed with the FreeSurfer image analysis suite version 5.3.0 (http://surfer.nmr.mgh.harvard.edu/). This includes segmentation of the subcortical white-matter and deep gray-matter volumetric structures²¹ and parcellation of the cortical surface. This labels cortical sulci and gyri, and thickness values are calculated in the regions of interest. Further, an automated procedure for subsegmentation of the hippocampus, as implemented in FreeSurfer, was employed.²² The thickness values of the ERC and the PRC were calculated using methods based on ultra-high resolution ex vivo applied to in vivo MRI, as implemented in FreeSurfer.^13,23

The WMH load was estimated by a recently developed in-house object-based supervised machine learning algorithm for automated segmentation and quantification of WMHs based on FLAIR image intensity and masks of tissue types.²⁴ This provides an automated quantitative assessment of WMH load (Figure 1). All segmentations were manually inspected for accuracy; areas radiologically considered to represent acute stroke were manually masked and excluded from the analyses.

Figure 1.

(A) An example of the hippocampal segmentation in one of the included control subjects. The presubiculum is shown in yellow, the subiculum in green, CA1 in red, and CA2/3 in blue. CA, cornu ammonis. (B) An example of the in-house white-matter hyperintensity (WMH) automated quantitative assessment of WMH load in one of the (high WMH load) stroke subjects.

Cerebrospinal Fluid Biomarkers

Cerebrospinal fluid samples were collected from all patients by lumbar puncture through the L3/L4 or L4/L5 intervertebral space. The lumbar puncture was performed between 0900 and 1200 hours, and consecutively after inclusion in the MCI group, 7 to 10 days after stroke or on the day of inclusion for the secondary control group. Cerebrospinal fluid Aβ42, T-tau, and P-tau were routinely examined. The CSF T-tau level was considered as abnormal if T-tau ≥ 300 ng/L for patients under 50 years, > 450 ng/L for patients from 50 to 69 years, and ≥ 500 ng/L for patients from 70 years and above. Cerebrospinal fluid P-tau was considered as pathologic if ≥ 80 ng/L, and CSF Aβ42 was considered as pathologic if < 700 ng/L.

Neuropsychological Assessment

Neuropsychological examination was performed 3 months after stroke or within 3 months of MRI and CSF sampling in the MCI and control groups. In addition to the general screening test MMSE, tests of memory, processing speed, and executive functions have been administered. The memory tests included Logical Memory Subtest from the Wechsler Memory Scale – Revised (WMS-R)²⁵ and the Rey Auditory Verbal Learning Test (RAVLT).²⁶ The tests of processing speed and executive functions comprise the Digit Symbol subtest of the Wechsler Adult Intelligence Scale – Third Edition (WAIS-III),²⁷ Delis-Kaplan Executive Function System (D-KEFS) Color-Word Interference Test,²⁸ and The Letter-Number Sequencing task from WAIS-III.

Statistical Analyses

For the hippocampus and the hippocampal subregions, volumes from the left and right hemispheres were added, and relative volumes (per mille of total intracranial volume) were computed. For the ERC and the PRC, mean thickness of the left and right hemispheres was computed. Analyses of variance were used to assess differences in age between the groups. Chi-square tests were used to assess gender composition within the groups. The Mann–Whitney test was used to assess group differences in WMH load.

RQ1: Are there differences in hippocampal complex substructure size between the three groups? Linear regression was used to correct for age and gender (across the three groups as a whole), and the residuals were used in further statistical computations. (ERC thickness was squared before regression to fulfill the assumptions of linear regression.) Visual inspection of the histograms and the Kolmogorov–Smirnov test was used to assess whether the resulting residuals were normally distributed. One-way analyses of variance with Dunnett one-way post hoc analyses were used to assess group differences between controls and MCI, and controls and stroke. Dunnett two-way post hoc was also performed to assess differences between MCI and poststroke.

RQ 2: Is white-matter hyperintensity load related to hippocampal complex substructure atrophy? The WMH load was skewed, and (nonparametric) one-way Spearman correlation was therefore used to assess correlations with hippocampal substructure sizes within the three groups separately. The volumetric variables were once again corrected for age and gender, but here this was done within each group.

RQ 3: Are the cerebrospinal fluid biomarkers (Aβ42, T-tau, and P-tau) related to hippocampal complex substructure size? Cerebrospinal fluid was not available for the primary control group; therefore, the secondary control group was used for this RQ. Further, because the amyloid-positive MCI group was already split according to CSF Aβ42 levels (and there are no indications that variations of already below threshold CSF Aβ42 should correlate with HCS size), we chose to use the all-cause MCI subgroup (i.e., MCI not split according to CSF Aβ42 levels) for this particular RQ.

RQ 3 is performed with the secondary control group, the all-cause MCI group, and the poststroke group. Demographic data of the secondary control group and the all-cause MCI subgroup are presented in Supplementary Tables 1 and 2.

Due to skewed CSF biomarker data, one-way Spearman correlation was used to assess relation to HCS size (corrected for age and gender as in RQ2).

RQ 4: Are measures of neuropsychological function related to white-matter hyperintensity load and total hippocampal volume? The WMH load variable was logarithmically transformed, and the resulting WMH variable was then corrected for the effects of age and gender by means of linear regression (separately within each group). (Visual inspection of the residual histograms and the Kolmogorov–Smirnov test confirmed normality.)

T scores, based on norms for each neuropsychological test (except MMSE) were used. Several of the neuropsychological variables were skewed and nonparametric correlation (Spearman) was used to assess correlations to WMH load (corrected for the effects of age and gender as described above), and the neuropsychological variables and relative hippocampal volume (corrected for the effects of age and gender as described in RQ 2).

RESULTS

There were no significant age differences between groups, but there were significantly more men in the poststroke group (Table 1). The WMH load was modest in MCI and controls, and significantly more extensive in stroke. By design, CSF Aβ42 was lower in amyloid-positive MCI than in poststroke. Cerebrospinal fluid P-tau was significantly higher in amyloid-positive MCI than in poststroke, but there were no significant difference in CSF T-tau. On MMSE, controls performed best (median score 30), MCI worse (28), and poststroke in between (29). These differences were statistically significant (Table 2). Both immediate and delayed word list memory were significantly worse in MCI compared with controls and poststroke. Story memory data were not available for controls, but MCI performed worse compared with poststroke. Mild cognitive impairment performed worse than controls on one measure of psychomotor speed whereas poststroke performed worse on two measures. Mild cognitive impairment performed worse than controls on all three measures of executive function, whereas poststroke performed worse on two (Table 2).

RQ 1: Are there Differences in Hippocampal Complex Substructure Size Between the Three Groups?

All the investigated HCSs (relative volumes of the presubiculum, subiculum, CA1, CA2/3, and CA4/DG, and mean thickness of ERC and PRC) was significantly smaller in MCI than in controls (after correction for age and gender). The presubiculum and the subiculum were significantly smaller in the stroke group than in controls (Table 1).

RQ 2: Is White-Matter Hyperintensity Load Related to Hippocampal Complex Substructure Atrophy?

In the stroke group, WMH load correlated with total hippocampal volume and volume of the presubiculum, PRC and ERC thickness. In MCI, WMH load did not correlate with size of any of the hippocampal complex structures, but in controls there were significant correlations with total hippocampal volume, CA2/3, and CA4/DG (Table 3).

RQ 3: Are the Cerebrospinal Fluid Biomarkers (Aβ42, T-tau, and P-tau) Related to Hippocampal Complex Substructure Size?

This RQ was performed in the secondary control group, all-cause MCI group and poststroke groups (as accounted for under Statistical analyses, RQ3). In the secondary control group, there were no significant correlations between any of the CSF biomarkers and the HCSs. In the (all cause) MCI group, hippocampal, presubicular, subicular, CA2/3, and CA4/DG volume correlated significantly with Aβ42 and T-tau. In the stroke group, there was a borderline significant correlation between CSF Aβ42 and total hippocampal volume (Table 4).

Table 4.

Within group Spearman correlations between CSF biomarkers and hippocampal complex structure sizes

Spearman	Secondary control group	All-cause MCI	Poststroke
Hippocampus; rV
Aβ42^a	rho = −0.071/P = 0.365	rho = 0.356/P = 0.006	rho = 0.313/P = 0.056*
p-Tau^b	rho = −0.036/P = 0.430	rho = −0.171/P = 0.196	rho = 0.304/P = 0.061
t-Tau^c	rho = 0.015/P = 0.471	rho = −0.283/P = 0.03	rho = 0.210/P = 0.146
Presubiculum; rV
Aβ42	rho = 0.161/P = 0.217	rho = 0.38/P = 0.003	rho = 0.245/P = 0.109
p-Tau	rho = 0.071/P = 0.364	rho = −0.171/P = 0.195	rho = 0.239/P = 0.115
t-Tau	rho = 0.117/P = 0.285	rho = −0.302/P = 0.02	rho = 0.115/P = 0.283
Subiculum; rV
Aβ42	rho = −0.046/P = 0.411	rho = 0.334/P = 0.010	rho = 0.145/P = 0.235
p-Tau	rho = 0.096/P = 0.320	rho = −0.196/P = 0.137	rho = 0.070/P = 0.365
t-Tau	rho = 0.119/P = 0.282	rho = −0.261/P = 0.046	rho = 0.079/P = 0.347
CA1; rV
Aβ42	rho = −0.078/P = 0.353	rho = 0.118/P = 0.375	rho = 0.237/P = 0.117
p-Tau	rho = 0.133/P = 0.258	rho = −0.199/P = 0.131	rho = −0.026/P = 0.449
t-Tau	rho = 0.001/P = 0.497	rho = −0.192/P = 0.145	rho = 0.111/P = 0.291
CA2/3; rV
Aβ42	rho = −0.068/P = 0.370	rho = 0.292/P = 0.025	rho = 0.145/P = 0.236
p-Tau	rho = 0.019/P = 0.464	rho = −0.197/P = 0.134	rho = 0.062/P = 0.379
t-Tau	rho = 0.076/P = 0.353	rho = −0.269/P = 0.039	rho = 0.059/P = 0.386
CA4/DG; rV
Aβ42	rho = −0.065/P = 0.376	rho = 0.344/P = 0.008	rho = 0.126/P = 0.266
p-Tau	rho = 0.056/P = 0.393	rho = −0.185/P = 0.16	rho = 0.096/P = 0.316
t-Tau	rho = 0.109/P = 0.298	rho = −0.265/P = 0.042	rho = 0.100/P = 0.310
Perirhinal cortex; mT
Aβ42	rho = −0.134/P = 0.257	rho = 0.032/P = 0.810	rho = −0.088/P = 0.275
p-Tau	rho = −0.164/P = 0.211	rho = −0.194/P = 0.142	rho = −0.031/P = 0.438
t-Tau	rho = −0.050/P = 0.405	rho = −0.164/P = 0.215	rho = 0.238/P = 0.116
Entorhinal cortex; mT
Aβ42	rho = −0.262/P = 0.098	rho = 0.242/P = 0.065	rho = 0.120/P = 0.275
p-Tau	rho = −0.324/P = 0.053	rho = −0.209/P = 0.111	rho = 0.209/P = 0.148
t-Tau	rho = −0.241/P = 0.117	rho = −0.209/P = 0.113	rho = 0.262/P = 0.093

Abbreviations: Aβ42, amyloid β 1–42; CA, cornu ammonis; CSF, cerebrospinal fluid; mT, mean thickness; MCI, mild cognitive impairment; rV, relative volume. The analyses in this research question were performed in the secondary control group, the all-cause MCI group, and the poststroke group. Negative correlations are denoted with a minus (−).

Cerebrospinal fluid amyloid β 1–42.

Cerebrospinal fluid total Tau and p-Tau.

Cerebrospinal fluid phosphorylated Tau. Significant correlations are shown in bold.

RQ 4: Are Measures of Neuropsychological Function Related to White-matter Hyperintensity Load and Total Hippocampal Volume?

In controls, there were neither significant correlations between WMH load and measures of neuropsychological function, nor hippocampal volume and measures of neuropsychological function. In MCI, there were correlations between WMH load and word list memory delayed recall, Stroop 1 and Letter-number sequencing. In poststroke, there were several near significant correlations between WMH load and memory. There were significant correlations between Stroops 1, 2, 4 and MMSE. In MCI, there were correlations between hippocampal volume and all tested neuropsychological variables of memory and executive function, but none of psychomotor speed. In CVD, there were correlations between hippocampal volume and word list memory immediate recall, story memory immediate and delayed recall as well as MMSE. Tables 5 and 6 summarize the results of RQ 4.

Table 5.

Within-group Spearman correlations between WMH load (corrected for age and gender) and neuropsychological function

	Controls	Amyloid-positive MCI	Poststroke
Memory
Word list memory; immediate recall (RAVLT)	rho = 0.059/P = 0.373	rho = 0.285/P = 0.083	rho = −0.324/P = 0.053
Word list memory; delayed recall (RAVLT)	rho = 0.019/P = 0.459	rho = 0.532/P = 0.003	rho = −0.182/P = 0.187
Story memory; immediate recall (WMS-R)	Data not available	rho = 0.301/P = 0.077	rho = −0.303/P = 0.066
Story memory; delayed recall (WMS-R)	Data not available	rho = 0.243/P = 0.126	rho = −0.303/P = 0.066
Psychomotor speed
D-KEFS Stroop 1 (Color naming)	rho = 0.051/P = 0.391	rho = −0.464/P = 0.010	rho = −0.359/P = 0.036
D-KEFS Stroop 2 (Word reading)	rho = 0.093/P = 0.307	rho = −0.318/P = 0.061	rho = −0.444/P = 0.011
Digit-symbol (WAIS-III)	Data not available	rho = −0.255/P = 0.110	rho = −0.323/P = 0.054
Executive functions
D-KEFS Stroop 3 (Color-Word response inhibition)	rho = 0.216/P = 0.117	rho = −0.216/P = 0.149	rho = −0.026/P = 0.450
D-KEFS Stroop 4 (Color-Word Switching)	rho = −0.198/P = 0.138	rho = −0.69/P = 0.374	rho = −0.420/P = 0.016
Letter-number sequencing (WAIS-III)	rho = 0.195/P = 0.139	rho = −0.437/P = 0.015	rho = −0.135/P = 0.256
General screening test
MMSE	rho = −0.20/P = 0.128	rho = 0.052/P = 0.402	rho = −0.497/P = 0.006

Abbreviations: D-KEFS, Delis-Kaplan Executive Function System; MCI, mild cognitive impairment; MMSE, mini-mental state examination; RAVLT, Rey Auditory Verbal Learning Test; WAIS-III, Wechsler Adult Intelligence Scale – Third Edition; WMS-R, Wechsler Memory Scale – Revised; WMH, white-matter hyperintensity. Neuropsychological measures (except MMSE) are normed (T scores). Negative correlations are denoted with a minus (−). Significant correlations are shown in bold.

Table 6.

Within-group Spearman correlations between relative hippocampal volume (corrected for age and gender) and T scores of neuropsychological function

	Controls	Amyloid-positive MCI	Poststroke
Memory
Word list memory; immediate recall (RAVLT)	rho = −0.071/P = 0.350	rho = 0.640/P < 0.001	rho = 0.377/P = 0.029
Word list memory; delayed recall (RAVLT)	rho = −0.157/P = 0.196	rho = 0.542/P = 0.002	rho = 0.275/P = 0.087
Story memory; immediate recall (WMS-R)	Data not available	rho = 0.361/P = 0.035	rho = 0.417/P = 0.017
Story memory; delayed recall (WMS-R)	Data not available	rho = 0.548/P = 0.002	rho = 0.578/P = 0.001
Psychomotor speed
D-KEFS Stroop 1 (Color naming)	rho = −0.237/P = 0.096	rho = 0.206/P = 0.151	rho = 0.097/P = 0.318
D-KEFS Stroop 2 (Word reading)	rho = −0.032/P = 0.431	rho = 0.140/P = 0.244	rho = 0.263/P = 0.097
Digit-symbol (WAIS-III)	Data not available	rho = 0.303/P = 0.062	rho = 0.236/P = 0.123
Executive functions
D-KEFS Stroop 3 (Color-Word response inhibition)	rho = −0.054/P = 0.385	rho = 0.422/P = 0.014	rho = 0.045/P = 0.413
D-KEFS Stroop 4 (Color-Word Switching)	rho = 0.032/P = 0.431	rho = 0.529/P = 0.003	rho = 0.168/P = 0.206
Letter-number sequencing (WAIS-III)	rho = −0.112/P = 0.268	rho = 0.346/P = 0.039	rho = 0.094/P = 0.324
General screening test
MMSE	rho = 0.229/P = 0.097	rho = 0.226/P = 0.129	rho = 0.389/P = 0.027

Abbreviations: D-KEFS, Delis-Kaplan Executive Function System; MCI, mild cognitive impairment; MMSE, mini-mental state examination; RAVLT, Rey Auditory Verbal Learning Test; WAIS-III, Wechsler Adult Intelligence Scale – Third Edition; WMS-R, Wechsler Memory Scale – Revised. Neuropsychological measures (except MMSE) are normed (T-scores). Negative correlations are denoted with a minus (−). Significant correlations are shown in bold.

DISCUSSION

To our knowledge, this is the first study to assess the relations between WMHs and HCSs in CVD, compared with relations in amyloid-positive MCI and controls. The main finding of this study is that the hippocampal complex is similarly affected poststroke and in early-stage AD, but there are different associations with WMH and CSF biomarkers. Also, we find a significantly higher WMH load poststroke compared with controls and amyloid-positive MCI. All examined HCSs were smaller in amyloid-positive MCI than in controls, and the presubiculum and the subiculum were smaller in poststroke than in controls. Neuropsychological measures of memory were reduced in amyloid-positive MCI as compared with controls, and measures of psychomotor speed and executive function were reduced in both poststroke and amyloid-positive MCI as compared with controls. Poststroke (but not MCI) WMH load correlated with total hippocampal and size of the presubiculum, PRC and ERC. In both poststroke and amyloid-positive MCI, there were several correlations between measures of neuropsychological function and both WMH load and hippocampal volume. Strikingly, in amyloid-positive MCI, there was a strong correlation between WMH load and the archetypal AD type impairment delayed recall.

The present results support previous reports that both CVD⁹ and AD cause medial temporal lobe atrophy. Although the patterns of medial temporal lobe atrophy in MCI and CVD are similar, there are differential associations with WMHs and CSF biomarkers. In terms of medial temporal atrophy, CVD is important in poststroke patients but has a minor role in amyloid-positive MCI. Vice versa, CSF biomarker pathology is important in MCI, but not in stroke.

The hippocampal complex is vulnerable due to plasticity, neuronal projections, transmitter systems, and oxygen demand. There are high concentrations of NMDA receptors, the presubiculum has consistently been identified with strong acetylcholine staining, and several other diseases (e.g., temporal lobe epilepsy, posttraumatic stress disorder, depression, and schizophrenia) are associated with hippocampal degeneration.²⁹ Neurofibrillary pathology appears early in the hippocampal complex,¹⁶ and soluble amyloid oligomers are neurotoxic.¹⁷ As such, pre-plaque Aβ induced hippocampal complex degeneration is possible (also in the poststroke group). Aβ accumulates around and affects perivascular cells, and amyloid angiopathy is present in a majority of cases. Brain amyloid is in equilibrium with amyloid in CSF and plasma, and can be transported both ways across the blood–brain barrier. Damage to the neurovascular unit and the blood–brain barrier may disrupt this equilibrium.⁵ Putatively (in CVD) this may cause or worsen pre-plaque amyloid dysmetabolism, or (in AD) amyloid dysmetabolism may cause or worsen impairment of cerebrovascular autoregulation; adding to the pathogenetic load. Further research must determine the exact underlying mechanisms and which factors predispose plaque deposition and which factors cause accelerated spreading of neurofibrillary pathology.

In line with the current findings, a recent neuropathological study did not find reduced neuronal volumes in the ERC (layer V) poststroke.³⁰ The current methodology has not allowed for testing the integrity of the perforant path per se, but with ERC atrophy in the MCI group, one should expect degeneration of the perforant path with ensuing partial hippocampal deafferentation. This may contribute to hippocampal subfield atrophy in the MCI group, but in the poststroke group, we see presubicular and subicular atrophy in the absence of discernible amyloid pathology and ERC atrophy, suggesting other mechanisms. We do, however, see a clear correlation between WMH load and PRC and ERC size in the poststroke group, suggesting that these anterior subhippocampal areas are affected in a manner not readily detected by the presently applied methodology. Alternatively, the observed hippocampal complex atrophy in the poststroke cases may be explained by impaired hippocampal microcirculation due to local capillary affection or reduced neocortical hippocampal input due to WMH-associated projection fiber damage not related to ERC atrophy. CA1 is the first hippocampal subfield in which neurofibrillary pathology appears,¹⁶ and several neuropathological (discussed in Lim et al³¹ and La Joie et al³²) studies have shown CA1 atrophy in Alzheimer's dementia, whereas not in preclinical (i.e., until death cognitively intact subjects fulfilling neuropathological criteria) AD.³³ One study even reported hypertrophy of CA1³⁴ in preclinical AD. Morphometric studies have shown that CA1 and subicular atrophy can predict progression to dementia from MCI and the cognitively normal.³⁵ Other studies using the same hippocampal segmentation technique in predementia populations^31,36 did not report CA1 atrophy. The present MCI group differs in being amyloid positive, and it is not surprising to see more extensive atrophy. We do, however, not find CA1 atrophy or a correlation between WMH and CA1 size in the poststroke group. Discernible CA1 atrophy may not occur at the earliest predementia stages, or this may be related to naming conventions and how different segmentation techniques define the boundaries between the different structures; parts of what herein is defined as subiculum in other techniques have been labeled CA1 (as discussed in Lim et al³¹ and La Joie et al³² and below). The presubiculum and the subiculum are nevertheless consistently found to be among the most affected hippocampal subfields in AD.³⁷ Two studies utilizing the same hippocampal segmentation as the present publication on amnestic MCI found subicular and both presubicular and subicular atrophy as compared with controls respectively,^31,36 similar to our findings in poststroke. When comparing neuropathological and imaging findings, it is important to note that the definition of the CA1-subiculum boundary is particularly inconsistent across neuroimaging studies.³² The parcellation procedure used here is among the conservative procedures in estimating CA1 volume. A wedge-shaped portion of the CA1 (often confusingly called the prosubiculum, and known for a particularly high density of neurofibrillary tangles and neuritic plaques) extends into the subiculum,¹⁶ and is likely herein segmented as part of the subiculum. This highlights central problems in MRI studies of the hippocampus and the most important limitation of the present study. Due to limited MRI resolution, it is difficult to accurately map histoanatomic areas, and naming conventions differ in terms of both microstructural histologic delineation and folding patterns. This makes cross-study comparison difficult particularly for hippocampal subsegmentation procedures. There is, however, an extensive effort of harmonizing a manual protocol for hippocampal segmentation that has just been validated and published by EADC-ADNI,³⁸ but there is still much work before a unified parcellation scheme can be agreed on. It should be noted that also manual delineation has its own problems, e.g., the segmentations are typically more consistent in the plane they were drawn and different operators make different but repeated errors. A recent study comparing the reproducibility of hippocampal volume change measurement between manual outlining and common automated methods³⁹ found similar measures of volume change but superior reproducibility in FreeSurfer.

Although all segmentations were manually inspected and the image and segmentation quality and consistency were deemed good, the segmentation procedure was automated and not performed on ultra-high field strength, high-resolution MRI and segmented according to a harmonized protocol. When choosing hippocampal subregions of interest, we only included the larger structures with the best Dice overlap measures with manual delineation.²² Nevertheless, when interpreting the results from the present study it should be noted that the subiculum is most likely overestimated on the expense of CA1, and that there does not exist a generally accepted protocol for subsegmenting the hippocampus.

In poststroke, the anterior part of the hippocampus is smaller than in controls and in particular the anterior subhippocampal areas (ERC and PRC) correlate (inversely) with WMH load. At the present stage, the hippocampal affection is more advanced than ERC and PRC, and there is a borderline significant correlation between delayed recall of story (context-rich) and WMH load, whereas not between WMH load and delayed word list recall (context free). This is in line with the supposition that anterior subhippocampal areas are important in context-free (or episodic) memory whereas the hippocampus is thought to be important in context-rich memory.⁴⁰ The observation that, in the poststroke group, delayed recall of story memory is strongly correlated with total hippocampal volume (with a smaller, only trend significant correlation for delayed word list memory) further underlines this (Table 6).

In MCI, the relatively moderate WMH load is most strongly (inversely) correlated with the early AD-specific context-free word list memory. Even though WMH does not seem to be a major factor in in the majority of MCI cases, its occurrence adds to the archetypal AD type impairment of context-free (episodic) memory. In line with this, a recent study⁹ found that an amyloid-PET positive subgroup of patients with subcortical vascular dementia performed worse on neuropsychological tests and had more extensive hippocampal atrophy (but were also older, mean age 79.5 vs. 71.9 years) than the amyloid-PET negative subgroup.

Although we have included CSF Tau biomarker data also for the stroke group, lumbar puncture was performed at a time when CSF Tau (but not Aβ42) levels are expected to be increased due to acute stroke.¹⁷ Also, it is not known whether very early atrophy presumptively due to neurofibrillary pathology would be reflected in CSF Tau levels.

A limitation that also needs to be addressed is the relatively moderate WMH load in the MCI group as compared with poststroke. It could be construed that with a more extensive WMH load, the effects on the hippocampal complex would be more pronounced. The relative sparseness of WMH in a memory-clinic recruited MCI population is nevertheless important to note, but even though on the same volume level as controls, WMH does affect neuropsychological function in MCI.

In conclusion, we show similar patterns of HCS atrophy in CVD and amyloid-positive MCI. Different associations with WMHs and CSF biomarkers underline the separate etiologies. The effects of WMHs on neuropsychological function are in line with mounting evidence for a mechanistic link between primary AD pathology and small vessel disease (e.g., putatively reciprocal effects of hypoperfusion and AD pathophysiology^2,3 or effects of hypoxia on amyloid metabolism^7,8), strengthening the connection between the two entities. The similarities in patterns of structural affection may, however, limit the applicability of structural scans for differential diagnosis of predementia cognitive affection.

Footnotes

PS contributed to the conception and design of the study, acquisition of data, analyzed the data, and contributed to the interpretation of the data. PS further drafted the article. RG contributed to the conception and design of the study, acquisition and interpretation of data (notably clinical and neuropsychological data), and revised the article. MR contributed to the acquisition and interpretation of data (notably MRI data), and revised the article. AB contributed to the conception and design of the study, acquisition and interpretation of data (notably MRI data), and revised the article. LG contributed to the conception and design of the study, interpretation of data, and revised the article. EH contributed to acquisition and interpretation of data (notably neuropsychological data), and revised the article. EA, KJ, BB, and IA contributed to the acquisition and interpretation of data (notably clinical data), and revised the article. PD-T contributed to the acquisition and interpretation of data, determined stroke location and size and revised the article. KV contributed to the acquisition and interpretation of data (notably MRI data), and revised the article. TF contributed to the conception and design of the study, acquisition and interpretation of data, and revised the article. All authors approved the final version.

MR reports grants from Iceland, Liechtenstein, and Norway through the EEA Financial Mechanism, during the conduct of the study. IA reports grants from South-Eastern Norway Regional Health Authority, grants from Research Council of Norway (NASATS-Nevronor), during the conduct of the study. TF reports grants from the Norwegian Research Council, during the conduct of the study. In addition, TF has a patent PCT/EP2009/001210 issued, and a patent PCT/US2011/062233 pending.

Notes

Supplementary Information accompanies the paper on the Journal of Cerebral Blood Flow & Metabolism website ()

References

Breteler

. Vascular involvement in cognitive decline and dementia. Epidemiologic evidence from the Rotterdam Study and the Rotterdam Scan Study. Ann NY Acad Sci 2000; 903: 457–465.

Østergaard

Aamand

Gutiérrez-Jiménez

Y-CL

Blicher

Madsen

The capillary dysfunction hypothesis of Alzheimer's disease. Neurobiol Aging 2013; 34: 1018–1031.

Østergaard

Jespersen

Mouridsen

The role of the cerebral capillaries in acute ischemic stroke: The extended penumbra model. J Cereb Blood Flow Metab 2013; 33: 635–648.

Honig

Kukull

Mayeux

Atherosclerosis and AD: Analysis of data from the US National Alzheimer's Coordinating Center. Neurology 2005; 64: 494–500.

Iadecola

The overlap between neurodegenerative and vascular factors in the pathogenesis of dementia. Acta Neuropathol 2010; 120: 287–296.

Wallin

Fladby

Do white matter hyperintensities on MRI matter clinically?

BMJ 2010; 341: c3400.

Selnes

Blennow

Zetterberg

Grambaite

Rosengren

Johnsen

Effects of cerebrovascular disease on amyloid precursor protein metabolites in cerebrospinal fluid. Cerebrospinal Fluid Res 2010; 7: 10.

Sun

Qing

Zhou

Dobie

Cai

Hypoxia facilitates Alzheimer's disease pathogenesis by up-regulating BACE1 gene expression. Proc Natl Acad Sci USA 2006; 103: 18727–18732.

Lee

Kim

Seo

Park

Identification of pure subcortical vascular dementia using 11C-Pittsburgh compound B. Neurology 2011; 77: 18–25.

10.

Pendlebury

Rothwell

. Prevalence, incidence, and factors associated with pre-stroke and post-stroke dementia: A systematic review and meta-analysis. Lancet Neurol 2009; 8: 1006–1018.

11.

Seeley

Crawford

Zhou

Miller

Greicius

. Neurodegenerative diseases target large-scale human brain networks. Neuron 2009; 62: 42–52.

12.

Gómez-Isla

Price

McKeel

Morris

Growdon

Hyman

. Profound loss of layer II entorhinal cortex neurons occurs in very mild Alzheimer's disease. J Neurosci 1996; 16: 4491–4500.

13.

Fischl

Stevens

Rajendran

Yeo

BTT

Greve

Van Leemput

Predicting the location of entorhinal cortex from MRI. Neuroimage 2009; 47: 8–17.

14.

Winblad

Palmer

Kivipelto

Jelic

Fratiglioni

Wahlund

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

15.

deToledo-Morrell

Stoub

Bulgakova

Wilson

Bennett

Leurgans

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

16.

Braak

Neuropathological stageing of Alzheimer-related changes. Acta Neuropathol 1991; 82: 239–259.

17.

Blennow

Hampel

Weiner

Zetterberg

Cerebrospinal fluid and plasma biomarkers in Alzheimer disease. Nat Rev Neurol 2010; 6: 131–144.

18.

Selnes

Aarsland

Bjørnerud

Gjerstad

Wallin

Hessen

Diffusion tensor imaging surpasses cerebrospinal fluid as predictor of cognitive decline and medial temporal lobe atrophy in subjective cognitive impairment and mild cognitive impairment. J Alzheimers Dis 2013; 33: 723–736.

19.

Folstein

McHugh

. ‘Mini-mental state’: A practical method for grading the cognitive state of patients for the clinician. J Psychiatr Res 1975; 12: 189–198.

20.

Zhang

Sachdev

Wen

Kochan

Crawford

Brodaty

Gray matter atrophy patterns of mild cognitive impairment subtypes. J Neurol Sci 2012; 315: 26–32.

21.

Fischl

Salat

Busa

Kennedy

Makris

Rosen

Whole brain segmentation: Automated labeling of neuroanatomical structures in the human brain. Neuron 2002; 33: 341–355.

22.

Van Leemput

Bakkour

Benner

Wiggins

Wald

Augustinack

Automated segmentation of hippocampal subfields from ultra-high resolution in vivo MRI. Hippocampus 2009; 19: 549–557.

23.

Augustinack

Huber

Stevens

Roy

Frosch

van der Kouwe

AJW

Predicting the location of human perirhinal cortex, Brodmann's area 35, from MRI. Neuroimage 2013; 64: 32–42.

24.

Rincon

Selnes

Larsson

Fladby

Bjørnerud

Automatic WML segmentation and quantification using a machine learning approach. Proc. Intl. Soc. Mag. Reson. Med. 2011.

25.

Wechsler DWechsler Memory Scale - Revised. Psychological Corporation: New York, 1987.

26.

Schmidt

Rey Auditory Verbal Learning Test. Western Psychological Services: Los Angeles, 1996.

27.

Wechsler

Wechsler Adult Intelligence Scale - Third Edition: Manual. The Psychological Corporation: San Antonio, TX, 2003.

28.

Delis

Kaplan

Kramer

. Delis-Kaplan Executive Function System (D-KEFS). The Psychological Corporation: San Antonio, TX, 2001.

29.

Small

Schobel

Buxton

Witter

Barnes

. A pathophysiological framework of hippocampal dysfunction in ageing and disease. Nat Rev Neurosci 2011; 12: 585–601.

30.

Gemmell

Bosomworth

Allan

Hall

Khundakar

Oakley

Hippocampal neuronal atrophy and cognitive function in delayed poststroke and aging-related dementias. Stroke 2012; 43: 808–814.

31.

Lim

Hong

Jung

Ahn

Won

Hahn

Automated hippocampal subfield segmentation in amnestic mild cognitive impairments. Dement Geriatr Cogn Disord 2012; 33: 327–333.

32.

La Joie

Fouquet

Mézenge

Landeau

Villain

Mevel

Differential effect of age on hippocampal subfields assessed using a new high-resolution 3T MR sequence. Neuroimage 2010; 53: 506–514.

33.

West

Kawas

Stewart

Rudow

Troncoso

. Hippocampal neurons in pre-clinical Alzheimer's disease. Neurobiol Aging 2004; 25: 1205–1212.

34.

Iacono

Markesbery

Gross

Pletnikova

Rudow

Zandi

The Nun study: Clinically silent AD, neuronal hypertrophy, and linguistic skills in early life. Neurology 2009; 73: 665–673.

35.

Apostolova

Mosconi

Thompson

Green

Hwang

Ramirez

Subregional hippocampal atrophy predicts Alzheimer's dementia in the cognitively normal. Neurobiol Aging 2010; 31: 1077–1088.

36.

Hanseeuw

Van Leemput

Kavec

Grandin

Seron

Ivanoiu

Mild cognitive impairment: Differential atrophy in the hippocampal subfields. AJNR Am J Neuroradiol 2011; 32: 1658–1661.

37.

Hyman

Van Hoesen

Damasio

Barnes

. Alzheimer's disease: Cell-specific pathology isolates the hippocampal formation. Science (New York, NY) 1984; 225: 1168–1170.

38.

Boccardi

Bocchetta

Ferrari

Duchesne

Jack

Frisoni

The EADC-ADNI harmonized protocol for hippocampal segmentation on magnetic resonance: Evidence of validity (S48.006). Neurology 2014; 82: S48.006–S48.006.

39.

Mulder

de Jong

Knol

van Schijndel

Cover

Visser

. Hippocampal volume change measurement: Quantitative assessment of the reproducibility of expert manual outlining and the automated methods FreeSurfer and FIRST. Neuroimage 2014; 92: 169–181.

40.

Didic

Barbeau

Felician

Tramoni

Guedj

Poncet

Which memory system is impaired first in Alzheimer's disease?

J Alzheimers Dis 2011; 27: 11–22.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.14 MB

Hippocampal Complex Atrophy in Poststroke and Mild Cognitive Impairment

Abstract

Keywords

INTRODUCTION

MATERIALS AND METHODS

Eligibility Criteria, Measures of Cognitive Impairment, and Ethical Conduct

Magnetic Resonance Imaging

Magnetic Resonance Imaging Segmentations and Analyses

Cerebrospinal Fluid Biomarkers

Neuropsychological Assessment

Statistical Analyses

RESULTS

DISCUSSION

Footnotes

Notes

References

Supplementary Material