Hippocampal Volume and Amyloid PET Status Three Years After Ischemic Stroke: A Pilot Study

Abstract

Hippocampal atrophy is seen in many neurodegenerative disorders and may be a cardinal feature of vascular neurodegeneration. We examined hippocampal volume (HV) in a group of ischemic stroke survivors with amyloid ¹⁸F-NAV4694 PET imaging three years after stroke. We compared HV between the amyloid-positive (n = 4) and amyloid-negative (n = 29) groups, and associations with co-morbidities using Charlson Comorbidity Indices and multi-way ANOVA. Amyloid status was not associated with verbal or visual delayed free recall memory indices or cognitive impairment. We found no association between amyloid status and HV in this group of ischemic stroke survivors.

Keywords

Amyloid imaging atrophy hippocampal volume ischemic stroke PET

INTRODUCTION

Hippocampal atrophy, long thought to be an imaging signature of Alzheimer’s disease (AD), is seen in a range of neurodegenerative disorders. It is also seen in association with vascular risk factors in people with and without dementia [1]. Vascular risk factors and brain injury are associated with neurodegeneration [2 –4], and there is increasing evidence that cognitive performance and brain atrophy may be impacted independently of amyloid status [1, 5]. Amyloid positivity associates with increasing age, and both independent [6] and synergistic [7] effects of ischemic and amyloid burden on cognition have been observed.

People with a diagnosis of subcortical vascular dementia are known to have smaller hippocampi [8], and amyloid positivity and ischemic brain burden appear to have dissociable effects on brain structure and cognitive performance [6]. Amyloid imaging studies in people with ischemic stroke have yielded varied results. Most studies have been done with carbon-11 Pittsburgh compound B positron emission tomography ([¹¹C] PiB-PET) imaging. [¹¹C] PiB uptake has been reported to be increased sub-acutely following stroke, both in peri-infarct regions [9] and in regions often associated with an AD-pattern of distribution [10], such as the precuneus and cingulate cortices. However, in most cross-sectional and longitudinal studies, no sustained increase in amyloid has been found in stroke populations above that expected for age [1 , 11]. [¹¹C] PiB-PET imaging has also been studied extensively in cerebral small vessel disease and subcortical vascular cognitive impairment [12]. However, few researchers have examined the associations between hippocampal volume (HV), a key marker of neurodegeneration, and amyloid status in people following ischemic stroke.

We have previously shown no association between ischemic stroke and increased [¹¹C] PiB-PET amyloid in a separate study [9]. Here, we examined HV and amyloid status using ¹⁸F-NAV4694 amyloid PET imaging in a group of ischemic stroke patients from the Cognition And Neocortical Volume After Stroke (CANVAS) study [13]. Given the complex interac-tions between normal brain aging, vascular risk factors and brain histopathology, we postulated that there would be no association between HV and amyloid status three years after ischemic stroke.

MATERIALS AND METHODS

CANVAS study design

The CANVAS study is an observational cohort study of 135 patients with ischemic stroke and 40 healthy stroke-free controls examining brain volume with MRI scans and cognitive assessments over 3 years [13]. Stroke patients were recruited from the Acute Stroke Units of three metropolitan university hospitals: Austin Health, Box Hill Hospital, and Royal Melbourne Hospital, Victoria, Australia. Adult patients were recruited with ischemic stroke confir-med on clinical imaging of any circulation, type, or etiology. Stroke survivors and control participants could have no history of dementia or any other neuro-degenerative condition. Ethical approval was granted by each hospital’s Human Research Ethics Commit-tee. All participants provided written consent in person in accordance with the Declaration of Helsinki.

Demographic and clinical details

Age, sex, years of education, vascular risk factors (VRFs) and medical co-morbidities were compiled. Charlson Comorbidity Index (CCI) scores were calculated for each participant as a general index of medical comorbidity [14]. The CCI is a validated wei-ghted comorbidity score of medical conditions associated with mortality, with a minimum score of 0 being no disease burden and 29 being maximal dis-ease burden. VRFs included a history of hypertension, hyperlipidemia, smoking, type II diabetes mellitus, and atrial fibrillation. We produced a composite vascular risk score for statistical analysis as previously published [15].

Venous blood was drawn for apolipoprotein E (APOE) genotype determination on participants who consented to DNA analyses and storage. Individuals were categorized as APOE ɛ4 carriers or noncarriers.

Amyloid imaging

Amyloid ¹⁸F-NAV4694 PET imaging was offered as a sub-study at three years for all CANVAS str-oke participants. We did not include this as a manda-tory part of the study due to potential concerns about ionizing radiation [13]. All scans were acquired on an Allegro PET camera (Philips) in 3-dimensional mode and were processed using a rotating ¹³⁷Cs point source for attenuation correction. The participants were injected with 250 MBq of ¹⁸F-NAV4694. Im-ages were reconstructed using a 3-dimensional row-action maximum likelihood algorithm and reported by a nuclear medicine physician as per a published centiloid protocol [16].

Centiloids rather than SUVRs were used for comparison of our results with others using different amy-loid PET tracers [16, 17]. We classified centiloid amyloid status with the following cut-offs: amyloid negative <15; uncertain or intermediate amyloid = 15–25; amyloid positive >25.

MRI acquisition

Whole brain images were acquired on a single 3T Siemens Tim Trio Scanner with a 12-channel head coil (Siemens, Erlangen, Germany)—see published protocol for imaging details [13]: T1-weighted MPRAGE sequences 160 slices; repetition time, TR = 1900 = ms; echo time, TE = 2.6 = ms; inversion time, TI = 900 = ms; flip angle = 9°; field of view 256×256 pixels; voxel size = 1 = mm³.

MRI imaging analysis

All images were visually inspected for quality control and excluded if degraded by motion or other artefacts before processing using automated pip-elines. Only those participants with evaluable MRI scans at three years and amyloid imaging were included in this analysis. Volumetric segmentations were performed using the longitudinal stream in FreeSurferV6.0 (http://surfer.nmr.mgh.harvard.edu/fswiki/LongitudinalProcessing) [18]. Left and right hippocampal results were averaged.

Cognitive testing

Participants underwent comprehensive neuropsychological testing using a cognitive test battery that has been previously described in full [13]. Age-appropriate normative values were used to calculate z-scores where available. The present analyses report on the following cognitive variables: 1) estimated premorbid full-scale IQ (FSIQ) assessed by the National Adult Reading Test (NART); 2) verbal delayed free recall memory assessed by the Hopkins Verbal Learning Test Revised (HVLT-R); 3) visual delayed free recall memory assessed by the Rey Complex Figure Test; 4) cognitive impairment status as judged by our Cognitive Outcome Evaluation Committee, where cognitive-impaired (CI) was determined by z-score < -1.5 in at least one cognitive domain [13].

Cognitive outcome evaluation committee

Outcome evaluation committee meetings were convened for allocation of cognitive status. De-ide-ntified information from their structured clinical interview (e.g., information regarding relevant recent psychosocial stressors, functional status), standardized cognitive z-scores, mood scores, and the Glo-bal CDR Score were presented. Participants were assigned their status of normal cognition (CN), cognitively impaired (CI), or dementia.

A participant was assigned CN if z-scores in all cognitive domains were within accepted age- and years of education-adjusted norms and there was no evidence of functional decline due to cognitive impairment (i.e., activities of daily living (ADLs) were unaffected). A participant was judged to be CI if 1) the z-score for at least one cognitive domain was lower than -1.5, and 2) there was no evidence of functional decline due to cognitive impairment (i.e., ADLs were unaffected). A participant was judged to have dementia if 1) z-scores for two or more cognitive domains were lower than -1.5 and 2) there was evidence of functional decline.

Statistical analysis

We compared amyloid positive and negative participants using the following demographic, clinical, cognitive, and imaging variables including age, sex, VRFs, CCI, years of education, verbal and visual delayed recall, cognitive-impaired status, APOE ɛ4 status, total intracranial volume (TIV), and HV (Table 1). TIV was used as a measure of skull volume and for inclusion in imaging analyses for adjustment. CCI was used as a continuous variable without a pre-specified cut-off for these analyses.

Table 1

Demographic, clinical, imaging, cognitive, and stroke characteristics in amyloid positive and negative stroke participants. Note that we have included statistical comparisons between positive and negative participants

	Amyloid positive	Amyloid negative	p
	(n = 4)	(n = 29)
Sociodemographic
Age, y, mean (SD)	78.3 (9.3)	71.3 (7.9)	0.24^*
Sex, men, no. (%)	3 (75)	23 (79.3)	>0.99^†
Clinical
APOEɛ4, no. (%)	1 (25)	5 (17.2)	>0.99^†
Depression, no. (%)	1 (25)	3 (10.3)	0.42^†
Charlson Comorbidity Scores	7 (5, 7.5)	4 (3, 5.3)	0.081^‡
(CCI), Md (Q1, Q3)
VRFs, no. (%)	3 (75)	20 (69)	>0.99^†
Hypertension, no. (%)	2 (50)	15 (51.7)	>0.99^†
Hypercholesterolemia, no (%)	2 (50)	13 (44.8)	>0.99^†
Type 2 diabetes mellitus, no. (%)	0 (0)	6 (20.7)	>0.99^†
Atrial fibrillation, no (%)	1 (25)	7 (24.1)	>0.99^†
Smoking, pack years, Md (Q1, Q3)	0 (0, 0)	1 (0, 14)	0.10^‡
Obese, no. (%)	3 (75)	6 (20.7)	0.052^†
Imaging
HV, mm³, mean (SD)	3622 (489)	3774 (425)	0.59^*
TIV, cm³, mean (SD)	1658 (80)	1557 (162)	0.083^*
HV/TIV, 10^-3, mean (SD)	2.2 (0.3)	2.4 (0.3)	0.16^*
White matter hyperintensities	11.5 (6.4, 16.4)	5.2 (1.7, 8.1)	0.13^‡
volume, cm³, Md (Q1, Q3)
Stroke lesion volume cm³,	1.9 (0.9, 10.8)	5.2 (0.4, 22.9)	0.53^‡
Md (Q1, Q3)
Cognitive
Education, y, Md (Q1, Q3)	16 (12, 20.5)	12 (10, 14)	0.11^‡
NART-FSIQ, Md (Q1, Q3)	119 (104, 125)	111 (104, 116)	0.46^‡
Delayed verbal free recall,	–0.96 (–3.06, –0.35)	–0.35 (–1.39, 0.13)	0.32^‡
z-scores, Md (Q1, Q3)^a
Delayed visual free recall, z-scores, Md (Q1, Q3)^a	1.01 (–1.82, 2.10)	0.35 (–0.18, 0.92)	0.67^‡
Cognitive impairment, no. (%)	2 (50)	5 (17.2)	0.071^†
Stroke characteristics
Previous stroke, no. (%)	0 (0)	3 (10.3)	>0.99^†
Thrombolysis, no. (%)	0 (0)	2 (6.9)	>0.99^†
NIHSS score, Md (Q1, Q3)	3 (2.5, 3)	3 (2, 4)	0.80^‡
NIHSS severity, mild (0–7), no. (%)	4 (100)	27 (93.1)	>0.99^†
mRS score, Md (Q1, Q3)	0.5 (0, 1)	1 (0, 2)	0.21^‡
mRS severity, mild (0–1), no. (%)	4 (100)	20 (69)	0.55^†
Stroke side, no. (%)			0.67^†
Right	2 (50)	19 (65.5)
Left	2 (50)	8 (27.6)
Bilateral	0 (0)	2 (6.9)
Oxfordshire, no. (%)			0.33^†
Lacunar	1 (25)	3 (10.3)
Posterior	0 (0)	11 (37.9)
Partial anterior	3 (75)	14 (48.3)
Total anterior	0 (0)	1 (3.5)

^*Two-sample t test; ^†Fisher exact test; ^‡Wilcoxon Rank Sum test. VRFs, vascular risk factors; APOE ɛ4, apolipoprotein ɛ4; Q1, Q3, 25th, 75th percentiles; HV, hippocampal volume; TIV, total intracranial volume; SD, standard deviation; NART-FSIQ, National Adult Reading Test Full-Scale IQ. ^aNote: only 3 out of 4 amyloid positive participants completed cognitive testing, as one participant had a diagnosis of dementia made by our hospital Memory Service following clinical assessment. Median, Q1 and Q3 values therefore represent the middle, lowest, and highest cognitive scores. All p-values should be interpreted with caution given low numbers and included as a guide to interpreting difference only.

Age, HV, and TIV were analyzed with 2-sample 2-tailed t-tests. Fisher exact tests were used to compare sex, APOE ɛ4, cognitively impaired status, and VRFs. Wilcoxon rank tests were used to compare CCI, years of education, verbal and visual del-ayed recall. We compared HV between the groups using multi-way (N-way) ANOVA (anovan, MATLAB 2019b) with CCI, VRFs, and TIV as covariates. We also investigated the association of HV with cen-tiloids after adjusting for CCI, VRFs, and TIV. Res-ults should be interpreted with caution given the small numbers in the group who were amyloid positive.

RESULTS

Participants

Thirty-three participants (26 men) with complete data sets (i.e., MRI HV plus amyloid status) were available. Given these small numbers, statistical analyses should be interpreted with caution.

Comparison of amyloid-negative and amyloid-positive participants

All participants were classified as positive or negative (i.e., no indeterminate). Only four of our 33 participants were amyloid positive (see Table 1). Two of the amyloid-positive participants were cognitively impaired (50%) as were five of the amyloid-negative (17%). No participant had dementia.

There were no significant differences in age, sex, years of education, cognitive status, verbal and visual delayed recall, recurrent stroke, location and side of stroke, APOE ɛ4 carriage, CCI, and VRFs between amyloid-positive and -negative participants (see Table 1 and Fig. 1). We found no association between amyloid status or centiloids scale or CCI and HV. CCI variability was associated with age. There was no difference in APOE ɛ4 carriage between groups (25% ɛ4 PET amyloid positive; 17.2% ɛ4 amyloid-negative).

Fig. 1

Hippocampal volume (HV) in amyloid-negative and amyloid–positive participants. Y-axis=hippocampal volume (HV). Midline = median HV. X-axis=group: n = 29 amyloid-negative versus n = 4 amyloid-positive.

DISCUSSION

In this small group of ischemic stroke survivors three years after their clinical event, hippocampal volume was not associated with amyloid burden. This is consistent with prior published associations between amyloid burden, increasing age and mid-life cardiovascular health [1]. While some researchers have proposed that hippocampal atrophy in stroke survivors is caused by incipient AD pathology, this has not been borne out in PET amyloid imaging studies of stroke [1 , 9]. There is increasing evidence that vascular risk factors and stroke exert a neurodegenerative effect independently of AD proteins. It appears more likely that the observed atrophy may be a result of vascular brain burden and associated secondary neurodegeneration [2, 3].

We found no association with amyloid status and the presence or absence of cognitive impairment or performance on memory tests. Given our small numbers, this could be due to Type II error, as hip-pocampal atrophy is associated both with a clinical diagnosis of AD dementia and episodic memory performance. However, hippocampal atrophy is also strongly associated with vascular risk factors and stroke, and entorhinal volume changes correlate with memory performance in unimpaired older adults [19]. Hence, hippocampal volume may be impacted by vascular brain burden and stroke effects independently of AD pathology [20].

Limitations of our study include the small sample size, very small numbers of amyloid positive participants, and absence of tau PET imaging. Tau imaging would have been particularly interesting in this cohort of patients, given that it is posited that neurodegeneration tracks more closely with tau than amyloid [21], and tau-PET positivity has been demonstrated to be associated with cognitive impairment in vascular dementia [22]. Vascular risk factors are associated with entorhinal cortical tau deposition [23], and APOE ɛ4 carriage is associated both with cardiovascular risk and medial temporal tau burden, independent of amyloid burden [24]. White matter hyperintensities have been shown to promote tau pathology in AD [25]. Indeed, hippocampal atrophy occurs independently of amyloid deposition and may be driven by vascular dysregulation [26]. The lack of association with amyloid may not surprise some readers, as hippocampal atrophy in people with a clinical diagnosis of AD is thought to be more associated with neurofibrillary tangle formation and tau burden.

Given our small group, these findings should be regarded as exploratory only, requiring validation in larger cohorts with amyloid PET and MRI. However, our findings suggest that factors other than amyloid may contribute to brain atrophy in vascular populations. We propose that hippocampal atrophy be considered a manifestation of vascular brain burden and associated with post-stroke neurodegeneration, rather than evidence of underlying AD pathology.

Footnotes

ACKNOWLEDGMENTS

The authors would like to thank the Victorian Life Sciences Computation Initiative at the University of Melbourne, National Imaging Facility at the Florey Node, the radiographers at Melbourne Brain Centre, the staff at the Molecular Imaging and Therapy Department of Austin Health, and all our participants, who so generously contributed their time to the study.

This work was supported by NHMRC grants GNT1020526 and GNT1094974 (AB) and Heart Foundation Future Leader Fellowship (AB, 100784).

Authors’ disclosures available online ().

References

Han

, Maillard

, Harvey

, Fletcher

, Martinez

, Johnson

, Olichney

, Farias

, Villeneuve

, Jagust

, Mungas

, DeCarli

(2020) Association of vascular brain injury, neurodegeneration, amyloid and cognitive trajectory. Neurology 95, e2622–e2634.

Brodtmann

, Khlif

, Egorova

, Veldsman

, Bird

, Werden

(2020) Dynamic regional brain atrophy rates in the first year after ischemic stroke. Stroke 51, e183–e192.

Egorova

, Dhollander

, Khlif

, Khan

, Werden

, Brodtmann

(2020) Pervasive white matter fiber degeneration in ischemic stroke. Stroke 51, 1507–1513.

Veldsman

, Cheng

H-J

, Fang

, Werden

, Khlif

, Ng

, Lim

JKW

, Qian

, Yu

, Zhou

, Brodtmann

(2020) Degeneration of structural brain networks is associated with cognitive decline after ischaemic stroke. Brain Commun 2, fcaa155.

Hagberg

, Ihle-Hansen

, Fure

, Thommessen

, Ihle-Hansen

, Oksengard

, Beyer

, Wyller

, Muller

, Pendlebury

, Selnes

(2020) No evidence for amyloid pathology as a key mediator of neurodegeneration post-stroke - a seven-year follow-up study. BMC Neurol 20, 174.

Park

, Seo

, Kim

, Jeon

, Lee

, Oh

, Kim

, Choe

, Lee

, Shin

, Kim

, Noh

, Cho

, Yoon

, Kim

, Ye

, Ewers

, Weiner

, Lee

, Werring

, Na

(2014) Effects of cerebrovascular disease and amyloid beta burden on cognition in subjects with subcortical vascular cognitive impairment. Neurobiol Aging 35, 254–260.

Lee

, Seo

, Na

, Kim

, Park

, Kim

, Noh

, Cho

, Kim

, Yoon

, Ye

, Chin

, Jeon

, Lee

, Choe

, Lee

, Kim

, Lee

, Ewers

, Werring

, Weiner

(2014) Synergistic effects of ischemia and beta-amyloid burden on cognitive decline in patients with subcortical vascular mild cognitive impairment. JAMA Psychiatry 71, 412–422.

Kim

, Lee

, Seo

, Kim

, Seong

, Ye

, Cho

, Noh

, Kim

, Yoon

, Oh

, Kim

, Choe

, Lee

, Kim

, Hwang

, Jeong

, Na

(2015) Hippocampal volume and shape in pure subcortical vascular dementia. Neurobiol Aging 36, 485–491.

Sahathevan

, Linden

, Villemagne

, Churilov

, Ly

, Rowe

, Donnan

, Brodtmann

(2016) Positron emission tomographic imaging in stroke: Cross-sectional and follow-up assessment of amyloid in ischemic stroke. Stroke 47, 113–119.

10.

Yasuno

, Kajimoto

, Ihara

, Taguchi

, Yamamoto

, Fukuda

, Kazui

, Iida

, Nagatsuka

(2019) Amyloid beta deposition in subcortical stroke patients and effects of educational achievement: A pilot study. Int J Geriatr Psychiatry 34, 1651–1657.

11.

Wong

, Lau

, Yang

, Wang

, Liu

, Lam

, Au

, Shi

, Wang

, Chu

, Xiong

, Lo

, Law

, Leung

, Lam

, Chan

, Soo

, Leung

, Wong

, Mok

(2016) Neuropsychiatric symptom clusters in stroke and transient ischemic attack by cognitive status and stroke subtype: Frequency and relationships with vascular lesions, brain atrophy and amyloid. PLoS One 11, e0162846.

12.

Kim

, Lee

, Seo

, Ye

, Cho

, Kim

, Noh

, Yoon

, Chin

, Oh

, Kim

, Choe

, Lee

, Kim

, Jeong

, Na

(2014) Seoul criteria for PiB(-) subcortical vascular dementia based on clinical and MRI variables. Neurology 82, 1529–1535.

13.

Brodtmann

, Werden

, Pardoe

, Li

, Jackson

, Donnan

, Cowie

, Bradshaw

, Darby

, Cumming

(2014) Charting cognitive and volumetric trajectories after stroke: Protocol for the Cognition And Neocortical Volume After Stroke (CANVAS) study. Int J Stroke 9, 824–828.

14.

D’Hoore

, Sicotte

, Tilquin

(1993) Risk adjustment in outcome assessment: The Charlson comorbidity index. Methods Inf Med 32, 382–387.

15.

Werden

, Cumming

, Li

, Bird

, Veldsman

, Pardoe

, Jackson

, Donnan

, Brodtmann

(2017) Structural MRI markers of brain aging early after ischemic stroke. Neurology 89, 116–124.

16.

Rowe

, Jones

, Dore

, Pejoska

, Margison

, Mulligan

, Chan

, Young

, Villemagne

(2016) Standardized expression of 18F-NAV4694 and 11C-PiB beta-amyloid PET results with the centiloid scale. J Nucl Med 57, 1233–1237.

17.

Bourgeat

, Dore

, Fripp

, Ames

, Masters

, Salvado

, Villemagne

, Rowe

, AIBL research group (2018) Implementing the centiloid transformation for (11)C-PiB and beta-amyloid (18)F-PET tracers using CapAIBL. Neuroimage 183, 387–393.

18.

Reuter

, Schmansky

, Rosas

, Fischl

(2012) Within-subject template estimation for unbiased longitudinal image analysis. Neuroimage 61, 1402–1418.

19.

Yeung

, Hale

, Rizvi

, Igwe

, Sloan

, Honig

, Small

, Brickman

(2020) Anterolateral entorhinal cortex volume is associated with memory retention in clinically unimpaired older adults. Neurobiol Aging 98, 134–145.

20.

Pelcher

, Puzo

, Tripodis

, Aparicio

, Steinberg

, Phelps

, Martin

, Palmisano

, Vassey

, Lindbergh

, McKee

, Stein

, Killiany

, Au

, Kowall

, Stern

, Mez

, Alosco

(2020) Revised Framingham Stroke Risk Profile: Association with cognitive status and MRI-derived volumetric measures. J Alzheimers Dis 78, 1393–1408.

21.

Jack

Jr. , Knopman

, Jagust

, Shaw

, Aisen

, Weiner

, Petersen

, Trojanowski

(2010) Hypothetical model of dynamic biomarkers of the Alzheimer’s pathological cascade. Lancet Neurol 9, 119–128.

22.

Kim

, Park

, Cho

, Jang

, San Lee

, Jang

, Kim

, Ryu

, Choi

, Moon

, Weiner

, Jagust

, Rabinovici

, DeCarli

, Lyoo

, Na

, Seo

(2018) Assessment of extent and role of tau in subcortical vascular cognitive impairment using 18F-AV1451 Positron emission tomography imaging. JAMA Neurol 75, 999–1007.

23.

Rabin

, Yang

, Schultz

, Hanseeuw

, Hedden

, Viswanathan

, Gatchel

, Marshall

, Kilpatrick

, Klein

, Rao

, Buckley

, Yau

, Kirn

, Rentz

, Johnson

, Sperling

, Chhatwal

(2019) Vascular risk and beta-amyloid are synergistically associated with cortical tau. Ann Neurol 85, 272–279.

24.

Therriault

, Benedet

, Pascoal

, Mathotaarachchi

, Chamoun

, Savard

, Thomas

, Kang

, Lussier

, Tissot

, Parsons

, Qureshi

MNI

, Vitali

, Massarweh

, Soucy

, Rej

, Saha-Chaudhuri

, Gauthier

, Rosa-Neto

(2020) Association of Apolipoprotein E epsilon4 with medial temporal tau independent of amyloid-beta. JAMA Neurol 77, 470–479.

25.

Laing

, Simoes

, Baena-Caldas

, Lao

, Kothiya

, Igwe

, Chesebro

, Houck

, Pedraza

, Hernandez

, Li

, Zimmerman

, Luchsinger

, Barone

, Moreno

, Brickman

, Alzheimer’s Disease Neuroimaging Initiative (2020) Cerebrovascular disease promotes tau pathology in Alzheimer’s disease. Brain Commun 2, fcaa132.

26.

Iturria-Medina

, Sotero

, Toussaint

, Mateos-Perez

, Evans

, Alzheimer’s Disease Neuroimaging Initiative (2016) Early role of vascular dysregulation on late-onset Alzheimer’s disease based on multifactorial data-driven analysis. Nat Commun 7, 11934.