Normal Cognition Research Articles

Florbetapir amyloid PET in cognitively impaired former professional and college American football players

AbstractBackgroundRepetitive head impact (RHI) exposure from the routine play of American football increases the risk of later‐life cognitive impairment and dementia from the neurodegenerative disease, chronic traumatic encephalopathy (CTE). Although neuropathological studies of CTE have found comorbid neurodegenerative and other pathologies, neuritic Aβ plaques are infrequent, particularly in early stages. However, amyloid PET imaging has not been fully characterized in cognitively impaired former American football players.MethodIn this cross‐sectional, observational cohort study from the DIAGNOSE CTE Research Project, florbetapir PET mean cortical standard uptake value ratios (SUVR) and positivity (defined by average SUVR >1.10, indicative of moderate‐to‐frequent neuritic Aβ plaques), were analyzed in 119 former professional (PRO) National Football League players, 60 former college football players (COL), and 58 men unexposed (UE) to RHI who were cognitively asymptomatic. Former players were categorized into one of four diagnostic groups: cognitively normal (CN), subjective memory complaints (SMC), mild cognitive impairment (MCI), and dementia (DEM). All participants were 45‐74 yo. Multivariable linear and logistic regressions were conducted, with race, education, age, and APOE4 genotype included as covariates. A priori power calculations demonstrated adequate power to detect significant findingsResultNo differences between the three exposure groups in average florbetapir SUVR or proportion of elevated florbetapir uptake were found. Pairwise group comparisons resulted in no differences in average SUVR between PRO and COL (‐0.01, 95% CI [‐0.033, 0.025], p = 0.95), PRO and UE (0.02, 95% CI [‐0.010, 0.029], p = 0.41), and COL and UE (0.02, 95% CI [0.0004, 0.039], p = 0.36). Analysis of covariance resulted in no differences among diagnostic groups (F = 0.94, p = 0.43). Neither years of playing football nor measures of cognition and daily functioning showed significant associations with florbetapir SUVR.ConclusionFormer professional and college American football players with cognitive impairment and dementia did not have elevated florbetapir amyloid PET compared to a group of men without RHI exposure. Their rates of florbetapir positivity in the football players were lower than expected for a group of men this age. These findings suggest that cognitive impairment in former elite football players may not be due to Alzheimer’s disease pathology.

Racial differences in the effect of APOE‐ε4 genotypes on trail making test B in Alzheimer's disease: A longitudinal study

AbstractObjectivesThe trail making test part B (TMT‐B) evaluates executive functions, memory, and sensorimotor functions. No previous study was found to examine the longitudinal effect of APOE‐ε4 genotypes on the TMT‐B scores in Alzheimer's disease (AD) across racial groups.MethodsThis study used the data from Alzheimer's Disease Neuroimaging Initiative (ADNI): 382 participants with AD, 503 with cognitive normal (CN), 1293 with mild cognitive impairment (MCI) at baseline and follow‐up of four years. The multivariable linear mixed model was used to investigate the effect of APOE‐ε4 genotypes on changes in TMT‐B scores.ResultsCompared with Whites, African Americans (AA) and Hispanics had higher TMT‐B scores (poor cognitive function). Furthermore, Whites subjects with 1 or 2 APOE‐ε4 alleles had significantly higher TMT‐B scores compared with individuals without APOE‐ε4 allele at baseline and four follow‐up visits; however, no differences in TMT‐B were found between APOE‐ε4 alleles in the Hispanic and AA groups. No APOE‐ε4 by visit interactions was found for 3 racial groups. Stratified by AD diagnosis, the APOE‐ε4 allele was associated with TMT‐B scores only in the MCI group, while there were significant interactions for visit by education, APOE‐ε4 allele, and the Mini Mental State Examination (MMSE) score in the MCI group. In addition, TMT‐B was significantly correlated with the MMSE, AD Assessment Scale‐cognitive subscale 13 (ADAS13), tTau, pTau, Aβ42, and hippocampus.ConclusionsAPOE‐ɛ4 allele is associated with TMT‐B scores in Whites subjects, but not in the Hispanic and AA groups. APOE‐ε4 showed interaction with visit in the MCI group.

Combining working memory and speech data in a digital app for early detection of cognitive impairment in primary care

AbstractBackgroundCognitive screening to detect mild cognitive impairment (MCI) and dementia in primary care settings has proven to be a challenging task. The ideal solution would be a brief, low‐burden assessment tool appropriate for use with individuals from diverse educational and cultural backgrounds that requires limited time and expertise from primary care providers. The purpose of this project was to develop a digital cognitive screening app derived from widely used cognitive and speech tasks using machine learning techniques.MethodParticipants were 104 old adults, 53 who were cognitively normal (CN) and 51 who were cognitively impaired (CI). All participants completed a working memory task (WM), followed by four speech‐language tasks, followed by a second administration of WM (WM2) to investigate the added utility of practice effects. Participants also completed the Quick Mild Cognitive Impairment (Qmci) screen. Bayesian adaptive regression trees were used to test 11 models. All models contained sex, age, education, and marital status. One model contained WM and the WM2‐WM difference. Four models contained 26 acoustic and 34 linguistic variables derived from each of the four speech tasks, and four more models were identical, except with the two WM variables added. For comparison, two models with Qmci were fitted: one with and one without the two WM variables.ResultThree models achieved cross‐validated classification accuracy of c≥0.90: WM by itself (estimated c = 0.91; Figure 1), WM and a personal narrative task (c = 0.94; Figure 2), and WM and a counting backwards task (c = 0.90). Models with picture description and procedural discourse tasks performed the worst. The QMCI‐only model yielded c = 0.91.ConclusionCombining WM, acoustic and linguistic variables involved in recounting personal memories, and the ability to benefit from prior exposure to a task discriminated CN from CI groups with a high degree of accuracy. This unique combination assesses key cognitive abilities known to be affected early in neurodegenerative diseases, executive functioning and memory, and includes both verbal and nonverbal tasks. Importantly, individuals with varying levels of cognitive functioning, including mild dementia, were able to use the app successfully, and the automated nature of the tool required minimal staff time or expertise.

Development of a triage algorithm for mild cognitive impairment using electronic health record data

AbstractBackgroundTimely detection and diagnosis of early‐stage Alzheimer’s disease (AD) become increasingly relevant with the introduction of new therapies. Electronic health records (EHR) data present an opportunity to identify cases with elevated risk of undiagnosed cognitive impairment and triage them for further evaluation. Such triage tools are critically important because population‐based screening of asymptomatic individuals is currently not recommended, but spontaneous detection rates are low. We previously developed a risk prediction algorithm to identify likely cases aged ≥50 years based on claims data.MethodWe will build on this work and develop an algorithm based on three data sources: claims, EHR data and the National Alzheimer’s Coordinating Centers (NACC) data. The claims database study, already completed, included newly‐diagnosed mild cognitive impairment (MCI) individuals and non‐MCI individual matched on age, gender and geographic region. Using data 2 years prior to diagnosis, medical conditions associated with MCI were used to build a predictive model. The EHR study will aim to augment the model with clinical data such as body mass index, smoking status, vital signs and laboratory results. The NACC study, including ∼5,400 MCI and ∼15,800 cognitively normal (CN) individuals, will be used to validate the algorithm.ResultThe claims database study identified 25 medical conditions that predict MCI risk (area under the curve, AUC = 0.78). The top five predictors for MCI were depression, stroke/transient ischemic attack, obstructive sleep apnea, hyperlipidemia, and hypertension. The predictive performance was better in the 50‐64 year than in the older age groups. Analyses of the EHR and NACC data are ongoing.ConclusionData routinely collected in claims and EHR may be accurate enough to triage individuals with potential MCI for cognitive assessment in primary care.

Effects of cerebrovascular disease on tau progression and cognitive decline in Alzheimer’s disease

AbstractBackgroundRecent post‐mortem studies suggest that cerebrovascular disease contributes to Alzheimer’s disease (AD) pathophysiology and clinical progression. In particular, cerebral amyloid angiopathy is highly common in AD and has been linked to faster cognitive decline. Yet, in‐vivo evidence on the contribution of cerebrovascular disease to pathobiological and/or clinical AD progression is rare. Thus, we assessed in‐vivo, whether i) higher cerebrovascular burden is associated with accelerated progression of primary AD pathology and ii) whether cerebrovascular burden synergistically accelerates AD‐related cognitive decline.MethodWe included n = 256 participants from the Alzheimer’s Disease Neuroimaging Initiative (ADNI) across the AD spectrum with longitudinal 18F‐flortaucipir tau‐PET (follow‐up∼2.30±1.32 years), 18F‐florbetapir or 18F‐florbetaben Aβ‐PET transformed to centiloid (Aβ–: n = 85 Cognitively normal (CN); Aβ+: n = 83 CN, n = 57 Mild Cognitive Impairment, n = 31 Dementia). We focused on global white matter hyperintensity (WMH) volume adjusted to intracranial volume and microhemorrhage (MCH) count as well‐established cerebrovascular disease markers. Cognition was probed through the AD Assessment Scale Cognition 13‐item scale (ADAS‐13). Subject‐specific annual changes in tau‐PET and cognition were calculated via linear mixed models, including random slope and intercept. To enhance reproducibility of our findings, statistical analyses were pre‐specified at AsPredicted.org. All statistical models were controlled for age, sex and APOE.ResultUsing linear regression, we found no main effect of cerebrovascular disease markers on baseline amyloid‐PET or tau‐PET (ps>0.05). Further, higher cerebrovascular burden was not associated with faster Aβ‐related tau‐PET SUVR increase (Fig. 1A+B). In contrast, higher cerebrovascular burden was associated with accelerated cognitive decline. Specifically, interaction models revealed that a higher MCH count was associated with faster cognitive decline at higher tau levels (Fig. 1C), while higher white matter hyperintensity burden was associated with faster cognitive decline at lower tau levels (Fig. 1D).ConclusionWe found no effect of cerebrovascular disease on primary AD pathology. In contrast, cerebrovascular disease was linked to accelerated tau‐associated cognitive decline. This suggests that cerebrovascular disease may not modulate pathobiological AD trajectories but rather add synergistically to AD symptom severity. This has important therapeutic implications, as treating cerebrovascular disease in addition to primary AD pathology (e.g. amyloid) may yield largest clinical efficacy.

Microbial infection and associated inflammasome in the retina of Alzheimer’s disease patients

AbstractBackgroundRecent advances in Alzheimer’s disease (AD) research revealed that AD pathology is not confined to the brain but also occur in the neurosensory retina. Data have further suggested parallels between retinal and brain AD biomarkers. While the exact etiology of AD pathogenesis is undefined, an emerging theory has linked AD with microbial infection as a plausible cause for AD progression. The presence of several microbes including Chlamydia pneumoniae (Cp), Borrelia burgdorferi, Porphyromonas gingivalis, and Toxoplasma gondii have been reported in the post‐mortem brain of AD patients and murine models. However, the existence and role of microbial infection in AD retina has never been studied.MethodWe employed a range of histological, biochemical, and genetic techniques to detect bacterial infection and other associated biomarkers in postmortem retinas of patients with mild cognitive impairment (MCI) or AD dementia, compared to retinas from age‐matched individuals who displayed normal cognition (NC). We assessed the presence and burden of Cp and correlated with gliosis, inflammasome, and apoptosis markers.ResultWe discovered the presence of Cp inclusions in postmortem retinas of patients with MCI and AD as compared to age‐ and sex‐matched NC individuals. Our experimental evidence revealed Cp‐infected retinal neuronal cells in AD patients, with no increase in retinal Cp infection in MCI versus CN and a significant 4‐fold increase in of Cp immunoreactivity in retinas of AD patients versus MCI and CN subjects. These data possibly indicate that Cp infection mainly occur during the dementia phase of AD and is less observable at the early functional impairment stages of AD. Retinal microglia were found to aggregate around and engulf Cp‐infected cells. Alongside Cp infection, retinal tissues from MCI and AD patients exhibited increased NLRP3 inflammasome and gliosis. Our data suggest Cp‐triggered inflammasome activation, that can lead to neuronal degeneration. We further found a relationship between the severity of retinal Cp infection and brain pathology; retinal Cp burden was associated with brain neurofibrillary tangles, neuropil threads, Braak stage.ConclusionCp‐triggered chronic inflammation could be a detrimental process leading to retinal degeneration in AD, suggesting potential new therapeutic avenues involving early antibiotic treatment or inflammasome attenuation.

Relationship of structural and functional visual biomarkers with measures of amyloid, tau, and neurodegeneration across the AD continuum

AbstractBackgroundThe eye often reflects changes seen in the brain in neurodegenerative diseases. This study sought to examine the relationship of contrast sensitivity measured using frequency doubling technology (FDT) and retinal thickness measured using optical coherence tomography (OCT) with temporal lobe neurodegeneration, cerebral tau deposition, and mean plasma ptau181 in older adults along the Alzheimer’s disease (AD) continuum.MethodParticipants included 28 individuals with subjective cognitive decline (SCD), 18 mild cognitive impairment (MCI), 4 with AD, and 30 age‐matched cognitive normal (CN) from the Indiana Memory and Aging Study at the Indiana ADRC. Participants were excluded from the study if they had significant eye disease determined to interfere with OCT, FDT, non‐AD dementia, or exclusion for MRI or PET. Mean deviation (MD) and pattern standard deviation (PSD) from FDT were used as measures of contrast sensitivity. OCT scans from each eye to measure retinal thickness in the retinal nerve fiber layer (RNFL) and macula. MRI scans were processed using Freesurfer v6 to measure medial (MTL) and lateral temporal lobe (LTL) volumes. LTL SUVR values were extracted from [18F]flortaucipir PET scans. Plasma ptau 181 was measured using a SIMOA assay. Finally, the association between contrast sensitivity, average RNFL thickness, macular volumes with temporal lobe volumes, LTL tau, and mean plasma ptau181 was assessed using a partial Pearson correlation, covaried for age, sex, and diagnosis. p<0.05 was considered significant.ResultRNFL thickness was decreased in patients with AD. Hippocampal volume correlated with RNFL thickness and macular volume in both eyes and with PSD in the left eye. The inferior temporal volume correlated with MD in both eyes, while mean plasma ptau181 with PSD in both eyes. Finally, lateral temporal lobe tau was correlated with the inner retinal layers of ganglion cell layer plus inner plexiform layer thickness.ConclusionContrast sensitivity correlates with temporal lobe volumes and mean plasma ptau181. RNFL thinning and macular volumes are associated with hippocampal atrophy, and inner retinal nuclear layers correlate with lateral temporal tau. Imaging of the eye may represent a useful biomarker to screen patients at risk for AD prior to more invasive and prolonged testing.

Prediction of Mild Cognitive Impairment using eye movements and other easily accessible measures

AbstractBackgroundThe early identification of patients at risk of dementia and timely medical intervention can be effective in preventing disease progression and mitigate burdens and costs of health insurance. Although the current diagnostic tests for preclinical and prodromal Alzheimer’s disease (AD) using neuropsychological evaluation and biomarkers can accurately predict the progression to AD dementia, the high cost, invasive nature, and limited availability restrict their use to wealthy clinics only. Therefore, we developed a model to predict cognitive impairment based on non‐invasive information, including eye movement (EM) data.MethodEye‐tracking data was extracted from 594 subjects, 428 cognitively normal (CN) controls, and 166 patients with mild cognitive impairment (MCI) while they performed Prosaccade/Antisaccade (PS/AS) and Go/No‐go tasks. EM behavior was compared by task after stratification by cognitive status and controlling for age, sex, and education. We used logistic regression to calculate odds ratios (OR) and 95% confidence intervals (CI) from the EM metrics.ResultLogistic regression models suggested increased OR of MCI from EM metrics, including wider latency variability, fixation duration variability, a higher number of errors committed, anticipations, and omissions with 95% CI of OR between 1.213‐1.621.ConclusionEM metric changes linked with MCI are associated with deficits in attentional and executive functions. The models built from non‐invasive data allow the discrimination of patients with MCI from cognitively normal adults. In summary, our results suggest that eye tracking, in combination with demographics and brief cognitive tests, might significantly improve the diagnostic prediction of MCI.

Human nasal beta‐amyloid 42 has a diagnostic value in discriminating Alzheimer’s disease

AbstractBackgroundThe key in Alzheimer’s disease (AD) therapy is a timely and accurate diagnosis for prompt drug intervention. However, due to the high cost and invasiveness of conventional biomarker analyses, including brain positron emission tomography (PET) imaging and cerebrospinal fluid (CSF)‐based assays, easy accessibility to these screening tests is often hindered. There is, therefore, a great need to develop a more accessible biomarker screening test using less invasive and cost‐effective peripheral body fluid biomarkers. Previous studies examined the non‐quantitative expression of beta‐amyloid (Aβ) in normal and AD patients' nasal discharge fluid. They identified higher expression of oligomeric Aβ in AD patients, showing a correlation with cognitive decline. However, the quantitative measurements of nasal Aβ42 levels, including the full AD continuum, remain unknown. Here, we assessed whether quantified human nasal Aβ42 levels have diagnostic performance and have the potential to be utilized as a biological risk factor in AD.Method161 subjects (cognitively normal (CN), n = 32; preclinical, n = 29; mild cognitive impairment (MCI), n = 73; AD, n = 27) underwent neuropsychological battery tests, including Mini‐Mental State Examination (MMSE), Clinical Dementia Rate (CDR), and Global Deterioration Scale (GDS), and amyloid‐PET scans (18F‐Flutemetamol). Their nasal discharge samples were collected, and nasal Aβ42 levels were measured via enzyme‐linked immunosorbent assay (ELISA).ResultWe found that the second‐highest quartile (Q3) group of nasal Aβ42 constituted the majority of patients with AD diagnosis (p = 0.036). The Q3 group also outnumbered the other groups in the most cognitively impaired subjects in all three neuropsychological battery tests (p = 0.008, p = 0.037, p = 0.023). Next we also found that subjects in the Q3 group exhibited most impaired cognitive function in all three neuropsychological battery tests with strong significance. When we examined the amyloid‐PET SUVR values, subjects with PET positive showed the highest SUVR.ConclusionQuantified nasal Aβ42 is strongly associated with cognition measurements. Nasal Aβ42 suggests the possibility for discriminating AD from non‐AD.

Subtle cognitive difficulties increase risk of neurodegeneration, incident mild cognitive impairment, and dementia in older women.

AbstractBackgroundIdentifying individuals with preclinical Alzheimer’s disease and related disorders (ADRD) is important for prevention and eventual treatment. Here we aimed to integrate overall test performance with episodic memory process scores in a cohort of older women to identify those with subtle cognitive difficulties (SCD; e.g., individuals with the earliest declines in cognitive ability preceding a diagnosis of mild cognitive impairment (MCI) or dementia) and determine subsequent risk of MCI, dementia, and neuroimaging indices of neurodegeneration.MethodParticipants included 2,186 community‐dwelling older women without MCI or dementia (Mage = 73.4±3.8 years) who completed a comprehensive neuropsychological assessment as part of the Women’s Health Initiative Study of Cognitive Aging (WHISCA: 1999‐2000). Women were assessed annually for MCI and dementia until 2021, and 1,024 women underwent structural magnetic resonance imaging (sMRI) of the brain in 2005‐6. Women were classified as cognitively normal (CN), early SCD (e‐SCD; two episodic memory process scores 1 SD below normative means or one process score and one performance score 1 SD below normative means), or late SCD (l‐SCD; two overall performance scores 1 SD below normative means) at WHISCA baseline using validated criteria from the Alzheimer’s Disease Neuroimaging Cohort (Thomas et al., 2018 Journal of Alzheimer’s Disease). We used Cox proportional hazards to examine associations between SCD status and risk of developing MCI/dementia. Linear regressions were used to examine associations between SCD status and MRI indices of AD‐related neurodegeneration, global cerebrovascular damage, and brain volumes of structures that make up the medial temporal lobe. All associations were adjusted for sociodemographic, lifestyle, and clinical characteristics.ResultWomen with e‐SCD and l‐SCD were 1.58 (p<0.001) and 3.01 (p<0.001) times as likely to develop MCI/dementia compared to CN women, respectively (Figure 1). Women with l‐SCD (β = 0.20; p<0.001) had more AD‐related neurodegeneration compared to CN women (Table 1). Women with e‐SCD had marginally significantly greater AD‐related neurodegeneration (β = 0.14; p = 0.073) and had smaller volumes in the entorhinal cortex (β = ‐0.22; p = 0.009) compared to CN women.ConclusionOur study demonstrates the usefulness of the SCD criteria for the early identification of cognitively normal women who are at increased risk of MCI, dementia, and neurodegeneration.

Spatial interplay between synaptic loss, UCH‐L1 deficiency, and pTau accumulation in the retina of MCI and AD patients

AbstractBackgroundSynaptic dysfunction occurs early during Alzheimer’s disease (AD) progression and is considered a strong predictor of cognitive decline. However, it has never been investigated in the AD retina. The ubiquitin‐proteasome system (UPS) is a major degradation pathway for damaged proteins and has been implicated in synaptic integrity and AD pathogenesis. Deficiency in ubiquitin hydrolase (UCH‐L1), a vital UPS component known for its neuroprotective and synaptoprotective properties, was previously linked to cerebral Aβ accumulation in AD, yet has never been explored in the AD retina.MethodWe employed a range of histological techniques to evaluate synaptic integrity and retinal pathology in postmortem retinas of patients with mild cognitive impairment (MCI) attributable to AD or AD dementia, compared to retinas from age‐matched individuals who displayed normal cognition (NC). We assessed the presence of amyloidosis, gliosis, tauopathy, apoptosis, in relation to UCH‐L1 levels, and visualized and quantified synaptic density through analysis of pre‐ and post‐synaptic markers while also assessing the spatial and cellular layer distribution of these markers.ResultWe detected extensive pre‐ and post‐synaptic losses in the IPL and OPL layers of MCI and AD patients, particularly in the far peripheral regions. This was accompanied by 2‐ to 4‐fold decrease in levels of UCH‐L1 in the retinal layers of MCI and AD patients, which directly correlated with pre‐synaptic loss (p<0.0001). These changes were inversely correlated with pTau accumulation (e.g., AT8, pS396), colocalizing at sites of synaptic loss in the IPL and OPL layers. Furthermore, there were significant (1.8‐4.2‐fold) increases in apoptotic cell death in the retinas of these patients (p<0.05–0.001). Remarkably, the pre‐synaptic losses significantly correlated with decreased cognitive scores.ConclusionOur results reveal a substantial retinal synaptic loss at sites of pTau accumulation in prodromal and clinical AD patients. Expression of retinal UCH‐L1 was significantly reduced in MCI and AD patients and strongly associated with synaptic loss and cognitive decline, signifying its potential role in protecting retinal neurons against AD‐related pathology.

Amyloid‐Beta PET Synthesis from Structural MRI: A Potential Alternative Method for Alzheimer’s Disease Screening

AbstractBackgroundAmyloid‐beta and structural brain atrophy are known to be hallmarks of Alzheimer’s Disease (AD), and can be quantified with Positron Emission Tomography (PET) and structural Magnetic Resonance (MRI), respectively. PET scans use radiotracers that binds to amyloid‐beta molecules, whereas MRI measures changes in structural morphology. PET scans are difficult to perform due to cost (∼$5000/scan), invasiveness, and ionizing radiation exposure, making them inaccessible for screening early‐onset AD. Conversely, MRI is a cheaper (∼$500/scan), non‐invasive, and free from ionizing radiation technique, however, it cannot provide molecular information obtained from PET. There is a known relationship between amyloid‐beta and brain atrophy, hence, amyloid‐beta PET images might be synthesized from structural MRI using image translation, which is an advanced form of machine learning.MethodAn image translation algorithm was developed using the Open Access Series of Imaging Studies (OASIS‐3) dataset that provides 929 subjects with pairs of T1‐weighted MRI and amyloid‐beta PET images, where 609 are cognitively normal (CN) and 489 at different stages of cognitive decline based on Clinical Dementia Rating scores. The image translation algorithm implemented with Conditional Generative Adversarial Networks, shown in Figure 1, was trained with 550 pairs with 156 axial slices each of amyloid‐beta and MRI images. Eight models where generated with different parameters to assess their impact in the generalization quality, then compared in a sample of 334 subjects using Structural Similarity Index (SSIM) and Peak Signal‐to‐Noise Ratio (PSNR).ResultFigure 2, compares the models performance based on SSIM and PSNR, where the best performing model generates synthetic PET images with a high degree of similarity with the real ones in Figure 3.ConclusionAn image translation pipeline was implemented and explored by changing parameters that can affect the model performance, the developed models are robust against the different parameters that were changed and it demonstrates that amyloid‐beta PET images can be synthesized from structural MRI, reducing cost, invasiveness and increasing availability. Such a method could enable non‐invasive screening of early onset AD.

CSF AβX‐34 ‐ a new biomarker for amyloid‐related small vessel pathology in early Alzheimer’s Disease

AbstractBackgroundBrain small vessel disease (SVD) increases Alzheimer’s disease (AD) risk. Impaired amyloid β (Aβ) clearance is linked to amyloid plaque formation and cerebral amyloid angiopathy (CAA). Aβ1‐40 is deposited in vessel walls, and Aβ with a C‐terminal 34' cut‐point is produced by perivascular cells. Evidence for SVD‐AD interactions in early AD is elusive, but incipient CAA putatively increases susceptibility for amyloid‐related imaging abnormalities (ARIA) in AD patients.Here we compare the SVD MRI marker “peak width of skeletonized mean diffusivity” (PSMD) to cerebrospinal fluid (CSF) Aβ species and the neurodegeneration marker Neurofilament light chain (Nfl), assessing links between SVD and AD pathologies and Aβ species as potential markers for incipient CAA.MethodCross‐sectional data from the Dementia Disease Initiation (DDI) cohort were available, comprising (n = 112) cognitively normal (CN) and (n = 115) cases with mild cognitive impairment (MCI) (n = 227). CSF markers for AD A/T/N‐staging identified 134 cases with negative markers (A‐/T‐/N‐), 21 with amyloid positive markers (A+/T‐/N‐) and 72 with amyloid and tau‐positive markers (A+/T or N+) (see table 1 for sample descriptives). Multiple linear regression models were used with PSMD by lobe as the dependent variable, and NfL, Aβ1‐40 (MesoScale), Aβx‐34 (Bioventix, Pre Diagnostic) (separately) and interactions with A/T/N group as independent variables. Variables were log‐transformed and standardized.ResultHigher Nfl was related to higher PSMD independently of the Aβ species in all lobes. Although Nfl correlated with Aβ1‐40 (figure 1), no multicollinearity was observed in the model. In A+, lower Aβ markers were related to higher PSMD in all lobes, but particularly for the temporal lobe (figure 2). Aβx‐34 showed more robust relationships with PSMD in both A+ groups particularly in the temporal lobes (A+/T‐/N‐ b = .62, p = <.001, A+/T or N+ b = .52, p<.001) than Aβ1‐40. (A+/T‐/N‐ b = .66, p = <.01, A+/T or N+ b = .33, p<.001). Neither t‐ nor p‐tau measures were associated with PSMD (not shown).ConclusionNfl independently related to PSMD, possibly reflecting amyloid‐independent neurodegeneration. CSF Aβx‐34 was in particular tightly linked to increased temporal lobe PSMD even in A+/T‐/N‐, pointing to early AD‐related SVD. This supports further evaluation of Aβx‐34 as a marker for CAA and susceptibility marker for ARIA.

Men show increased brain aging with respect to women among Cognitively Normal Individuals

AbstractBackgroundThere is an increasing interest in using machine learning and artificial intelligence to estimate chronological age using neuroimaging data. The gap between chronological age and estimated brain age (brain age gap, BAG) is used as a measure of accelerated/resilient brain aging. Previously, BAG has been associated with cognitive status. However, whether the BAG varies across sex and cognitive status have been less explored. The present study examines these associations and validates a voxel‐based machine learning approach based on the elastic net regression (ENR) for BAG calculation.MethodUsing data from the Atherosclerosis Risk in Communities Study (ARIC) study, the Wake Forest School of Medicine Alzheimer’s Disease Research Center (WFSM‐ADRC) clinical cohort and Alzheimer’s Disease Neuroimaging Initiative (ADNI), we examined associations of BAG across cognitive status and sex. We used structural MRI scans from 1853 ARIC participants (ages 67‐90, 60% females), 508 from the WFSM‐ADRC (55‐95 yo., 66% females) and 584 ADNI cognitively normal (CN) participants (55‐90 yo., 57% females). All images were aligned into a common template and the derived gray matter (GM) probability maps from ADNI MRIs were used as input to train the machine learning algorithm. Once the model was fitted the ARIC and WFSM‐ADRC GM probability maps were provided as input to the algorithm to estimate the BAG values. Finally, an age bias correction was applied. Linear regression methods were used to investigate differences between groups. Age, race, education, sex, and cognitive status were included in the model.ResultWe found in both ARIC and WFSM‐ADRC participants that differences in BAG values between CN‐MCI and MCI‐Dementia participants were highly significant. In addition, when we examined differences in BAG values across sex per cognitive status, we found again in both cohorts that differences were only significant for CN individuals. See Table 1 for details.ConclusionOur analyses show that our approach to estimate chronological age using high‐dimensional ENR, produces BAG values which are strongly associated with cognitive status. The increased severity of cognitive impairment is related to accelerated brain aging. Differences in BAG between men and women were significant for CN individuals only.

Structural brain network degeneration associated with agitation in dementia

AbstractBackgroundAgitation is a behavioral syndrome involving increased motor activity, restlessness, aggressiveness and emotional distress. Its has a 30% prevalence across multiple types of dementia and is associated with negative outcomes, including reduced quality of life, caregiver distress, and mortality. Pharmacological treatment risks serious side effects, including mortality. Understanding brain network topology could provide insights into novel treatment methods.MethodParticipants comprised 600 subjects from 3 existing databases: the Alzheimer’s Disease Neuroimaging Initiative (ADNI), Frontotemporal Lobar Degeneration Neuroimaging Initiative (NIFD) and the Mesulam Center for Cognitive Neurology and Alzheimer’s Disease at Northwestern University. Participants were clinically diagnosed with either behavioral variant frontotemporal dementia (bvFTD), mild cognitive impairment (MCI), dementia of the Alzeheimer type (DAT) or cognitively normal (CN). The Neuropsychiatric Inventory Questionnaire (NPI‐Q) caregiver rating was used to determine whether individuals were agitated or not. Morphometric similarity networks (MSNs) were generated from Freesurfer statistics calculated from T1‐weighted MRIs. 7 surface‐based cortical metrics (e.g. gray matter volume, surface area) were calculated for each of 360 parcels. Pairwise inter‐parcel Pearson correlations of feature vectors were calculated to produce a morphometric similarity matrix for each individual. We calculated sub‐matrices for salience (SN) cognitive control (CCN) and default mode networks (DMN). Brain Connectivity Toolbox calculated transitivity and global efficiency of each network. Metrics were calculated at different thresholds to ensure the results were not threshold‐dependent. We calculated 2 (Agitation present/absent) x 4 (Diagnosis) repeated measures ANCOVAS for transitivity and global efficiency for each network. Covariates were age, sex, race, database, education, ICV, CDR‐SB and days MRI‐NPI‐Q.ResultFor the SN, people with Agitation had significantly lower global efficiency than people without Agitation (Figure 1). There were no significant effects of diagnosis or interaction. For the CCN, people with agitation had significantly lower transitivity than people without Agitation (Figure 2). There were no significant effect of diagnosis or interaction.ConclusionAcross different forms of dementia, Agitation is associated with reduced integration (global efficiency) of the salience network, and reduced segregation (transitivity) of the cognitive control network. Interventions that alter these topological network features may be effective in reducing agitation in dementia, regardless of clinical diagnosis.

Performance Evaluation of CNN Models for Alzheimer’s Disease Detection with MRI Scans

The primary symptom of Alzheimer's disease is memory impairment, which is a neurodegenerative condition. The manifestation of these symptoms can be attributed to the impairment of the cerebral nerve responsible for cognitive functions such as learning, thinking and memory. Alzheimer’s disease is a prominent cause of mortality and lacks a definitive curve. However, appropriate medicinal interventions have demonstrated the potential to mitigate the progression and severity of the condition. This study presents the comparison of Convolutional Neural Network (CNN) models, namely ResNet50, VGG19 and VGG16 architectures, as an approach to construct an automated classification system for Alzheimer’s disease in future. The study utilises Magnetic Resonance Imaging (MRI) datasets to identify MRI datasets of individual with Alzheimer’s disease (AD), Cognitively normal (CN), mild cognitive impairment (MCI), early mild cognitive impairment (EMCI), and late mild cognitive impairment (LMCI). In the conducted experiment, the study achieved accuracy rates of 91.18% and 94.56% while utilising an epoch size of 2. The accuracy results indicate that the VGG16 model outperforms the ResNet50 model. The utilisation of automated Alzheimer’s disease classification holds potential as an auxiliary tool for healthcare professionals in determining the stage of Alzheimer’s disease hence facilitating the administration of suitable medicinal interventions.

Influence of APOE on the Alzheimer’s disease polygenic risk scores based on hippocampal subfield volumes

AbstractBackgroundNumerous studies have shown the association between Alzheimer’s disease, hippocampal subfield volume atrophy, and the role of APOE. In our previous study, we showed the differential atrophy in the hippocampal subfield volumes (HSV) identified using Alzheimer’s disease polygenic risk scores (AD PGRS). Here, we evaluate the influence of the APOE 𝛆4 on the AD PGRS of the HSV.MethodWe calculated the AD PGRS for HSV atrophy in 1,086 individuals (319 cognitively normal (CN), 591 mild cognitively impaired (MCI), and 176 Alzheimer’s disease dementia (ADD)). We then stratified the samples based on the APOE 𝛆4 carriers and non‐carriers. We studied the AD PGRS for the HSV for the left‐right hemispheres.ResultThe analysis was performed for the left and right hemispheres separately. The effects of APOE 𝛆4 were larger on the left hemisphere than on the right hemisphere. The effects of the APOE 𝛆4 were severe on most of the subfield volumes. Although, certain subfield volumes showed larger effects for 𝛆4 non‐carriers and certain subfield volumes showed no effect between 𝛆4 carriers and non‐carriers.ConclusionThe APOE 𝛆4 carriers showed a higher AD PGRS for hippocampal subfield volume atrophy than the APOE 𝛆4 non‐carriers. The AD PGRS and the APOE 𝛆4 showed the differential risk for the subfield volumes bilaterally. These risk scores can be used in preclinical Alzheimer’s risk modeling. We also hypothesize that hippocampal subfield volume atrophy occurs in both normal and pathological aging. However, atrophy is intense in pathological aging due to the underlying processes and the functioning of the risk genes. A clear understanding of this will enhance the specificity and sensitivity for the early diagnosis of Alzheimer’s disease.

Comparison of white matter hyperintensities between Latin American and American individuals with Alzheimer’s disease and cognitively normal older adults

AbstractBackgroundWhite matter hyperintensities (WMH) are Flair signal abnormalities frequently seen in periventricular and subcortical brain regions in the elderly. They are presumed to be, at least in part, a sign of cardiovascular disease (CVD). WMHs are associated with stroke, cognitive decline, and dementia. Given that most of the literature on dementia comes from high‐income countries, little is known about dementia in Latin American countries (LAC), which have an increased risk of dementia and CVD due to unique demographic, socioeconomic, cultural and ethnic factors. Our study aims to characterize the effects of white matter hyperintensities (WMH) in patients with AD within LAC and to compare these findings with a sample from the USA.MethodA total of 491 latin american subjects (235 cognitively normal (CN), 256 AD) and 271 american subjects (165 CN, 106 AD), underwent a brain MRI. WMH were extracted from T2‐FLAIR images using the Lesion Segmentation Toolbox and were divided into periventricular (pWMH) or deep (dWMH) depending on their distance from the ventricles. Linear mixed‐effects were used to model the volumes of WMH as a function of age, gender, years of education, intracranial volume, average cortical thickness, and region (LAC or USA).ResultAD groups from both regions showed larger total, dWMH and pWMH volumes compared to HC. Higher pWMH from AD and HC was associated with age, female sex and inversely associated with years of education. On the other hand, the increase in dWMH was found to be associated with female sex for the HC group and with age for AD. Cortical thinning was associated with an increase in total, subcortical and periventricular WMH for both groups. No statistically significant association was found between the region (USA or LAC) and WMH.ConclusionAD showed higher WMH compared to HC. Our findings suggest that pWMH and dWMH have different clinical correlates. No associations between region and WMH load or distribution were found. Future studies need to elucidate if there is an increase in other signs of CVD (such as cerebral microbleeds, brain lacunae, enlarged perivascular spaces) in Latin America and if they are linked to cognition and social determinants of health.

Which cortical regions affect the prediction of cerebrospinal fluid amyloid beta from PET images by novel residual neural network (ArcheD)

AbstractBackgroundAmyloid‐β (Aβ) aggregation, as a biological hallmark of Alzheimer’s disease (AD), is a key component of biological classification of AD in the AT(N) framework. This research aimed to develop a new deep learning model for predicting Aβ cerebrospinal fluid (CSF) measures from amyloid PET scans and explore which cortical regions are most important contributors in the model’s prediction.MethodWe developed a Residual Network model (ArcheD) to predict Aβ CSF from amyloid PET images, and trained it on 1,870 PET scans from the Alzheimer’s Disease Neuroimaging Initiative. We registered PET scans to MNI152 template and used the Neuromorphometrics atlas to derive cortical gray matter (GM) regions. We extracted brain regions’ contributions to model decision‐making by guided backpropagation, and visualized these contributions as relevance heatmaps. We used bootstrapping, Welch’s t‐test and FDR correction in testing significance of brain regions’ relevance to predictions in the whole cohort, and separately in clinical diagnostic classes of AD, mild cognitive impairment (MCI), and cognitively normal (CN) cases.ResultArcheD showed high accuracy in predicting Aβ CSF (r = 0.81; p<0.00001). According to the relevance analysis for all samples (n = 1,390) temporal (mean 0.106; 95%CI [0.105, 0.108]) and parietal (0.105; 95%CI [0.104, 0.107]) lobes were the most important GM regions for the model decision‐making. Occipital (0.098; 95%CI [0.098, 0.100]) and frontal lobes (0.091; 95%CI [0.089, 0.92]) showed noticeably fewer relevance values. Based on between‐classes comparison, these regions contribute significantly more to CSF prediction in AD and MCI individuals compared to CN (p<0.05). Analysis of 49 smaller regions showed an almost 7‐fold difference between the most and least contributing regions.ConclusionWe developed a deep residual neural network model that predicted Aβ CSF from amyloid PET scans with high accuracy, and found that temporal and parietal lobes have the highest influence on Aβ CSF predictions. In the next step, we will validate the model in independent data and aim to develop the model further for clinical use.

Which cortical regions affect the prediction of cerebrospinal fluid amyloid beta from PET images by novel residual neural network (ArcheD)

AbstractBackgroundAmyloid‐ß (Aß) aggregation, as a biological hallmark of Alzheimer’s disease (AD), is a key component of biological classification of AD in the AT(N) framework. This research aimed to develop a new deep learning model for predicting Aß cerebrospinal fluid (CSF) measures from amyloid PET scans and explore which cortical regions are most important contributors in the model’s prediction.MethodWe developed a Residual Network model (ArcheD) to predict Aß CSF from amyloid PET images, and trained it on 1,870 PET scans from the Alzheimer’s Disease Neuroimaging Initiative. We registered PET scans to MNI152 template and used the Neuromorphometrics atlas to derive cortical gray matter (GM) regions. We extracted brain regions’ contributions to model decision‐making by guided backpropagation, and visualized these contributions as relevance heatmaps. We used bootstrapping, Welch’s t‐test and FDR correction in testing significance of brain regions’ relevance to predictions in the whole cohort, and separately in clinical diagnostic classes of AD, mild cognitive impairment (MCI), and cognitively normal (CN) cases.ResultArcheD showed high accuracy in predicting Aß CSF (r = 0.81; p<0.00001). According to the relevance analysis for all samples (n = 1,390) temporal (mean 0.106; 95%CI [0.105, 0.108]) and parietal (0.105; 95%CI [0.104, 0.107]) lobes were the most important GM regions for the model decision‐making. Occipital (0.098; 95%CI [0.098, 0.100]) and frontal lobes (0.091; 95%CI [0.089, 0.92]) showed noticeably fewer relevance values. Based on between‐classes comparison, these regions contribute significantly more to CSF prediction in AD and MCI individuals compared to CN (p<0.05). Analysis of 49 smaller regions showed an almost 7‐fold difference between the most and least contributing regions.ConclusionWe developed a deep residual neural network model that predicted Aß CSF from amyloid PET scans with high accuracy, and found that temporal and parietal lobes have the highest influence on Aß CSF predictions. In the next step, we will validate the model in independent data and aim to develop the model further for clinical use.