Predicting cognitive decline in a low-dimensional representation of brain morphology.

Identifying early signs of neurodegeneration due to Alzheimer's disease (AD) is a necessary first step towards preventing cognitive decline. Individual cortical thickness measures, available after processing anatomical magnetic resonance imaging (MRI), are sensitive markers of neurodegeneration. However, normal aging cortical decline and high inter-individual variability complicate the comparison and statistical determination of the impact of AD-related neurodegeneration on trajectories. In this paper, we computed trajectories in a 2D representation of a 62-dimensional manifold of individual cortical thickness measures. To compute this representation, we used a novel, nonlinear dimension reduction algorithm called Uniform Manifold Approximation and Projection (UMAP). We trained two embeddings, one on cortical thickness measurements of 6237 cognitively healthy participants aged 18-100 years old and the other on 233 mild cognitively impaired (MCI) and AD participants from the longitudinal database, the Alzheimer's Disease Neuroimaging Initiative database (ADNI). Each participant had multiple visits ([Formula: see text]), one year apart. The first embedding's principal axis was shown to be positively associated ([Formula: see text]) with participants' age. Data from ADNI is projected into these 2D spaces. After clustering the data, average trajectories between clusters were shown to be significantly different between MCI and AD subjects. Moreover, some clusters and trajectories between clusters were more prone to host AD subjects. This study was able to differentiate AD and MCI subjects based on their trajectory in a 2D space with an AUC of 0.80 with 10-fold cross-validation.

The Influence of Birth Cohorts on Future Cognitive Decline.

BACKGROUND: Slowed rates of cognitive decline have been reported in individuals with higher cognitive reserve (CR), but interindividual discrepancies remain unexplained. Few studies have reported a birth cohort effect, favoring later-born individuals, but these studies remain scarce. OBJECTIVE: We aimed to predict cognitive decline in older adults using birth cohorts and CR. METHODS: Within the Alzheimer's Disease Neuroimaging Initiative, 1,041 dementia-free participants were assessed on four cognitive domains (verbal episodic memory; language and semantic memory; attention; executive functions) at each follow-up visit up to 14 years. Four birth cohorts were formed according to the major historical events of the 20th century (1916-1928; 1929-1938; 1939-1945; 1946-1962). CR was operationalized by merging education, complexity of occupation, and verbal IQ. We used linear mixed-effect models to evaluate the effects of CR and birth cohorts on rate of performance change over time. Age at baseline, baseline structural brain health (total brain and total white matter hyperintensities volumes), and baseline vascular risk factors burden were used as covariates. RESULTS: CR was only associated with slower decline in verbal episodic memory. However, more recent birth cohorts predicted slower annual cognitive decline in all domains, except for executive functions. This effect increased as the birth cohort became more recent. CONCLUSION: We found that both CR and birth cohorts influence future cognitive decline, which has strong public policy implications.

Obesity-Associated Neurodegeneration Pattern Mimics Alzheimer's Disease in an Observational Cohort Study.

BACKGROUND: Excess weight in adulthood leads to health complications such as diabetes, hypertension, or dyslipidemia. Recently, excess weight has also been related to brain atrophy and cognitive decline. Reports show that obesity is linked with Alzheimer's disease (AD)-related changes, such as cerebrovascular damage or amyloid-ß accumulation. However, to date no research has conducted a direct comparison between brain atrophy patterns in AD and obesity. OBJECTIVE: Here, we compared patterns of brain atrophy and amyloid-ß/tau protein accumulation in obesity and AD using a sample of over 1,300 individuals from four groups: AD patients, healthy controls, obese otherwise healthy individuals, and lean individuals. METHODS: We age- and sex-matched all groups to the AD-patients group and created cortical thickness maps of AD and obesity. This was done by comparing AD patients with healthy controls, and obese individuals with lean individuals. We then compared the AD and obesity maps using correlation analyses and permutation-based tests that account for spatial autocorrelation. Similarly, we compared obesity brain maps with amyloid-ß and tau protein maps from other studies. RESULTS: Obesity maps were highly correlated with AD maps but were not correlated with amyloid-ß/tau protein maps. This effect was not accounted for by the presence of obesity in the AD group. CONCLUSION: Our research confirms that obesity-related grey matter atrophy resembles that of AD. Excess weight management could lead to improved health outcomes, slow down cognitive decline in aging, and lower the risk for AD.

Normative data for the story recall subtest of the BEM-144 in the Quebec-French population aged 50 years and over.

OBJECTIVE: The story recall subtest of the Batterie d'Efficience Mnésique (BEM-144) is a verbal episodic memory test that assesses immediate and episodic memory. Variables such as age, sex, and education level can impact performance on this type of memory test, as can cultural differences. Therefore, the purpose of this study was to establish normative data for the story recall subtest of the BEM-144 in the elderly French-Quebec population. METHOD: The normative sample consisted of 260 healthy individuals aged 50-90 years, all from the province of Quebec, Canada. Analyses were performed to estimate the association between age, sex, and education level on one hand, and immediate and delayed recall performance, on the other hand. RESULTS: The results show that all sociodemographic variables are significantly associated with story recall performance. Normative data are proposed in the form of regression equations. CONCLUSIONS: Overall, these norms will be beneficial for the evaluation and detection of episodic memory impairment in middle-aged and older adults.

Data-Driven Analyses of Longitudinal Hippocampal Imaging Trajectories: Discrimination and Biomarker Prediction of Change Classes.

BACKGROUND: Hippocampal atrophy is a well-known biomarker of neurodegeneration, such as that observed in Alzheimer's disease (AD). Although distributions of hippocampal volume trajectories for asymptomatic individuals often reveal substantial heterogeneity, it is unclear whether interpretable trajectory classes can be objectively detected and used for prediction analyses. OBJECTIVE: To detect and predict hippocampal trajectory classes in a computationally competitive context using established AD-related risk factors/biomarkers. METHODS: We used biomarker/risk factor and longitudinal MRI data in asymptomatic adults from the AD Neuroimaging Initiative (n = 351; Mean = 75 years; 48.7% female). First, we applied latent class growth analyses to left (LHC) and right (RHC) hippocampal trajectory distributions to identify distinct classes. Second, using random forest analyses, we tested 38 multi-modal biomarkers/risk factors for their relative importance in discriminating the lower (potentially elevated atrophy risk) from the higher (potentially reduced risk) class. RESULTS: For both LHC and RHC trajectory distribution analyses, we observed three distinct trajectory classes. Three biomarkers/risk factors predicted membership in LHC and RHC lower classes: male sex, higher education, and lower plasma Aß1-42. Four additional factors selectively predicted membership in the lower LHC class: lower plasma tau and Aß1-40, higher depressive symptomology, and lower body mass index. CONCLUSION: Data-driven analyses of LHC and RHC trajectories detected three classes underlying the heterogeneous distributions. Machine learning analyses determined three common and four unique biomarkers/risk factors discriminating the higher and lower LHC/RHC classes. Our sequential analytic approach produced evidence that the dynamics of preclinical hippocampal trajectories can be predicted by AD-related biomarkers/risk factors from multiple modalities.

Deep learning of chest X-rays can predict mechanical ventilation outcome in ICU-admitted COVID-19 patients.

The COVID-19 pandemic repeatedly overwhelms healthcare systems capacity and forced the development and implementation of triage guidelines in ICU for scarce resources (e.g. mechanical ventilation). These guidelines were often based on known risk factors for COVID-19. It is proposed that image data, specifically bedside computed X-ray (CXR), provide additional predictive information on mortality following mechanical ventilation that can be incorporated in the guidelines. Deep transfer learning was used to extract convolutional features from a systematically collected, multi-institutional dataset of COVID-19 ICU patients. A model predicting outcome of mechanical ventilation (remission or mortality) was trained on the extracted features and compared to a model based on known, aggregated risk factors. The model reached a 0.702 area under the curve (95% CI 0.707-0.694) at predicting mechanical ventilation outcome from pre-intubation CXRs, higher than the risk factor model. Combining imaging data and risk factors increased model performance to 0.743 AUC (95% CI 0.746-0.732). Additionally, a post-hoc analysis showed an increase performance on high-quality than low-quality CXRs, suggesting that using only high-quality images would result in an even stronger model.

Tracking and predicting COVID-19 radiological trajectory on chest X-rays using deep learning.

Radiological findings on chest X-ray (CXR) have shown to be essential for the proper management of COVID-19 patients as the maximum severity over the course of the disease is closely linked to the outcome. As such, evaluation of future severity from current CXR would be highly desirable. We trained a repurposed deep learning algorithm on the CheXnet open dataset (224,316 chest X-ray images of 65,240 unique patients) to extract features that mapped to radiological labels. We collected CXRs of COVID-19-positive patients from an open-source dataset (COVID-19 image data collection) and from a multi-institutional local ICU dataset. The data was grouped into pairs of sequential CXRs and were categorized into three categories: 'Worse', 'Stable', or 'Improved' on the basis of radiological evolution ascertained from images and reports. Classical machine-learning algorithms were trained on the deep learning extracted features to perform immediate severity evaluation and prediction of future radiological trajectory. Receiver operating characteristic analyses and Mann-Whitney tests were performed. Deep learning predictions between "Worse" and "Improved" outcome categories and for severity stratification were significantly different for three radiological signs and one diagnostic ('Consolidation', 'Lung Lesion', 'Pleural effusion' and 'Pneumonia'; all P < 0.05). Features from the first CXR of each pair could correctly predict the outcome category between 'Worse' and 'Improved' cases with a 0.81 (0.74-0.83 95% CI) AUC in the open-access dataset and with a 0.66 (0.67-0.64 95% CI) AUC in the ICU dataset. Features extracted from the CXR could predict disease severity with a 52.3% accuracy in a 4-way classification. Severity evaluation trained on the COVID-19 image data collection had good out-of-distribution generalization when testing on the local dataset, with 81.6% of intubated ICU patients being classified as critically ill, and the predicted severity was correlated with the clinical outcome with a 0.639 AUC. CXR deep learning features show promise for classifying disease severity and trajectory. Once validated in studies incorporating clinical data and with larger sample sizes, this information may be considered to inform triage decisions.

[Perpetrators and victims of intimate partner violence: Personological profiles of people with borderline personality disorder]. / Violence conjugale commise et subie : profils personnologiques de personnes avec un trouble de personnalité limite.

Objective Personality disorders and intimate partner violence (IPV) are two problems recognized as major public health issues associated with serious individual and societal repercussions. Several studies have documented the links between borderline personality disorder (BPD) and IPV; however, we know very little about the specific pathological traits contributing to IPV. The study aims to document the phenomenon of IPV committed and suffered in persons with BPD and to draw profiles from the personality facets of the DSM-5 Alternative Model for Personality Disorders (AMPD). Method One hundred and eight BPD participants (83.3% female; Mage = 32.39, SD = 9.00) referred to a day hospital program following a crisis episode completed a battery of questionnaires including the French versions of the Revised Conflict Tactics Scales, evaluating physical and psychological IPV committed and suffered, and the Personality Inventory for the DSM-5- Faceted Brief Form, evaluating 25 pathological facets of personality. Results Among the participants, 78.7% report having committed psychological IPV, while 68.5% have been victims, which is more than the estimates published by the World Health Organization (27%). In addition, 31.5% would have committed physical IPV, while 22.2% would have been victims. IPV appears to be bidirectional since 85.9% of participants who are perpetrators of psychological IPV also report suffering from it and 52.9% of participants who are perpetrators of physical IPV report being also victims. Nonparametric group comparisons indicate that Hostility, Suspiciousness, Duplicity, Risk-Taking, and Irresponsibility facets distinguish physically and psychologically violent participants from nonviolent participants. High results on Hostility, Callousness, Manipulation, and Risk-taking facets characterize participants who are victims of psychological IPV, while an elevation in Hostility, Withdrawal, Avoidance of intimacy, and Risk-taking facets and a low result on the Submission facet distinguish participants who are victims of physical IPV from non-victims. Regression analyzes show that the Hostility facet alone explains a significant variance in the results of IPV perpetrated, while the Irresponsibility facet contributes substantially to the variance of the results of IPV experienced. Conclusion Results show the high prevalence of IPV in a sample of persons with BPD, as well as its bidirectional nature. Beyond the diagnosis of BPD, certain specific facets of the personality (including Hostility and Irresponsability) make it possible to target persons at greater risk of committing and suffering from psychological and physical IPV.

Histopathological Analysis of Cerebrovascular Lesions Associated With Aging.

Cerebrovascular disease (CVD) has been associated with cognitive impairment. Yet, our understanding of vascular contribution to cognitive decline has been limited by heterogeneity of definitions and assessment, as well as its occurrence in cognitively healthy aging. Therefore, we aimed to establish the natural progression of CVD associated with aging. We conducted a retrospective observational study of 63 cognitively healthy participants aged 19-84 years selected through the histological archives of the CHU de Québec. Assessment of CVD lesions was performed independently by 3 observers blinded to clinical data using the Vascular Cognitive Impairment Neuropathology Guidelines (VCING). We found moderate to almost perfect interobserver agreement for most regional CVD scores. Atherosclerosis (ρ = 0.758) and arteriolosclerosis (ρ = 0.708) showed the greatest significant association with age, followed by perivascular hemosiderin deposits (ρ = 0.432) and cerebral amyloid angiopathy (CAA; ρ = 0.392). Amyloid and tau pathologies were both associated with higher CVD load, but only CAA remained significantly associated with amyloid plaques after controlling for age. Altogether, these findings support the presence of multiple CVD lesions in the brains of cognitively healthy adults, the burden of which increases with age and can be quantified in a reproducible manner using standardized histological scales such as the VCING.

Birth Cohorts and Cognitive Reserve Influence Cognitive Performances in Older Adults.

BACKGROUND: Evidence suggests birth cohort differences in cognitive performance of older adults. Proxies of cognitive reserve (CR), such as educational attainment and occupational complexity, could also partly account for these differences as they are influenced by the sociocultural environment of the birth cohorts. OBJECTIVE: To predict cognitive performance using birth cohorts and CR and examine the moderating influence of CR on cognitive performance and structural brain health association. METHODS: Using ADNI data (n = 1628), four birth cohorts were defined (1915-1928; 1929-1938; 1939-1945; 1946-1964). CR proxies were education, occupational complexity, and verbal IQ. We predicted baseline cognitive performances (verbal episodic memory; language and semantic memory; attention capacities; executive functions) using multiple linear regressions with CR, birth cohorts, age, structural brain health (total brain volume; total white matter hyperintensities volume) and vascular risk factors burden as predictors. Sex and CR interactions were also explored. RESULTS: Recent birth cohorts, higher CR, and healthier brain structures predicted better performance in verbal episodic memory, language and semantic memory, and attention capacities, with large effect sizes. Better performance in executive functions was predicted by a higher CR and a larger total brain volume, with a small effect size. With equal score of CR, women outperformed men in verbal episodic memory and language and semantic memory in all cohorts. Higher level of CR predicted better performance in verbal episodic memory, only when total brain volume was lower. CONCLUSION: Cohort differences in cognitive performance favor more recent birth cohorts and suggests that this association may be partly explained by proxies of CR.

Childhood Socioeconomic Status Does Not Predict Late-Life Cognitive Decline in the 1936 Lothian Birth Cohort.

This study examined childhood socioeconomic status (SES) as a predictor of later life cognitive decline. Data came from 519 participants in the Lothian Birth Cohort 1936 (LBC1936) study. SES measures at 11 years of age included parental educational attainment, father's occupational status, household characteristics and a composite measure of global childhood SES (i.e., a total of low SES childhood indicators). Cognitive abilities were assessed by the Mini-Mental State Exam at ages 69.8, 72.8 and 76.7 years. Most indicators of low childhood SES (i.e., father manual worker, less than secondary school father education, household overcrowding, exterior located toilet, and global childhood SES) did not predict cognitive decline between the ages of 69.8 and 76.7. Participants with less educated mothers showed an increase in cognitive decline (ß = -0.132, p = 0.048, and CI = -0.80, -0.00). The relationship between maternal educational attainment and cognitive decline became non-significant when controlling for adult SES (i.e., participant educational attainment and occupation). Adult SES did not mediate the latter relationship. This study provides new evidence that childhood SES alone is not strongly associated with cognitive decline. New knowledge is critical to improving population health by identifying life span stages in which interventions might be effective in preventing cognitive decline.

Beware of white matter hyperintensities causing systematic errors in FreeSurfer gray matter segmentations!

Volumetric estimates of subcortical and cortical structures, extracted from T1-weighted MRIs, are widely used in many clinical and research applications. Here, we investigate the impact of the presence of white matter hyperintensities (WMHs) on FreeSurfer gray matter (GM) structure volumes and its possible bias on functional relationships. T1-weighted images from 1,077 participants (4,321 timepoints) from the Alzheimer's Disease Neuroimaging Initiative were processed with FreeSurfer version 6.0.0. WMHs were segmented using a previously validated algorithm on either T2-weighted or Fluid-attenuated inversion recovery images. Mixed-effects models were used to assess the relationships between overlapping WMHs and GM structure volumes and overall WMH burden, as well as to investigate whether such overlaps impact associations with age, diagnosis, and cognitive performance. Participants with higher WMH volumes had higher overlaps with GM volumes of bilateral caudate, cerebral cortex, putamen, thalamus, pallidum, and accumbens areas (p < .0001). When not corrected for WMHs, caudate volumes increased with age (p < .0001) and were not different between cognitively healthy individuals and age-matched probable Alzheimer's disease patients. After correcting for WMHs, caudate volumes decreased with age (p < .0001), and Alzheimer's disease patients had lower caudate volumes than cognitively healthy individuals (p < .01). Uncorrected caudate volume was not associated with ADAS13 scores, whereas corrected lower caudate volumes were significantly associated with poorer cognitive performance (p < .0001). Presence of WMHs leads to systematic inaccuracies in GM segmentations, particularly for the caudate, which can also change clinical associations. While specifically measured for the Freesurfer toolkit, this problem likely affects other algorithms.

Disappearing metabolic youthfulness in the cognitively impaired female brain.

Sex differences play a vital role in human brain structure and physiology. Previous reports have proposed evidence hinting at a metabolic advantage in female brains across adulthood. It remained to be determined whether this advantage would be maintained across the spectrum of cognitive impairment, up to and including dementia due to Alzheimer's disease (AD). Here, using a machine-learning algorithm, we explore sex differences in metabolic brain-age derived from fluorodeoxyglucose positron emission tomography imaging among cognitively healthy individuals and those affected by mild cognitive impairment and clinically probable AD. First, we report that cognitively healthy male participants showed a persistently "older" looking brains when compared to healthy female participants in term of metabolic brain age, confirming earlier reports. However, this distinction disappeared among MCI individuals and probable AD patients, and this loss could not be explained by an accompanying neurodegeneration. This would seem to indicate that females have a higher rate of decline in brain glucose metabolism when cognitively impaired to negate their prior advantage.

Patch-wise brain age longitudinal reliability.

We recently introduced a patch-wise technique to estimate brain age from anatomical T1-weighted magnetic resonance imaging (T1w MRI) data. Here, we sought to assess its longitudinal reliability by leveraging a unique dataset of 99 longitudinal MRI scans from a single, cognitively healthy volunteer acquired over a period of 17 years (aged 29-46 years) at multiple sites. We built a robust patch-wise brain age estimation framework on the basis of 100 cognitively healthy individuals from the MindBoggle dataset (aged 19-61 years) using the Desikan-Killiany-Tourville atlas, then applied the model to the volunteer dataset. The results show a high prediction accuracy on the independent test set (R2 = .94, mean absolute error of 0.63 years) and no statistically significant difference between manufacturers, suggesting that the patch-wise technique has high reliability and can be used for longitudinal multi-centric studies.

Associating Type 2 Diabetes Risk Factor Genes and FDG-PET Brain Metabolism in Normal Aging and Alzheimer's Disease.

Background: Several studies have linked type 2 diabetes (T2D) to an increased risk of developing Alzheimer's disease (AD). This has led to an interest in using antidiabetic treatments for the prevention of AD. However, the underlying mechanisms explaining the relationship between T2D and AD have not been completely elucidated. Objective: Our objective was to examine cerebral 18F-fluorodeoxyglucose (FDG) uptake during normal aging and in AD patients in regions associated with diabetes genetic risk factor expression to highlight which genes may serve as potential targets for pharmaceutical intervention. Methods: We calculated regional glucose metabolism differences in units of standardized uptake values (SUVR) for 386 cognitively healthy adults and 335 clinically probable AD patients. We then proceeded to extract gene-expression data from the publicly available Allen Human Brain Atlas (HBA) database. We used the nearest genes to 46 AD- and T2D-associated SNPs previously identified in the literature, and mapped their expression to the same 34 cortical regions in which we calculated SUVRs. SNPs with a donor consistency of 0.40 or greater were selected for further analysis. We evaluated the associations between SUVR and gene-expression across the brain. Results: Of the 46 risk-factor genes, 15 were found to be significantly correlated with FDG-PET brain metabolism in healthy adults and probable AD patients after correction for multiple comparisons. Using multiple regression, we found that five genes explained a total of 72.5% of the SUVR variance across the healthy adult group regions, while four genes explained a total of 79.3% of the SUVR variance across the probable AD group regions. There were significant differences in whole-brain SUVR as a function of allele frequencies for two genes. Conclusions: These results highlight the association between risk factor genes for T2D and regional glucose metabolism during both normal aging and in probable AD. Highlighted genes were associated with mitochondrial stability, vascular maintenance, and glucose intolerance. Pharmacological intervention of these pathways has the potential to improve glucose metabolism during normal again as well as in AD patients.

Selection of the optimal intensity normalization region for FDG-PET studies of normal aging and Alzheimer's disease.

The primary method for measuring brain metabolism in humans is positron emission tomography (PET) imaging using the tracer 18F-fluorodeoxyglucose (FDG). Standardized uptake value ratios (SUVR) are commonly calculated from FDG-PET images to examine intra- and inter-subject effects. Various reference regions are used in the literature of FDG-PET studies of normal aging, making comparison between studies difficult. Our primary objective was to determine the optimal SUVR reference region in the context of healthy aging, using partial volume effect (PVE) and non-PVE corrected data. We calculated quantitative cerebral metabolic rates of glucose (CMRg) from PVE-corrected and non-corrected images from young and older adults. We also investigated regional atrophy using magnetic resonance (MR) images. FreeSurfer 6.0 atlases were used to explore possible reference regions of interest (ROI). Multiple regression was used to predict CMRg data, in each FreeSurfer ROI, with age and sex as predictors. Age had the least effect in predicting CMRg for PVE corrected data in the pons (r2 = 2.83 × 10-3, p = 0.67). For non-PVE corrected data age also had the least effect in predicting CMRg in the pons (r2 = 3.12 × 10-3, p = 0.67). We compared the effects of using the whole brain or the pons as a reference region in PVE corrected data in two regions susceptible to hypometabolism in Alzheimer's disease, the posterior cingulate and precuneus. Using the whole brain as a reference region resulted in non-significant group differences in the posterior cingulate while there were significant differences between all three groups in the precuneus (all p < 0.004). When using the pons as a reference region there was significant differences between all groups for both the posterior cingulate and the precuneus (all p < 0.001). Therefore, the use of the pons as a reference region is more sensitive to hypometabism changes associated with Alzheimer's disease than the whole brain.

A dataset of long-term consistency values of resting-state fMRI connectivity maps in a single individual derived at multiple sites and vendors using the Canadian Dementia Imaging Protocol.

The impact of multisite acquisition on resting-state functional MRI (rsfMRI) connectivity has recently gained attention. We provide consistency values (Pearson's correlation) between rsfMRI connectivity maps of an adult volunteer (Csub) scanned 25 times over 3.5 years at 13 sites using the Canadian Dementia Imaging Protocol (CDIP, www.cdip-pcid.ca). This dataset was generated as part of the following article: Multivariate consistency of resting-state fMRI connectivity maps acquired on a single individual over 2.5 years, 13 sites and 3 vendors [1]. Acquired on three 3T scanner vendors (GE, Siemens and Philips), the Csub dataset is part of an ongoing effort to monitor the quality and comparability of MRI data collected across the Canadian Consortium on Neurodegeneration in Aging (CCNA) imaging network. The participant was scanned 25 times in the above-mentioned article: multiple times at six sites over a period of 2.5 years, and once at the remaining seven sites. Since then the participant was scanned an additional 45 times, allowing us to extend the dataset to 70 rsfMRI scans over a period of >4 years. In addition, we provide intra- and inter-subject consistency values of rsfMRI connectivity maps derived from 26 adult participants belonging to the publicly released Hangzhou Normal University dataset (HNU1). All HNU1 participants underwent 10 rsfMRI scans over one month on a single 3T scanner (GE). Connectivity maps of seven canonical networks were generated for each scan in the two datasets (Csub and HNU1). All consistency values, along with the scripts used to preprocess the rsfMRI data and generate connectivity maps and pairwise consistency values, have been made available on two public repositories, Github and Zenodo. We have also made available four Jupyter notebooks that use the provided consistency values to (a) generate interactive graphical summaries - 1 notebook, (b) perform statistical analyses - 2 notebooks, and (c) perform data-driven cluster analysis for the recovery of subject identity (i.e. rsfMRI fingerprinting) - 1 notebook. In addition, we provide two interactive dashboards that allow visualization of individual connectivity maps from the two datasets. Finally, we also provide minimally preprocessed rsfMRI data in Brain Imaging Data Standard (BIDS) format on all 70 scans in the extended dataset.

Multivariate consistency of resting-state fMRI connectivity maps acquired on a single individual over 2.5 years, 13 sites and 3 vendors.

Studies using resting-state functional magnetic resonance imaging (rsfMRI) are increasingly collecting data at multiple sites in order to speed up recruitment or increase sample size. The main objective of this study was to assess the long-term consistency of rsfMRI connectivity maps derived at multiple sites and vendors using the Canadian Dementia Imaging Protocol (CDIP, www.cdip-pcid.ca). Nine to 10 min of functional BOLD images were acquired from an adult cognitively healthy volunteer scanned repeatedly at 13 Canadian sites on three scanner makes (General Electric, Philips and Siemens) over the course of 2.5 years. The consistency (spatial Pearson's correlation) of rsfMRI connectivity maps for seven canonical networks ranged from 0.3 to 0.8, with a negligible effect of time, but significant site and vendor effects. We noted systematic differences in data quality (i.e. head motion, number of useable time frames, temporal signal-to-noise ratio) across vendors, which may also confound some of these results, and could not be disentangled in this sample. We also pooled the long-term longitudinal data with a single-site, short-term (1 month) data sample acquired on 26 subjects (10 scans per subject), called HNU1. Using randomly selected pairs of scans from each subject, we quantified the ability of a data-driven unsupervised cluster analysis to match two scans of the same subjects. In this "fingerprinting" experiment, we found that scans from the Canadian subject (Csub) could be matched with high accuracy intra-site (>95% for some networks), but that the accuracy decreased substantially for scans drawn from different sites and vendors, even falling outside of the range of accuracies observed in HNU1. Overall, our results demonstrate good multivariate stability of rsfMRI measures over several years, but substantial impact of scanning site and vendors. How detrimental these effects are will depend on the application, yet our results demonstrate that new methods for harmonizing multisite analysis represent an important area for future work.

Bias-adjustment in neuroimaging-based brain age frameworks: A robust scheme.

The level of prediction error in the brain age estimation frameworks is associated with the authenticity of statistical inference on the basis of regression models. In this paper, we present an efficacious and plain bias-adjustment scheme using chronological age as a covariate through the training set for downgrading the prediction bias in a Brain-age estimation framework. We applied proposed bias-adjustment scheme coupled by a machine learning-based brain age framework on a large set of metabolic brain features acquired from 675 cognitively unimpaired adults through fluorodeoxyglucose positron emission tomography data as the training set to build a robust Brain-age estimation framework. Then, we tested the reliability of proposed bias-adjustment scheme on 75 cognitively unimpaired adults, 561 mild cognitive impairment patients as well as 362 Alzheimer's disease patients as independent test sets. Using the proposed method, we gained a strong R2 of 0.81 between the chronological age and brain estimated age, as well as an excellent mean absolute error of 2.66 years on 75 cognitively unimpaired adults as an independent set; whereas an R2 of 0.24 and a mean absolute error of 4.71 years was achieved without bias-adjustment. The simulation results demonstrated that the proposed bias-adjustment scheme has a strong capability to diminish prediction error in brain age estimation frameworks for clinical settings.