Baseline and longitudinal patterns of brain atrophy in MCI patients, and their use in prediction of short-term conversion to AD: results from ADNI

Longitudinal dementia trajectories for Alzheimer's Disease characterization and prediction

BACKGROUND

Alzheimer's Disease (AD) remains one of the most significant neurodegenerative diseases globally, affecting approximately 38.5 million people in 2023. Early identification of individuals at risk of developing AD is essential to managing disease progression and implementing timely interventions. Despite extensive research on predicting individual AD progression and diagnosis conversion, the dynamic and multimodal characterization of AD evolution remains an open challenge.

APPROACH

This study introduces a model leveraging multimodal clinical data to capture and predict both population-level and individual trajectories of dementia over extended timescales. We define a Static Dementia Score (SDS) as a metric representing the current state of dementia, with its temporal evolution modeled using a Bayesian nonlinear mixed-effects approach. This method generates a single continuous curve that delineates the average dementia progression pattern across the population. Additionally, our approach allows for patient-specific SDS progression modeling, enabling the prediction of individual dementia trajectories over a specified time horizon.

RESULTS

The model was trained on a dataset comprising 5,033 observations, including MP-RAGE MRI scans, cognitive assessments, and demographic data from 883 individuals with at least four observations and stable diagnostic trajectories. Population-level curves produced by our model align with established AD progression trends, capturing key stages of disease evolution. Individual-specific deformation parameters were effective in characterizing personalized disease progression, achieving a theoretical diagnosis prediction accuracy of 0.952±0.013 three years in advance and an AD conversion prediction accuracy of 0.916±0.022 within the same period.

CONCLUSION

Our findings underscore the potential of this approach for AD classification and forecasting disease progression. These results emphasize its utility for clinicians and researchers in detecting atypical disease trajectories early in patients with longitudinal follow-up, enhancing decision-making and therapeutic planning.

Cumulative anticholinergic exposure and white matter hyperintensity burden in community-dwelling older adults

BACKGROUND

Anticholinergic exposure is associated with dementia risk; however, the mechanisms for this association remain unclear. The objective of this study was to examine the association between anticholinergic exposure and white matter hyperintensity (WMH) burden.

METHODS

This was a retrospective analysis of data from the Adult Changes in Thought (ACT) study, a prospective cohort study among adults aged ≥65 years on dementia risk factors. We used data collected through March 2020 for this analysis. The sample included ACT participants who were referred for and had a clinical magnetic resonance imaging (MRI) scan and ≥10 years of continuous healthcare enrollment prior to the scan. Our primary exposure was total standardized daily dose (TSDD) of anticholinergics. Outcomes included three semi-quantitative ratings of WMH volume. We used separate linear regression models for each outcome to estimate and compare covariate-adjusted mean values of WMH ratings in each exposure group.

RESULTS

Of the 1043 individuals included in the analyses, 28% had no use, 33% had 1-90 TSDD, 15% had 91-365 TSDD, 7% had 366-1095 TSDD, and 17% had ≥1096 TSDD. The mean age was 81 years, most were female (58%) and White race (88%). Compared to those with no use, the ≥1096 TSDD group had a higher (worse) adjusted mean [95% confidence intervals] Fazekas (4.0 [3.8, 4.2] vs. 3.4 [3.2, 3.5]; p: <0.001), Modified Scheltens (14.3 [13.4, 15.2] vs. 12.2 [11.5, 12.9]; p: <0.001), and Age-Related White Matter Changes (5.6 [5.3, 6.0] vs. 4.8 [4.5, 5.1]; p = 0.001). A dose-response relationship was not found.

CONCLUSIONS

The highest anticholinergic exposure was associated with greater WMH burden. Future studies should focus on longitudinal changes of WMH burden to better understand the biological mechanisms underlying the link between anticholinergics and dementia risk.

Weighted Multi-Modal Contrastive Learning Based Hybrid Network for Alzheimer's Disease Diagnosis

Multiple imaging modalities and specific proteins in the cerebrospinal fluid, providing a comprehensive understanding of neurodegenerative disorders, have been widely used for computer-aided diagnosis of Alzheimer's disease (AD). Given the proven effectiveness of contrastive learning in aligning multi-modal representation, in this paper, we investigate effective contrastive learning strategies to learn better cross-modal representations for the integration of multi-modal complementary information. To enhance the overall performance in AD diagnosis, we construct a unified hybrid network that integrates feature learning and classifier learning into an end-to-end framework. Specifically, we propose a weighted multi-modal contrastive learning based on hybrid network (WMCL-HN) method. Firstly, an adaptive weighted strategy is implemented on the multi-modal contrastive learning to dynamically regulate the degree of information exchange across modalities. It assigns higher weights to more important modality pair, thus the most important underlying relationships across modalities can be captured. Secondly, we construct a hybrid network, which employs a curriculum learning strategy that gradually transitions the training from feature learning to classifier learning, ensuring that the learned features are tailored to the diagnostic task. Experimental results on ADNI dataset demonstrate the effectiveness of the proposed WMCL-HN in AD-related diagnosis tasks. The source code is available at https://github.com/pcehnago/WMCL-HN.

Baseline Brain Volumes Predict Future Brain Atrophy in Mild Cognitive Impairment: A Tensor-based Morphometry Study of the Alzheimer Continuum

OBJECTIVE

Prognostic evaluation of patients with mild cognitive impairment (MCI) is of great importance, and magnetic resonance imaging, as a readily available modality, can play a pivotal role in this field.

METHODS

Using the Alzheimer Disease Neuroimaging Initiative database, we conducted a retrospective longitudinal study of the associations between volumetric brain magnetic resonance imaging and cognitive composite scores in all domains (memory, executive function, language, and visuospatial) with annual whole-brain atrophy based on tensor-based morphometry (TBM) scores among patients with MCI and healthy controls (HCs). The Reliable Change Index was further used to categorize patients into 2 groups including (1) patients with meaningful 1-year reliable cognitive changes [reliable change (RC) group] and (2) patients without (non-RC).

RESULTS

One hundred thirty-seven patients with MCI and 132 HCs were enrolled. The 2 groups showed no significant differences in age, sex, and apolipoprotein E4 expression (P > 0.05). Based on the TBM score, patients with MCI had more significant 1-year brain volume loss than HCs (P < 0.001). After multiple comparison corrections, the 1-year TBM atrophy score was positively correlated with baseline whole brain (P = 0.03), hippocampus (P < 0.0001), entorhinal (P < 0.0001), and middle temporal (P < 0.0001) volumes among MCI patients, indicating that lower volumes in these regions were associated with greater 1-year atrophy rates. Regression analyses showed a positive correlation between baseline and 1-year memory composite scores and annual brain atrophy rate in MCI patients (P = 0.01, 0.04), demonstrating that lower cognitive scores were associated with a greater annual atrophy rate. However, the correlations no longer held significance after correction for multiple comparison (P = 0.05, 0.17). MCI participants with RCs in language composite scores initially had significantly greater brain atrophy than those without (P = 0.03, corrected P = 0.06). However, TBM scores showed no significant differences between RC and non-RC groups for other composite scores (P > 0.05).

CONCLUSIONS

Lower baseline volumes in multiple brain regions of MCI are associated with greater annual brain volume loss based on TBM, suggesting TBM as a potential imaging marker for conventional volumetric studies in MCI. Further research is needed to explore the link between cognitive scores and the application of Reliable Change Index in TBM imaging across the Alzheimer disease spectrum.

BPEN: Brain Posterior Evidential Network for trustworthy brain imaging analysis

The application of deep learning techniques to analyze brain functional magnetic resonance imaging (fMRI) data has led to significant advancements in identifying prospective biomarkers associated with various clinical phenotypes and neurological conditions. Despite these achievements, the aspect of prediction uncertainty has been relatively underexplored in brain fMRI data analysis. Accurate uncertainty estimation is essential for trustworthy learning, given the challenges associated with brain fMRI data acquisition and the potential diagnostic implications for patients. To address this gap, we introduce a novel posterior evidential network, named the Brain Posterior Evidential Network (BPEN), designed to capture both aleatoric and epistemic uncertainty in the analysis of brain fMRI data. We conducted comprehensive experiments using data from the Alzheimer's Disease Neuroimaging Initiative (ADNI) and ADNI-depression (ADNI-D) cohorts, focusing on predictions for mild cognitive impairment (MCI) and depression across various diagnostic groups. Our experiments not only unequivocally demonstrate the superior predictive performance of our BPEN model compared to existing state-of-the-art methods but also underscore the importance of uncertainty estimation in predictive models.

The impact of sleep and exercise on brain atrophy in mild cognitive impairment

Chronic sleep deprivation and lack of physical exercise may have detrimental effects on overall health, particularly in terms of brain health, with significant implications for cognitive function and well-being. This review explores the impact of chronic sleep deprivation and physical exercise on brain atrophy in mild cognitive impairment (MCI) and Alzheimer's disease (AD). Drawing insights from 40 selected studies, the review synthesizes evidence on these lifestyle factors' correlations with neurodegenerative changes. Chronic sleep deprivation disrupts circadian rhythms and neurochemical pathways, potentially accelerating brain atrophy, while physical exercise preserves brain structure by enhancing vascular health, reducing inflammation, and supporting synaptic plasticity, particularly in regions like the hippocampus. Results highlight distinct patterns of brain atrophy in AD and MCI, underscoring the potential for targeted interventions to mitigate cognitive decline. Understanding the relationship between sleep disruption and brain health provides insights into strategies for possibly delaying neurodegenerative diseases like MCI, which represents a milder form of Alzheimer's, and AD. The findings underscore the potential utility of integrating sleep therapy and physical exercise interventions in clinical practice for early detection of mild cognitive impairment and potentially delaying disease progression. This integrated approach has been found to promote healthy aging, reduce atrophy rates, and enhance cognitive resilience across aging populations.

Changes in the Parietal Lobe Subregion Volume at Various Stages of Alzheimer's Disease and the Role in Cognitively Normal and Mild Cognitive Impairment Conversion

BACKGROUND

Volume alterations in the parietal subregion have received less attention in Alzheimer's disease (AD), and their role in predicting conversion of mild cognitive impairment (MCI) to AD and cognitively normal (CN) to MCI remains unclear. In this study, we aimed to assess the volumetric variation of the parietal subregion at different cognitive stages in AD and to determine the role of parietal subregions in CN and MCI conversion.

METHODS

We included 662 participants from the Alzheimer's Disease Neuroimaging Initiative (ADNI) database, including 228 CN, 221 early MCI (EMCI), 112 late MCI (LMCI), and 101 AD participants. We measured the volume of the parietal subregion based on the Human Brainnetome Atlas (BNA-246) using voxel-based morphometry among individuals at various stages of AD and the progressive and stable individuals in CN and MCI. We then calculated the area under the curve (AUC) of the receiver operating characteristic (ROC) curve to test the ability of parietal subregions to discriminate between different cognitive groups. The Cox proportional hazard model was constructed to determine which specific parietal subregions, alone or in combination, could be used to predict progression from MCI to AD and CN to MCI. Finally, we examined the relationship between the cognitive scores and parietal subregion volume in the diagnostic groups.

RESULTS

The left inferior parietal lobule (IPL)_6_5 (rostroventral area 39) showed the best ability to discriminate between patients with AD and those with CN (AUC = 0.688). The model consisting of the left IPL_6_4 (caudal area 40) and bilateral IPL_6_5 showed the best combination for predicting the CN progression to MCI. The left IPL_6_1 (caudal area 39) showed the best predictive power in predicting the progression of MCI to AD. Certain subregions of the volume correlated with cognitive scales.

CONCLUSION

Subregions of the angular gyrus are essential in the early onset and subsequent development of AD, and early detection of the volume of these regions may be useful in identifying the tendency to develop the disease and its treatment.

LiveDrive AI: A Pilot Study of a Machine Learning-Powered Diagnostic System for Real-Time, Non-Invasive Detection of Mild Cognitive Impairment

Alzheimer's disease (AD) represents a significant global health issue, affecting over 55 million individuals worldwide, with a progressive impact on cognitive and functional abilities. Early detection, particularly of mild cognitive impairment (MCI) as an indicator of potential AD onset, is crucial yet challenging, given the limitations of current diagnostic biomarkers and the need for non-invasive, accessible tools. This study aims to address these gaps by exploring driving performance as a novel, non-invasive biomarker for MCI detection. Using the LiveDrive AI system, equipped with multimodal sensing (MMS) technology and a driving performance assessment strategy, the proposed work analyzes the predictive capacity of driving patterns in indicating cognitive decline. Machine learning models, trained on an expert-annotated in-house dataset, were employed to detect MCI status from driving performance. Key findings demonstrate the feasibility of using nuanced driving features, such as velocity and acceleration during turning, as indicators of cognitive decline. This approach holds promise for integration into smartphone or car applications, enabling real-time, continuous cognitive health monitoring. The implications of this work suggest a transformative step towards scalable, real-world solutions for early AD diagnosis, with the potential to improve patient outcomes and disease management.

Early diagnosis of Alzheimer's disease and mild cognitive impairment using MRI analysis and machine learning algorithms

Early diagnosis of Alzheimer's disease (AD) and mild cognitive impairment (MCI) is crucial to prevent their progression. In this study, we proposed the analysis of magnetic resonance imaging (MRI) based on features including; hippocampus (HC) area size, HC grayscale statistics and texture features (mean, standard deviation, skewness, kurtosis, contrast, correlation, energy, homogeneity, entropy), lateral ventricle (LV) area size, gray matter area size, white matter area size, cerebrospinal fluid area size, patient age, weight, and cognitive score. Five machine learning classifiers; K-nearest neighborhood (KNN), support vector machine (SVM), random forest (RF), decision tree (DT), and multi-layer perception (MLP) were used to distinguish between groups: cognitively normal (CN) vs AD, early MCI (EMCI) vs late MCI (LMCI), CN vs EMCI, CN vs LMCI, AD vs EMCI, and AD vs LMCI. Additionally, the correlation and dependence were calculated to examine the strength and direction of association between each extracted feature and each classification of the group. The average classification accuracies in 20 trials were 95% (SVM), 71.50% (RF), 82.58% (RF), 84.91% (SVM), 85.83% (RF), and 85.08% (RF), respectively, with the best accuracies being 100% (SVM, RF, and MLP), 83.33% (RF), 91.66% (RF), 95% (SVM, and MLP), 96.66% (RF), and 93.33% (DT). Cognitive scores, HC and LV area sizes, and HC texture features demonstrated significant potential for diagnosing AD and its subtypes for all groups. RF and SVM showed better performance in distinguishing between groups. These findings highlight the importance of using 2D-MRI to identify key features containing critical information for early diagnosis of AD.

Supplementary Information

The online version contains supplementary material available at 10.1007/s42452-024-06440-w.

Jointly constrained group sparse connectivity representation improves early diagnosis of Alzheimer's disease on routinely acquired T1-weighted imaging-based brain network

Background

Radiomics-based morphological brain networks (radMBN) constructed from routinely acquired structural MRI (sMRI) data have gained attention in Alzheimer's disease (AD). However, the radMBN suffers from limited characterization of AD because sMRI only characterizes anatomical changes and is not a direct measure of neuronal pathology or brain activity.

Purpose

To establish a group sparse representation of the radMBN under a joint constraint of group-level white matter fiber connectivity and individual-level sMRI regional similarity (JCGS-radMBN).

Methods

Two publicly available datasets were adopted, including 120 subjects from ADNI with both T1-weighted image (T1WI) and diffusion MRI (dMRI) for JCGS-radMBN construction, 818 subjects from ADNI and 200 subjects solely with T1WI from AIBL for validation in early AD diagnosis. Specifically, the JCGS-radMBN was conducted by jointly estimating non-zero connections among subjects, with the regularization term constrained by group-level white matter fiber connectivity and individual-level sMRI regional similarity. Then, a triplet graph convolutional network was adopted for early AD diagnosis. The discriminative brain connections were identified using a two-sample t-test, and the neurobiological interpretation was validated by correlating the discriminative brain connections with cognitive scores.

Results

The JCGS-radMBN exhibited superior classification performance over five brain network construction methods. For the typical NC vs. AD classification, the JCGS-radMBN increased by 1-30% in accuracy over the alternatives on ADNI and AIBL. The discriminative brain connections exhibited a strong connectivity to hippocampus, parahippocampal gyrus, and basal ganglia, and had significant correlation with MMSE scores.

Conclusion

The proposed JCGS-radMBN facilitated the AD characterization of brain network established on routinely acquired imaging modality of sMRI.

Supplementary Information

The online version of this article (10.1007/s13755-023-00269-0) contains supplementary material, which is available to authorized users.

Functional Connectome Controllability in Patients with Mild Cognitive Impairment after Repetitive Transcranial Magnetic Stimulation of the Dorsolateral Prefrontal Cortex

Background: Repetitive transcranial magnetic stimulation (rTMS) has shown therapeutic effects in neurological patients by inducing neural plasticity. In this pilot study, we analyzed the modifying effects of high-frequency (HF-)rTMS applied to the dorsolateral prefrontal cortex (DLPFC) of patients with mild cognitive impairment (MCI) using an advanced approach of functional connectome analysis based on network control theory (NCT). Methods: Using local-to-global functional parcellation, average and modal controllability (AC/MC) were estimated for DLPFC nodes of prefrontal-lateral control networks (R/LH_Cont_PFCl_3/4) from a resting-state fMRI series acquired at three time points (T0 = baseline, T1 = T0 + 4 weeks, T2 = T1 + 20 weeks) in MCI patients receiving regular daily sessions of 10 Hz HF-rTMS (n = 10, 68.00 ± 8.16 y, 4 males) or sham (n = 10, 63.80 ± 9.95 y, 5 males) stimulation, between T0 and T1. Longitudinal (group) effects on AC/MC were assessed with non-parametric statistics. Spearman correlations (ρ) of AC/MC vs. neuropsychological (RBANS) score %change (at T1, T2 vs. T0) were calculated. Results: AC median was reduced in MCI-rTMS, compared to the control group, for RH_Cont_PFCl_3/4 at T1 and T2 (vs. T0). In MCI-rTMS patients, for RH_Cont_PFCl_3, AC % change at T1 (vs. T0) was negatively correlated with semantic fluency (ρ = -0.7939, p = 0.045) and MC % change at T2 (vs. T0) was positively correlated with story memory (ρ = 0.7416, p = 0.045). Conclusions: HF-rTMS stimulation of DLFC nodes significantly affects the controllability of the functional connectome in MCI patients. Emerging correlations between AC/MC controllability and cognitive performance changes, immediately (T1 vs. T0) and six months (T2 vs. T0) after treatment, suggest NCT could help explain the HF-rTMS impact on prefrontal-lateral control network, monitoring induced neural plasticity effects in MCI patients.

Tensor Completion for Alzheimer's Disease Prediction From Diffusion Tensor Imaging

OBJECTIVE

Alzheimer's disease (AD) is a slowly progressive neurodegenerative disorder with insidious onset. Accurate prediction of the disease progression has received increasing attention. Cognitive scores that reflect patients' cognitive status have become important criteria for predicting AD. Most existing methods consider the relationship between neuroimages and cognitive scores to improve prediction results. However, the inherent structure information in interrelated cognitive scores is rarely considered.

METHOD

In this article, we propose a relation-aware tensor completion multitask learning method (RATC-MTL), in which the cognitive scores are represented as a third-order tensor to preserve the global structure information in clinical scores. We combine both tensor completion and linear regression into a unified framework, which allows us to capture both inter and intra modes correlations in cognitive tensor with a low-rank constraint, as well as incorporate the relationship between biological features and cognitive status by imposing a regression model on multiple cognitive scores.

RESULT

Compared to the single-task and state-of-the-art multi-task algorithms, our proposed method obtains the best results for predicting cognitive scores in terms of four commonly used metrics. Furthermore, the overall performance of our method in classifying AD progress is also the best.

CONCLUSION

Our results demonstrate the effectiveness of the proposed framework in fully exploring the global structure information in cognitive scores.

SIGNIFICANCE

This study introduces a novel concept of leveraging tensor completion to assist in disease diagnoses, potentially offering a solution to the issue of data scarcity encountered in prolonged monitoring scenarios.

Predicting conversion from mild cognitive impairment to Alzheimer's disease: a multimodal approach

Successively predicting whether mild cognitive impairment patients will progress to Alzheimer's disease is of significant clinical relevance. This ability may provide information that can be leveraged by emerging intervention approaches and thus mitigate some of the negative effects of the disease. Neuroimaging biomarkers have gained some attention in recent years and may be useful in predicting the conversion of mild cognitive impairment to Alzheimer's disease. We implemented a novel multi-modal approach that allowed us to evaluate the potential of different imaging modalities, both alone and in different degrees of combinations, in predicting the conversion to Alzheimer's disease of mild cognitive impairment patients. We applied this approach to the imaging data from the Alzheimer's Disease Neuroimaging Initiative that is a multi-modal imaging dataset comprised of MRI, Fluorodeoxyglucose PET, Florbetapir PET and diffusion tensor imaging. We included a total of 480 mild cognitive impairment patients that were split into two groups: converted and stable. Imaging data were segmented into atlas-based regions of interest, from which relevant features were extracted for the different imaging modalities and used to construct machine-learning models to classify mild cognitive impairment patients into converted or stable, using each of the different imaging modalities independently. The models were then combined, using a simple weight fusion ensemble strategy, to evaluate the complementarity of different imaging modalities and their contribution to the prediction accuracy of the models. The single-modality findings revealed that the model, utilizing features extracted from Florbetapir PET, demonstrated the highest performance with a balanced accuracy of 83.51%. Concerning multi-modality models, not all combinations enhanced mild cognitive impairment conversion prediction. Notably, the combination of MRI with Fluorodeoxyglucose PET emerged as the most promising, exhibiting an overall improvement in predictive capabilities, achieving a balanced accuracy of 78.43%. This indicates synergy and complementarity between the two imaging modalities in predicting mild cognitive impairment conversion. These findings suggest that β-amyloid accumulation provides robust predictive capabilities, while the combination of multiple imaging modalities has the potential to surpass certain single-modality approaches. Exploring modality-specific biomarkers, we identified the brainstem as a sensitive biomarker for both MRI and Fluorodeoxyglucose PET modalities, implicating its involvement in early Alzheimer's pathology. Notably, the corpus callosum and adjacent cortical regions emerged as potential biomarkers, warranting further study into their role in the early stages of Alzheimer's disease.

Implementing ABCD studyⓇ MRI sequences for multi-site cohort studies: Practical guide to necessary steps, preprocessing methods, and challenges

Large multi-site studies that combine magnetic resonance imaging (MRI) data across research sites present exceptional opportunities to advance neuroscience research. However, scanner or site variability and non-standardised image acquisition protocols, data processing and analysis pipelines can adversely affect the reliability and repeatability of MRI derived brain measures. We implemented a standardised MRI protocol based on that used in the Adolescent Brain Cognition Development (ABCD)Ⓡ study in two sites, and across four MRI scanners. Twice repeated measurements of a single healthy volunteer were obtained in two sites and in four 3T MRI scanners (vendors: Siemens, Philips, and GE). Imaging data included anatomical scans (T1 weighted, T2 weighted), diffusion weighted imaging (DWI) and resting state functional MRI (rs-fMRI). Standardised containerized pipelines were utilised to pre-process the data and different image quality metrics and test-retest variability of different brain metrics were evaluated. The implementation of the MRI protocols was possible with minor adjustments in acquisition (e.g. repetition time (TR), higher b-values) and exporting (DICOM formats) of images due to different technical performance of the scanners. This study provides practical insights into the implementation of standardised sequences and data processing for multisite studies, showcase the benefits of containerised preprocessing tools, and highlights the need for careful optimisation of multisite image acquisition.

A multimodal machine learning model for predicting dementia conversion in Alzheimer's disease

Alzheimer's disease (AD) accounts for 60-70% of the population with dementia. Mild cognitive impairment (MCI) is a diagnostic entity defined as an intermediate stage between subjective cognitive decline and dementia, and about 10-15% of people annually convert to AD. We aimed to investigate the most robust model and modality combination by combining multi-modality image features based on demographic characteristics in six machine learning models. A total of 196 subjects were enrolled from four hospitals and the Alzheimer's Disease Neuroimaging Initiative dataset. During the four-year follow-up period, 47 (24%) patients progressed from MCI to AD. Volumes of the regions of interest, white matter hyperintensity, and regional Standardized Uptake Value Ratio (SUVR) were analyzed using T1, T2-weighted-Fluid-Attenuated Inversion Recovery (T2-FLAIR) MRIs, and amyloid PET (αPET), along with automatically provided hippocampal occupancy scores (HOC) and Fazekas scales. As a result of testing the robustness of the model, the GBM model was the most stable, and in modality combination, model performance was further improved in the absence of T2-FLAIR image features. Our study predicts the probability of AD conversion in MCI patients, which is expected to be useful information for clinician's early diagnosis and treatment plan design.

Mahmoud L, Sadek I. AI-based tool for early detection of Alzheimer's disease

In the context of Alzheimer's disease (AD), timely identification is paramount for effective management, acknowledging its chronic and irreversible nature, where medications can only impede its progression. Our study introduces a holistic solution, leveraging the hippocampus and the VGG16 model with transfer learning for early AD detection. The hippocampus, a pivotal early affected region linked to memory, plays a central role in classifying patients into three categories: cognitively normal (CN), representing individuals without cognitive impairment; mild cognitive impairment (MCI), indicative of a subtle decline in cognitive abilities; and AD, denoting Alzheimer's disease. Employing the Alzheimer's Disease Neuroimaging Initiative (ADNI) dataset, our model undergoes training enriched by advanced image preprocessing techniques, achieving outstanding accuracy (testing 98.17 %, validation 97.52 %, training 99.62 %). The strategic use of transfer learning fortifies our competitive edge, incorporating the hippocampus approach and, notably, a progressive data augmentation technique. This innovative augmentation strategy gradually introduces augmentation factors during training, significantly elevating accuracy and enhancing the model's generalization ability. The study emphasizes practical application with a user-friendly website, empowering radiologists to predict class probabilities, track disease progression, and visualize patient images in both 2D and 3D formats, contributing significantly to the advancement of early AD detection.

Multi-response Regression for Block-missing Multi-modal Data without Imputation

Multi-modal data are prevalent in many scientific fields. In this study, we consider the parameter estimation and variable selection for a multi-response regression using block-missing multi-modal data. Our method allows the dimensions of both the responses and the predictors to be large, and the responses to be incomplete and correlated, a common practical problem in high-dimensional settings. Our proposed method uses two steps to make a prediction from a multi-response linear regression model with block-missing multi-modal predictors. In the first step, without imputing missing data, we use all available data to estimate the covariance matrix of the predictors and the cross-covariance matrix between the predictors and the responses. In the second step, we use these matrices and a penalized method to simultaneously estimate the precision matrix of the response vector, given the predictors, and the sparse regression parameter matrix. Lastly, we demonstrate the effectiveness of the proposed method using theoretical studies, simulated examples, and an analysis of a multi-modal imaging data set from the Alzheimer's Disease Neuroimaging Initiative.

Increased Likelihood of Dementia with Coexisting Atrophy of Multiple Regions of Interest

BACKGROUND

Brain volume is associated with cognitive decline in later life, and cortical brain atrophy exceeding the normal range is related to inferior cognitive and behavioral outcomes in later life.

OBJECTIVE

To investigate the likelihood of cognitive decline, mild cognitive impairment (MCI), or dementia, when regional atrophy is present in participants' magnetic resonance imaging (MRI).

METHODS

Multi-center MRI data of 2,545 adults were utilized to measure regional volumes using NEUROPHET AQUA. Four lobes (frontal, parietal, temporal, and occipital), four Alzheimer's disease-related regions (entorhinal, fusiform, inferior temporal, and middle temporal area), and the hippocampus in the left and right hemispheres were measured and analyzed. The presence of regional atrophy from brain MRI was defined as ≤1.5 standard deviation (SD) compared to the age- and sex-matched cognitively normal population. The risk ratio for cognitive decline was investigated for participants with regional atrophy in contrast to those without regional atrophy.

RESULTS

The risk ratio for cognitive decline was significantly higher when hippocampal atrophy was present (MCI, 1.84, p < 0.001; dementia, 4.17, p < 0.001). Additionally, participants with joint atrophy in multiple regions showed a higher risk ratio for dementia, e.g., 9.6 risk ratio (95% confidence interval, 8.0-11.5), with atrophy identified in the frontal, temporal, and hippocampal gray matter, than those without atrophy.

CONCLUSIONS

Our study showed that individuals with multiple regional atrophy (either lobar or AD-specific regions) have a higher likelihood of developing dementia compared to the age- and sex-matched population without atrophy. Thus, further consideration is needed when assessing MRI findings.

Distinct subtypes of spatial brain metabolism patterns in Alzheimer's disease identified by deep learning-based FDG PET clusters

PURPOSE

Alzheimer's disease (AD) is a heterogeneous disease that presents a broad spectrum of clinicopathologic profiles. To date, objective subtyping of AD independent of disease progression using brain imaging has been required. Our study aimed to extract representations of unique brain metabolism patterns different from disease progression to identify objective subtypes of AD.

METHODS

A total of 3620 FDG brain PET images with AD, mild cognitive impairment (MCI), and cognitively normal (CN) were obtained from the ADNI database from 1607 participants at enrollment and follow-up visits. A conditional variational autoencoder model was trained on FDG brain PET images of AD patients with the corresponding condition of AD severity score. The k-means algorithm was applied to generate clusters from the encoded representations. The trained deep learning-based cluster model was also transferred to FDG PET of MCI patients and predicted the prognosis of subtypes for conversion from MCI to AD. Spatial metabolism patterns, clinical and biological characteristics, and conversion rate from MCI to AD were compared across the subtypes.

RESULTS

Four distinct subtypes of spatial metabolism patterns in AD with different brain pathologies and clinical profiles were identified: (i) angular, (ii) occipital, (iii) orbitofrontal, and (iv) minimal hypometabolic patterns. The deep learning model was also successfully transferred for subtyping MCI, and significant differences in frequency (P < 0.001) and risk of conversion (log-rank P < 0.0001) from MCI to AD were observed across the subtypes, highest in S2 (35.7%) followed by S1 (23.4%).

CONCLUSION

We identified distinct subtypes of AD with different clinicopathologic features. The deep learning-based approach to distinguish AD subtypes on FDG PET could have implications for predicting individual outcomes and provide a clue to understanding the heterogeneous pathophysiology of AD.

Predicting Four-Year's Alzheimer's Disease Onset Using Longitudinal Neurocognitive Tests and MRI Data Using Explainable Deep Convolutional Neural Networks

BACKGROUND

Prognosis of future risk of dementia from neuroimaging and cognitive data is important for optimizing clinical management for patients at early stage of Alzheimer's disease (AD). However, existing studies lack an efficient way to integrate longitudinal information from both modalities to improve prognosis performance.

OBJECTIVE

In this study, we aim to develop and evaluate an explainable deep learning-based framework to predict mild cognitive impairment (MCI) to AD conversion within four years using longitudinal whole-brain 3D MRI and neurocognitive tests.

METHODS

We proposed a two-stage framework that first uses a 3D convolutional neural network to extract single-timepoint MRI-based AD-related latent features, followed by multi-modal longitudinal feature concatenation and a 1D convolutional neural network to predict the risk of future dementia onset in four years.

RESULTS

The proposed deep learning framework showed promising to predict MCI to AD conversion within 4 years using longitudinal whole-brain 3D MRI and cognitive data without extracting regional brain volumes or cortical thickness, reaching a balanced accuracy of 0.834, significantly improved from models trained from single timepoint or single modality. The post hoc model explainability revealed heatmap indicating regions that are important for predicting future risk of AD.

CONCLUSIONS

The proposed framework sets the stage for future studies for using multi-modal longitudinal data to achieve optimal prediction for prognosis of AD onset, leading to better management of the diseases, thereby improving the quality of life.

Prediction of Progression from Mild Cognitive Impairment to Alzheimer's disease with Longitudinal and Multimodal Data

Introduction

Accurate prediction of the progression from mild cognitive impairment (MCI) to Alzheimer's disease (AD) within a certain time frame is crucial for appropriate therapeutic interventions. However, it is challenging to capture the dynamic changes in cognitive and functional abilities over time, resulting in limited predictive performance. Our study aimed to investigate whether incorporating longitudinal multimodal data with advanced analytical methods could improve the capability to predict the risk of progressing to AD.

Methods

This study included participants from the Alzheimer's Disease Neuroimaging Initiative (ADNI), a large-scale multi-center longitudinal study. Three data modalities, including demographic variables, neuropsychological tests, and neuroimaging measures were considered. A Long Short-Term Memory (LSTM) model using data collected at five-time points (baseline, 6-month, 12-month, 18-month, and 24-month) was developed to predict the risk of progression from MCI to AD within two years from the index exam (the exam at 24-month). In contrast, a random forest model was developed to predict the risk of progression just based on the data collected at the index exam.

Results

The study included 347 participants with MCI at 24-month (age: mean 75, SD 7 years; 39.8% women) from ADNI, of whom 77 converted to AD over a 2-year follow-up period. The longitudinal LSTM model showed superior prediction performance of MCI-to-AD progression (AUC 0.93±0.06) compared to the random forest model (AUC 0.90±0.09). A similar pattern was also observed across different age groups.

Discussion

Our study suggests that the incorporation of longitudinal data can provide better predictive performance for 2-year MCI-to-AD progression risk than relying solely on cross-sectional data. Therefore, repeated or multiple times routine health surveillance of MCI patients are essential in the early detection and intervention of AD.

Detecting conversion from mild cognitive impairment to Alzheimer's disease using FLAIR MRI biomarkers

Mild cognitive impairment (MCI) is the prodromal phase of Alzheimer's disease (AD) and while it presents as an imperative intervention window, it is difficult to detect which subjects convert to AD (cMCI) and which ones remain stable (sMCI). The objective of this work was to investigate fluid-attenuated inversion recovery (FLAIR) MRI biomarkers and their ability to differentiate between sMCI and cMCI subjects in cross-sectional and longitudinal data. Three types of biomarkers were investigated: volume, intensity and texture. Volume biomarkers included total brain volume, cerebrospinal fluid volume (CSF), lateral ventricular volume, white matter lesion volume, subarachnoid CSF, and grey matter (GM) and white matter (WM), all normalized to intracranial volume. The mean intensity, kurtosis, and skewness of the GM and WM made up the intensity features. Texture features quantified homogeneity and microstructural tissue changes of GM and WM regions. Composite indices were also considered, which are biomarkers that represent an aggregate sum (z-score normalization and summation) of all biomarkers. The FLAIR MRI biomarkers successfully identified high-risk subjects as significant differences (p < 0.05) were found between the means of the sMCI and cMCI groups and the rate of change over time for several individual biomarkers as well as the composite indices for both cross-sectional and longitudinal analyses. Classification accuracy and feature importance analysis showed volume biomarkers to be most predictive, however, best performance was obtained when complimenting the volume biomarkers with the intensity and texture features. Using all the biomarkers, accuracy of 86.2 % and 69.2 % was achieved for normal control-AD and sMCI-cMCI classification respectively. Survival analysis demonstrated that the majority of the biomarkers showed a noticeable impact on the AD conversion probability 4 years prior to conversion. Composite indices were the top performers for all analyses including feature importance, classification, and survival analysis. This demonstrated their ability to summarize various dimensions of disease into single-valued metrics. Significant correlation (p < 0.05) with phosphorylated-tau and amyloid-beta CSF biomarkers was found with all the FLAIR biomarkers. The proposed biomarker system is easily attained as FLAIR is routinely acquired, models are not computationally intensive and the results are explainable, thus making this pipeline easily integrated into clinical workflow.

Evaluation of MRI-based machine learning approaches for computer-aided diagnosis of dementia in a clinical data warehouse

A variety of algorithms have been proposed for computer-aided diagnosis of dementia from anatomical brain MRI. These approaches achieve high accuracy when applied to research data sets but their performance on real-life clinical routine data has not been evaluated yet. The aim of this work was to study the performance of such approaches on clinical routine data, based on a hospital data warehouse, and to compare the results to those obtained on a research data set. The clinical data set was extracted from the hospital data warehouse of the Greater Paris area, which includes 39 different hospitals. The research set was composed of data from the Alzheimer's Disease Neuroimaging Initiative data set. In the clinical set, the population of interest was identified by exploiting the diagnostic codes from the 10th revision of the International Classification of Diseases that are assigned to each patient. We studied how the imbalance of the training sets, in terms of contrast agent injection and image quality, may bias the results. We demonstrated that computer-aided diagnosis performance was strongly biased upwards (over 17 percent points of balanced accuracy) by the confounders of image quality and contrast agent injection, a phenomenon known as the Clever Hans effect or shortcut learning. When these biases were removed, the performance was very poor. In any case, the performance was considerably lower than on the research data set. Our study highlights that there are still considerable challenges for translating dementia computer-aided diagnosis systems to clinical routine.

Enhanced Harris hawks optimization-based fuzzy k-nearest neighbor algorithm for diagnosis of Alzheimer's disease

In order to stop deterioration and give patients with Alzheimer's disease (AD) early therapy, it is crucial to correctly diagnose AD and its early stage, mild cognitive impairment (MCI). A framework for diagnosing AD is presented in this paper, which includes magnetic resonance imaging (MRI) image preprocessing, feature extraction, and the Fuzzy k-nearest neighbor algorithm (FKNN) model. In particular, the framework's novelty lies in the use of an improved Harris Hawks Optimization (HHO) algorithm named SSFSHHO, which integrates the Sobol sequence and Stochastic Fractal Search (SFS) mechanisms for optimizing the parameters of FKNN. The HHO method improves the quality of the initial population overall by incorporating the Sobol sequence, and the SFS mechanism increases the algorithm's capacity to get out of the local optimum solution. Comparisons with other classical meta-heuristic algorithms, state-of-the-art HHO variants in low and high dimensions, and enhanced meta-heuristic algorithms on 30 typical IEEE CEC2014 benchmark test problems show that the overall performance of SSFSHHO is significantly better than other comparative algorithms. Moreover, the created framework based on the SSFSHHO-FKNN model is employed to classify AD and MCI using MRI scans from the ADNI dataset, achieving high classification performance for 6 representative cases. Experimental findings indicate that the proposed algorithm performs better than a number of high-performance optimization algorithms and classical machine learning algorithms, thus offering a promising approach for AD classification. Additionally, the proposed strategy can successfully identify relevant features and enhance classification performance for AD diagnosis.

18F-FDG-PET Radiomics Based on White Matter Predicts The Progression of Mild Cognitive Impairment to Alzheimer Disease: A Machine Learning Study

RATIONALE AND OBJECTIVES

To build a model using white-matter radiomics features on positron-emission tomography (PET) and machine learning methods to predict progression from mild cognitive impairment (MCI) to Alzheimer disease (AD).

MATERIALS AND METHODS

We analyzed the data of 341 MCI patients from the Alzheimer's Disease Neuroimaging Initiative, of whom 102 progressed to AD during an 8-year follow-up. The patients were divided into the training (238 patients) and test groups (103 patients). PET-based radiomics features were extracted from the white matter in the training group, and dimensionally reduced to construct a psychoradiomics signature (PS), which was combined with multimodal data using machine learning methods to construct an integrated model. Model performance was evaluated using receiver operating characteristic curves in the test group.

RESULTS

Clinical Dementia Rating (CDR) scores, Alzheimer's Disease Assessment Scale (ADAS) scores, and PS independently predicted MCI progression to AD on multivariate logistic regression. The areas under the curve (AUCs) of the CDR, ADAS and PS in the training and test groups were 0.683, 0.755, 0.747 and 0.737, 0.743, 0.719 respectively, and were combined using a support vector machine to construct an integrated model. The AUC of the integrated model in the training and test groups was 0.868 and 0.865, respectively (sensitivity, 0.873 and 0.839, respectively; specificity, 0.784 and 0.806, respectively). The AUCs of the integrated model significantly differed from those of other predictors in both groups (p < 0.05, Delong test).

CONCLUSION

Our psych radiomics signature based on white-matter PET data predicted MCI progression to AD. The integrated model built using multimodal data and machine learning identified MCI patients at a high risk of progression to AD.

Accumulation of network redundancy marks the early stage of Alzheimer's disease

Brain wiring redundancy counteracts aging-related cognitive decline by reserving additional communication channels as a neuroprotective mechanism. Such a mechanism plays a potentially important role in maintaining cognitive function during the early stages of neurodegenerative disorders such as Alzheimer's disease (AD). AD is characterized by severe cognitive decline and involves a long prodromal stage of mild cognitive impairment (MCI). Since MCI subjects are at high risk of converting to AD, identifying MCI individuals is essential for early intervention. To delineate the redundancy profile during AD progression and enable better MCI diagnosis, we define a metric that reflects redundant disjoint connections between brain regions and extract redundancy features in three high-order brain networks-medial frontal, frontoparietal, and default mode networks-based on dynamic functional connectivity (dFC) captured by resting-state functional magnetic resonance imaging (rs-fMRI). We show that redundancy increases significantly from normal control (NC) to MCI individuals and decreases slightly from MCI to AD individuals. We further demonstrate that statistical features of redundancy are highly discriminative and yield state-of-the-art accuracy of up to 96.8 ± 1.0% in support vector machine (SVM) classification between NC and MCI individuals. This study provides evidence supporting the notion that redundancy serves as a crucial neuroprotective mechanism in MCI.

Convergent and distinct neural structural and functional patterns of mild cognitive impairment: a multimodal meta-analysis

Mild cognitive impairment (MCI) is regarded as a transitional stage between normal aging and Alzheimer's disease. Numerous voxel-based morphometry (VBM) and resting-state fMRI (rs-fMRI) studies have provided strong evidence of abnormalities in the structure and intrinsic function of brain regions in MCI. Studies have recently begun to explore their association but have not employed systematic information in this pursuit. Herein, a multimodal meta-analysis was performed, which included 43 VBM datasets (1,247 patients and 1,352 controls) of gray matter volume (GMV) and 42 rs-fMRI datasets (1,468 patients and 1,605 controls) that combined 3 metrics: amplitude of low-frequency fluctuation, the fractional amplitude of low-frequency fluctuation, and regional homogeneity. Compared to controls, patients with MCI displayed convergent reduced regional GMV and altered intrinsic activity, mainly in the default mode network and salience network. Decreased GMV alone in ventral medial prefrontal cortex and altered intrinsic function alone in bilateral dorsal anterior cingulate/paracingulate gyri, right lingual gyrus, and cerebellum were identified, respectively. This meta-analysis investigated complex patterns of convergent and distinct brain alterations impacting different neural networks in MCI patients, which contributes to a further understanding of the pathophysiology of MCI.

AANet: Attentive All-level Fusion Deep Neural Network Approach for Multi-modality Early Alzheimer's Disease Diagnosis

Multi-modality deep learning models have recently been used for disease diagnosis; however, effectively integrating diverse, complex, and heterogeneous data remains a challenge. In this study, we propose a novel system, attentive All-level Fusion(AANet), to fuse multi-level and multi-modality patient data, including 3D brain images, patient demographics, genetics, and blood biomarkers into a deep-learning framework for disease diagnosis, and tested it for early Alzheimer's disease diagnosis. We first constructed a deep learning feature pyramid network for whole-brain brain magnetic resonance imaging (MRI) feature extraction. We then leveraged the self-attention-based all-level fusion method by automatically adjusting weights of all-level MRI image features, patient demographics, blood biomarkers, and genetic data. We trained and tested AANet on data from the Alzheimer's Disease Neuroimaging Initiative for the task of classifying mild cognitive impairment from Alzheimer's disease, a challenging task in early Alzheimer's disease diagnosis. AANet achieved an accuracy of 90.5%, outperformed several state-of-the-art methods. In summary, AANet provides an advanced methodological framework for multi-modality-based disease diagnosis.

Baseline structural MRI and plasma biomarkers predict longitudinal structural atrophy and cognitive decline in early Alzheimer's disease

BACKGROUND

Crucial to the success of clinical trials targeting early Alzheimer's disease (AD) is recruiting participants who are more likely to progress over the course of the trials. We hypothesize that a combination of plasma and structural MRI biomarkers, which are less costly and non-invasive, is predictive of longitudinal progression measured by atrophy and cognitive decline in early AD, providing a practical alternative to PET or cerebrospinal fluid biomarkers.

METHODS

Longitudinal T1-weighted MRI, cognitive (memory-related test scores and clinical dementia rating scale), and plasma measurements of 245 cognitively normal (CN) and 361 mild cognitive impairment (MCI) patients from ADNI were included. Subjects were further divided into β-amyloid positive/negative (Aβ+/Aβ-)] subgroups. Baseline plasma (p-tau181 and neurofilament light chain) and MRI-based structural medial temporal lobe subregional measurements and their association with longitudinal measures of atrophy and cognitive decline were tested using stepwise linear mixed effect modeling in CN and MCI, as well as separately in the Aβ+/Aβ- subgroups. Receiver operating characteristic (ROC) analyses were performed to investigate the discriminative power of each model in separating fast and slow progressors (first and last terciles) of each longitudinal measurement.

RESULTS

A total of 245 CN (35.0% Aβ+) and 361 MCI (53.2% Aβ+) participants were included. In the CN and MCI groups, both baseline plasma and structural MRI biomarkers were included in most models. These relationships were maintained when limited to the Aβ+ and Aβ- subgroups, including Aβ- CN (normal aging). ROC analyses demonstrated reliable discriminative power in identifying fast from slow progressors in MCI [area under the curve (AUC): 0.78-0.93] and more modestly in CN (0.65-0.73).

CONCLUSIONS

The present data support the notion that plasma and MRI biomarkers, which are relatively easy to obtain, provide a prediction for the rate of future cognitive and neurodegenerative progression that may be particularly useful in clinical trial stratification and prognosis. Additionally, the effect in Aβ- CN indicates the potential use of these biomarkers in predicting a normal age-related decline.

Predicting Alzheimer's Disease and Quantifying Asymmetric Degeneration of the Hippocampus Using Deep Learning of Magnetic Resonance Imaging Data

In order to quantify lateral asymmetric degeneration of the hippocampus for early predicting Alzheimer's disease (AD), we develop a deep learning (DL) model to learn informative features from the hippocampal magnetic resonance imaging (MRI) data for predicting AD conversion in a time-to-event prediction modeling framework. The DL model is trained on unilateral hippocampal data with an autoencoder based regularizer, facilitating quantification of lateral asymmetry in the hippocampal prediction power of AD conversion and identification of the optimal strategy to integrate the bilateral hippocampal MRI data for predicting AD. Experimental results on MRI scans of 1307 subjects (817 for training and 490 for validation) have demonstrated that the left hippocampus can better predict AD than the right hippocampus, and an integration of the bilateral hippocampal data with the instance based DL method improved AD prediction, compared with alternative predictive modeling strategies.

Characterization of Depressive Symptoms in Dementia and Examination of Possible Risk Factors

Background

Depression in individuals with Alzheimer's disease (AD) is common, distressing, difficult to treat, and inadequately understood. It occurs more frequently in AD than in older adults without dementia. The reasons why some patients develop depression during AD and others do not remain obscure.

Objective

We aimed to characterize depression in AD and to identify risk factors.

Methods

We used data from three large dementia focused cohorts: ADNI (n = 665 with AD, 669 normal cognition), NACC (n = 698 with AD, 711 normal cognition), and BDR (n = 757 with AD). Depression ratings were available using the GDS and NPI and in addition for BDR the Cornell. A cut-off of≥8 was used for the GDS and the Cornell Scale for Depression in Dementia,≥6 for the NPI depression sub-scale, and≥2 for the NPI-Q depression sub-scale. We used logistic regression to examine potential risk factors and random effects meta-analysis and an interaction term to look for interactions between each risk factor and the presence of cognitive impairment.

Results

In individual studies there was no evidence of a difference in risk factors for depressive symptoms in AD. In the meta-analysis the only risk factor which increased the risk of depressive symptoms in AD was previous depression, but information on this was only available from one study (OR 7.78 95% CI 4.03-15.03).

Conclusion

Risk factors for depression in AD appear to differ to those for depression per se supporting suggestions of a different pathological process, although a past history of depression was the strongest individual risk factor.

Two Routes to Alzheimer's Disease Based on Differential Structural Changes in Key Brain Regions

BACKGROUND

Alzheimer's disease (AD) is a neurodegenerative disorder with homogenous disease patterns. Neuropathological changes precede symptoms by up to two decades making neuroimaging biomarkers a prime candidate for early diagnosis, prognosis, and patient stratification.

OBJECTIVE

The goal of the study was to discern intermediate AD stages and their precursors based on neuroanatomical features for stratifying patients on their progression through different stages.

METHODS

Data include grey matter features from 14 brain regions extracted from longitudinal structural MRI and cognitive data obtained from 1,017 healthy controls and AD patients of ADNI. AD progression was modeled with a Hidden Markov Model, whose hidden states signify disease stages derived from the neuroanatomical data. To tie the progression in brain atrophy to a behavioral marker, we analyzed the ADAS-cog sub-scores in the stages.

RESULTS

The optimal model consists of eight states with differentiable neuroanatomical features, forming two routes crossing once at a very early point and merging at the final state. The cortical route is characterized by early and sustained atrophy in cortical regions. The limbic route is characterized by early decrease in limbic regions. Cognitive differences between the two routes are most noticeable in the memory domain with subjects from the limbic route experiencing stronger memory impairments.

CONCLUSION

Our findings corroborate that more than one pattern of grey matter deterioration with several discernable stages can be identified in the progression of AD. These neuroanatomical subtypes are behaviorally meaningful and provide a door into early diagnosis of AD and prognosis of the disease's progression.

Increased Hippocampal-Inferior Temporal Gyrus White Matter Connectivity following Donepezil Treatment in Patients with Early Alzheimer's Disease: A Diffusion Tensor Probabilistic Tractography Study

The incidence of Alzheimer's disease (AD) has been increasing each year, and a defective hippocampus has been primarily associated with an early stage of AD. However, the effect of donepezil treatment on hippocampus-related networks is unknown. Thus, in the current study, we evaluated the hippocampal white matter (WM) connectivity in patients with early-stage AD before and after donepezil treatment using probabilistic tractography, and we further determined the WM integrity and changes in brain volume. Ten patients with early-stage AD (mean age = 72.4 ± 7.9 years; seven females and three males) and nine healthy controls (HC; mean age = 70.7 ± 3.5 years; six females and three males) underwent a magnetic resonance (MR) examination. After performing the first MR examination, the patients received donepezil treatment for 6 months. The brain volumes and diffusion tensor imaging scalars of 11 regions of interest (the superior/middle/inferior frontal gyrus, the superior/middle/inferior temporal gyrus, the amygdala, the caudate nucleus, the hippocampus, the putamen, and the thalamus) were measured using MR imaging and DTI, respectively. Seed-based structural connectivity analyses were focused on the hippocampus. The patients with early AD had a lower hippocampal volume and WM connectivity with the superior frontal gyrus and higher mean diffusivity (MD) and radial diffusivity (RD) in the amygdala than HC (p < 0.05, Bonferroni-corrected). However, brain areas with a higher (or lower) brain volume and WM connectivity were not observed in the HC compared with the patients with early AD. After six months of donepezil treatment, the patients with early AD showed increased hippocampal-inferior temporal gyrus (ITG) WM connectivity (p < 0.05, Bonferroni-corrected).

A review on Alzheimer's disease classification from normal controls and mild cognitive impairment using structural MR images

Alzheimer's disease (AD) is an irreversible neurodegenerative brain disorder that degrades the memory and cognitive ability in elderly people. The main reason for memory loss and reduction in cognitive ability is the structural changes in the brain that occur due to neuronal loss. These structural changes are most conspicuous in the hippocampus, cortex, and grey matter and can be assessed by using neuroimaging techniques viz. Positron Emission Tomography (PET), structural Magnetic Resonance Imaging (MRI) and functional MRI (fMRI), etc. Out of these neuroimaging techniques, structural MRI has evolved as the best technique as it indicates the best soft tissue contrast and high spatial resolution which is important for AD detection. Currently, the focus of researchers is on predicting the conversion of Mild Cognitive Impairment (MCI) into AD. MCI represents the transition state between expected cognitive changes with normal aging and Alzheimer's disease. Not every MCI patient progresses into Alzheimer's disease. MCI can develop into stable MCI (sMCI, patients are called non-converters) or into progressive MCI (pMCI, patients are diagnosed as MCI converters). This paper discusses the prognosis of MCI to AD conversion and presents a review of structural MRI-based studies for AD detection. AD detection framework includes feature extraction, feature selection, and classification process. This paper reviews the studies for AD detection based on different feature extraction methods and machine learning algorithms for classification. The performance of various feature extraction methods has been compared and it has been observed that the wavelet transform-based feature extraction method would give promising results for AD classification. The present study indicates that researchers are successful in classifying AD from Normal Controls (N_rmC) but, it still requires a lot of work to be done for MCI/ N_rmC and MCI/AD, which would help in detecting AD at its early stage.

Application of Machine Learning Techniques in the HELIAD Study Data for the Development of Diagnostic Models in MCI and Dementia

The increase in the population's life expectancy leads to an increase in the incidence of dementia and, therefore, in diseases such as Alzheimer's. Towards this direction, the HELIAD1 study is the first large-scale epidemiological study aimed at assessing epidemiological data on dementia, mild mental decline, and other neuropsychiatric disorders associated with old age. This is a huge study with several computational challenges, most of which can be addressed by machine learning processes. The objectives of this study were to detect patterns in the HELIAD clinical data that classify with high accuracy various levels of cognitive impairment by training ML algorithms and hence apply derived model on future clinical data to predict with the same accuracy the class variable. We propose a machine learning method based on RUSBoost classifier to identify a critical subset of biomarkers that classify accurately between neurological patients with mild cognitive impairment (MCI) or dementia of the Alzheimer's type (DAT) and the cognitively healthy control (CHC) group. In this study we used a highly skewed (imbalanced) dataset with most observations (majority class) belonging to the CHC group. The method proposed predicts accurately the clinical diagnosis label and effectively classifies the neurological patients from the CHC class. In particular, the classification accuracy (actual vs predicted) for the three classes of the clinical diagnosis was 97%, 78%, and 91% for control, MCI, and dementia class, respectively.

Persistent depressive symptoms are associated with frontal regional atrophy in patients with Alzheimer's disease

BACKGROUND

Depression in individuals with Alzheimer's disease (AD) is common, difficult to treat and inadequately understood. Previous studies have identified possible differences in regional brain atrophy in individuals with AD and depression, but the results have been inconsistent and some studies had less robust definitions of depression. We aimed to examine regional brain atrophy in two large dementia focused cohorts.

METHODS

We used data from Alzheimer's disease neuroimaging initiative (ADNI) and the National Alzheimer's Co-ordinating Center (NACC), for those with data from at least one MRI scan. Depression ratings were available using the Geriatric Depression Scale (GDS) and Neuropsychiatric Inventory (NPI). Intermittent depressive symptoms were defined as one episode above threshold (≥8 on GDS, ≥6 on NPI depression subscale and ≥2 on the Neuropsychiatric Inventory version Q depression sub-scale) and persistent as ≥2 episodes. Derived regional volumetric data was available from ADNI and the NACC.

RESULTS

Data was available from 698 individuals with AD in NACC and from 666 individuals in ADNI. We found no evidence of between group differences in regional brain volume at baseline, or of differential atrophy in NACC. In ADNI we found evidence of increased brain atrophy in several frontal brain areas.

LIMITATIONS

Because this study was limited to those with MRI data, the numbers in some analyses were low. MRI parcellation differed between studies making direct comparison difficult. For some individuals only the NPI was used to rate depression.

CONCLUSIONS

We have found mixed evidence of increased regional atrophy in depression in AD, mainly in frontal brain regions. We found no evidence to support a vascular basis for depression in AD.

Asymmetry-enhanced attention network for Alzheimer's diagnosis with structural Magnetic Resonance Imaging

BACKGROUND AND OBJECTIVE

With the aging of the global population becoming severe, Alzheimer's disease (AD) has become one of the world's most common senile diseases. Many studies have suggested that the brain's left-right asymmetry is one of the possible diagnostic landmarks for AD. However, most published approaches to classification problems may not adequately explore the asymmetry between the left and right hemispheres. At the same time, the relationship between asymmetry traits and other classifier features remains understudied.

METHODS

In this paper, we proposed an asymmetry enhanced attention network (ASEAN) for AD diagnosis that effectively combines the anatomical asymmetry characteristics of the brain to enhance the accuracy and stability of classification tasks. First, we proposed a multi-scale asymmetry feature extraction module (MSAF) that can extract the asymmetry features of the brain from various scales. Second, we proposed an asymmetry refinement module (ARM) that considers the dependency between feature maps to suppress the irrelevant regions of the asymmetric feature maps. In addition, a parameter-free attention module was introduced to infer 4D attention weights and improve the network's representation capability.

RESULTS

The proposed method achieved performance improvements on two databases: Alzheimer's Disease Neuroimaging Initiative (ADNI) and Australian Imaging, Biomarkers and Lifestyle (AIBL). For the classification tasks on ADNI, the proposed method achieves 92.1% accuracy, 96.2% sensitivity, and 91.3% specificity on the AD vs. CN (Cognitively Normal) task. Compared with state-of-the-art methods, the proposed method could achieve comparable results.

CONCLUSION

The proposed model can extract long-range left-right brain similarity as complementary information and improve the model's diagnostic performance. A large number of experiments also support the model's validity. At the same time, this work provides a valuable reference for other neurological diseases, particularly those that exhibit left-right brain asymmetry during development.

Early dementia diagnosis, MCI-to-dementia risk prediction, and the role of machine learning methods for feature extraction from integrated biomarkers, in particular for EEG signal analysis

INTRODUCTION

Dementia in its various forms represents one of the most frightening emergencies for the aging population. Cognitive decline-including Alzheimer's disease (AD) dementia-does not develop in few days; disease mechanisms act progressively for several years before clinical evidence.

METHODS

A preclinical stage, characterized by measurable cognitive impairment, but not overt dementia, is represented by mild cognitive impairment (MCI), which progresses to-or, more accurately, is already in a prodromal form of-AD in about half cases; people with MCI are therefore considered the population at risk for AD deserving special attention for validating screening methods.

RESULTS

Graph analysis tools, combined with machine learning methods, represent an interesting probe to identify the distinctive features of physiological/pathological brain aging focusing on functional connectivity networks evaluated on electroencephalographic data and neuropsychological/imaging/genetic/metabolic/cerebrospinal fluid/blood biomarkers.

DISCUSSION

On clinical data, this innovative approach for early diagnosis might provide more insight into pathophysiological processes underlying degenerative changes, as well as toward a personalized risk evaluation for pharmacological, nonpharmacological, and rehabilitation treatments.

Multi-label classification of Alzheimer's disease stages from resting-state fMRI-based correlation connectivity data and deep learning

Alzheimer's disease is a neurodegenerative condition that gradually impairs cognitive abilities. Recently, various neuroimaging modalities and machine learning methods have surfaced to diagnose Alzheimer's disease. Resting-state fMRI is a neuroimaging modality that has been widely utilized to study brain activity related to neurodegenerative diseases. In literature, the previous studies are limited to the binary classification of Alzheimer's disease and Mild Cognitive Impairment. The application of computer-aided diagnosis for the numerous advancing phases of Alzheimer's disease, on the other hand, remains understudied. This research analyzes and presents methods for multi-label classification of six Alzheimer's stages using rs-fMRI and deep learning. The proposed model solves the multi-class classification problem by extracting the brain's functional connectivity networks from rs-fMRI data and employing two deep learning approaches, Stacked Sparse Autoencoder and Brain Connectivity Graph Convolutional Network. The suggested models' results were assessed using the k-fold cross-validation approach, and an average accuracy of 77.13% and 84.03% was reached for multi-label classification using Stacked Sparse Autoencoders and Brain Connectivity Based Convolutional Network, respectively. An analysis of brain regions was also performed by using the network's learned weights, leading to the conclusion that the precentral gyrus, frontal gyrus, lingual gyrus, and supplementary motor area are the significant brain regions of interest.

Regularized Buckley-James method for right-censored outcomes with block-missing multimodal covariates

High-dimensional data with censored outcomes of interest are prevalent in medical research. To analyze such data, the regularized Buckley-James estimator has been successfully applied to build accurate predictive models and conduct variable selection. In this paper, we consider the problem of parameter estimation and variable selection for the semiparametric accelerated failure time model for high-dimensional block-missing multimodal neuroimaging data with censored outcomes. We propose a penalized Buckley-James method that can simultaneously handle block-wise missing covariates and censored outcomes. This method can also perform variable selection. The proposed method is evaluated by simulations and applied to a multimodal neuroimaging dataset and obtains meaningful results.

Transfer learning-based modified inception model for the diagnosis of Alzheimer's disease

Alzheimer's disease (AD) is a neurodegenerative ailment, which gradually deteriorates memory and weakens the cognitive functions and capacities of the body, such as recall and logic. To diagnose this disease, CT, MRI, PET, etc. are used. However, these methods are time-consuming and sometimes yield inaccurate results. Thus, deep learning models are utilized, which are less time-consuming and yield results with better accuracy, and could be used with ease. This article proposes a transfer learning-based modified inception model with pre-processing methods of normalization and data addition. The proposed model achieved an accuracy of 94.92 and a sensitivity of 94.94. It is concluded from the results that the proposed model performs better than other state-of-the-art models. For training purposes, a Kaggle dataset was used comprising 6,200 images, with 896 mild demented (M.D) images, 64 moderate demented (Mod.D) images, and 3,200 non-demented (N.D) images, and 1,966 veritably mild demented (V.M.D) images. These models could be employed for developing clinically useful results that are suitable to descry announcements in MRI images.

A Causal Analysis of the Effect of Age and Sex Differences on Brain Atrophy in the Elderly Brain

We studied how brain volume loss in old age is affected by age, the APOE gene, sex, and the level of education completed. The quantitative characterization of brain volume loss at an old age relative to a young age requires-at least in principle-two MRI scans, one performed at a young age and one at an old age. There is, however, a way to address this problem when having only one MRI scan obtained at an old age. We computed the total brain losses of elderly subjects as a ratio between the estimated brain volume and the estimated total intracranial volume. Magnetic resonance imaging (MRI) scans of 890 healthy subjects aged 70 to 85 years were assessed. A causal analysis of factors affecting brain atrophy was performed using probabilistic Bayesian modelling and the mathematics of causal inference. We found that both age and sex were causally related to brain atrophy, with women reaching an elderly age with a 1% larger brain volume relative to their intracranial volume than men. How the brain ages and the rationale for sex differences in brain volume losses during the adult lifespan are questions that need to be addressed with causal inference and empirical data. The graphical causal modelling presented here can be instrumental in understanding a puzzling scientific area of study-the biological aging of the brain.

Relationship between finger movement characteristics and brain voxel-based morphometry

BACKGROUND

Aging is the most significant risk factor for dementia. Alzheimer's disease (AD) accounts for approximately 60-80% of all dementia cases in older adults. This study aimed to examine the relationship between finger movements and brain volume in AD patients using a voxel-based reginal analysis system for Alzheimer's disease (VSRAD) software.

METHODS

Patients diagnosed with AD at the Center for Comprehensive Care and Research on Memory Disorders were included. The diagnostic criteria were based on the National Institute on Aging-Alzheimer's Association. A finger-tapping device was used for all measurements. Participants performed the tasks in the following order: with their non-dominant hand, dominant hand, both hands simultaneously, and alternate hands. Movements were measured for 15 s each. The relationship between distance and output was measured. Magnetic resonance imaging measurements were performed, and VSRAD was conducted using sagittal section 3D T1-weighted images. The Z-score was used to calculate the severity of medial temporal lobe atrophy. Pearson's product-moment correlation coefficient analyzed the relationship between the severity of medial temporal lobe atrophy and mean values of the parameters in the finger-tapping movements. The statistical significance level was set at <5%. The calculated p-values were corrected using the Bonferroni method.

RESULTS

Sixty-two patients were included in the study. Comparison between VSRAD and MoCA-J scores corrected for p-values showed a significant negative correlation with the extent of gray matter atrophy (r = -0. 52; p< 0.001). A positive correlation was observed between the severity of medial temporal lobe atrophy and standard deviation (SD) of the distance rate of velocity peak in extending movements in the non-dominant hand (r = 0. 51; p< 0.001).

CONCLUSIONS

The SD of distance rate of velocity peak in extending movements extracted from finger taps may be a useful parameter for the early detection of AD and diagnosis of its severity.

Brain atrophy pattern in patients with mild cognitive impairment: MRI study

We evaluated the accuracy of the quantitative and semiquantitative analysis in detecting regional atrophy patterns and differentiating mild cognitive impairment patients who remain stable (aMCI-S) from patients who develop Alzheimer’s disease (aMCI-AD) at clinical follow-up. Baseline magnetic resonance imaging was used for quantitative and semiquantitative analysis using visual rating scales. Visual rating scores were related to gray matter thicknesses or volume measures of some structures belonging to the same brain regions. Receiver operating characteristic (ROC) analysis was performed to assess measures’ accuracy in differentiating aMCI-S from aMCI-AD. Comparing aMCI-S and aMCI-AD patients, significant differences were found for specific rating scales, for cortical thickness belonging to the middle temporal lobe (MTL), anterior temporal (AT), and fronto-insular (FI) regions, for gray matter volumes belonging to MTL and AT regions. ROC curve analysis showed that middle temporal atrophy, AT, and FI visual scales showed better diagnostic accuracy than quantitative measures also when thickness measures were combined with hippocampal volumes. Semiquantitative evaluation, performed by trained observers, is a fast and reliable tool in differentiating, at the early stage of disease, aMCI patients that remain stable from those patients that may progress to AD since visual rating scales may be informative both about early hippocampal volume loss and cortical thickness reduction.

Stage-dependent differential influence of metabolic and structural networks on memory across Alzheimer's disease continuum

Background

Large-scale neuronal network breakdown underlies memory impairment in Alzheimer's disease (AD). However, the differential trajectories of the relationships between network organisation and memory across pathology and cognitive stages in AD remain elusive. We determined whether and how the influences of individual-level structural and metabolic covariance network integrity on memory varied with amyloid pathology across clinical stages without assuming a constant relationship.

Methods

Seven hundred and eight participants from the Alzheimer's Disease Neuroimaging Initiative were studied. Individual-level structural and metabolic covariance scores in higher-level cognitive and hippocampal networks were derived from magnetic resonance imaging and [18F] fluorodeoxyglucose positron emission tomography using seed-based partial least square analyses. The non-linear associations between network scores and memory across cognitive stages in each pathology group were examined using sparse varying coefficient modelling.

Results

We showed that the associations of memory with structural and metabolic networks in the hippocampal and default mode regions exhibited pathology-dependent differential trajectories across cognitive stages using sparse varying coefficient modelling. In amyloid pathology group, there was an early influence of hippocampal structural network deterioration on memory impairment in the preclinical stage, and a biphasic influence of the angular gyrus-seeded default mode metabolic network on memory in both preclinical and dementia stages. In non-amyloid pathology groups, in contrast, the trajectory of the hippocampus-memory association was opposite and weaker overall, while no metabolism covariance networks were related to memory. Key findings were replicated in a larger cohort of 1280 participants.

Conclusions

Our findings highlight potential windows of early intervention targeting network breakdown at the preclinical AD stage.

Funding

Data collection and sharing for this project was funded by the Alzheimer's Disease Neuroimaging Initiative (ADNI) (National Institutes of Health Grant U01 AG024904) and DOD ADNI (Department of Defense award number W81XWH-12-2-0012). We also acknowledge the funding support from the Duke NUS/Khoo Bridge Funding Award (KBrFA/2019-0020) and NMRC Open Fund Large Collaborative Grant (OFLCG09May0035).

Theta Band-Power Shapes Amyloid-Driven Longitudinal EEG Changes in Elderly Subjective Memory Complainers At-Risk for Alzheimer's Disease

BACKGROUND

Alzheimer's disease (AD) includes progressive symptoms spread along a continuum of preclinical and clinical stages. Although numerous studies uncovered the neuro-cognitive changes of AD, very little is known on the natural history of brain lesions and modifications of brain networks in elderly cognitively-healthy memory complainers at risk of AD for carrying pathophysiological biomarkers (amyloidopathy and tauopathy).

OBJECTIVE

We analyzed resting-state electroencephalography (EEG) of 318 cognitively-healthy subjective memory complainers from the INSIGHT-preAD cohort at the time of their first visit (M0) and two-years later (M24).

METHODS

Using 18F-florbetapir PET-scanner, subjects were stratified between amyloid negative (A-; n = 230) and positive (A+; n = 88) groups. Differences between A+ and A-were estimated at source-level in each band-power of the EEG spectrum.

RESULTS

At M0, we found an increase of theta power in the mid-frontal cortex in A+ compared to A-. No significant association was found between mid-frontal theta and the individuals' cognitive performance. At M24, theta power increased in A+ relative to A-individuals in the posterior cingulate cortex and the pre-cuneus. Alpha band revealed a peculiar decremental trend in posterior brain regions in the A+ relative to the A-group only at M24. Theta power increase over the mid-frontal and mid-posterior cortices suggests an hypoactivation of the default-mode network in the A+ individuals and a non-linear longitudinal progression at M24.

CONCLUSION

We provide the first source-level longitudinal evidence on the impact of brain amyloidosis on the EEG dynamics of a large-scale, monocentric cohort of elderly individuals at-risk for AD.

Deep Spatio-Temporal Attention-based Recurrent Network from Dynamic Adaptive Functional Connectivity for MCI Identification

Most existing methods of constructing dynamic functional connectivity (dFC) network obtain the connectivity strength via the sliding window correlation (SWC) method, which estimates the connectivity strength at each time segment, rather than at each time point, and thus is difficult to produce accurate dFC network due to the influence of the window type and window width. Furthermore, the deep learning methods may not capture the discriminative spatio-temporal information that is closely related to disease, thus impacting the performance of (mild cognitive impairment) MCI identification. In this paper, a novel spatio-temporal attention-based bidirectional gated recurrent unit (STA-BiGRU) network is proposed to extract inherent spatio-temporal information from a dynamic adaptive functional connectivity (dAFC) network for MCI diagnosis. Specifically, we adopt a group lasso-based Kalman filter algorithm to obtain the dAFC network with more accurate connectivity strength at each time step. Then a spatial attention module with self-attention and a temporal attention module with multiple temporal attention vectors are incorporated into the BiGRU network to extract more discriminative disease-related spatio-temporal information. Finally, the spatio-temporal regularizations are employed to better guide the attention learning of STA-BiGRU network to enhance the robustness of the deep network. Experimental results show that the proposed framework achieves mean accuracies of 90.2%, 90.0%, and 81.5%, respectively, for three MCI classification tasks. This study provides a more effective deep spatio-temporal attention-based recurrent network and obtains good performance and interpretability of deep learning for psychiatry diagnosis research.

Validation of the Alzheimer's disease-resemblance atrophy index in classifying and predicting progression in Alzheimer's disease

Background

Automated tools for characterising dementia risk have the potential to aid in the diagnosis, prognosis, and treatment of Alzheimer's disease (AD). Here, we examined a novel machine learning-based brain atrophy marker, the AD-resemblance atrophy index (AD-RAI), to assess its test-retest reliability and further validate its use in disease classification and prediction.

Methods

Age- and sex-matched 44 probable AD (Age: 69.13 ± 7.13; MMSE: 27-30) and 22 non-demented control (Age: 69.38 ± 7.21; MMSE: 27-30) participants were obtained from the Minimal Interval Resonance Imaging in Alzheimer's Disease (MIRIAD) dataset. Serial T1-weighted images (n = 678) from up to nine time points over a 2-year period, including 179 pairs of back-to-back scans acquired on same participants on the same day and 40 pairs of scans acquired at 2-week intervals were included. All images were automatically processed with AccuBrain® to calculate the AD-RAI. Its same-day repeatability and 2-week reproducibility were first assessed. The discriminative performance of AD-RAI was evaluated using the receiver operating characteristic curve, where DeLong's test was used to evaluate its performance against quantitative medial temporal lobe atrophy (QMTA) and hippocampal volume adjusted by intracranial volume (ICV)-proportions and ICV-residuals methods, respectively (HVR and HRV). Linear mixed-effects modelling was used to investigate longitudinal trajectories of AD-RAI and baseline AD-RAI prediction of cognitive decline. Finally, the longitudinal associations between AD-RAI and MMSE scores were assessed.

Results

AD-RAI had excellent same-day repeatability and excellent 2-week reproducibility. AD-RAI's AUC (99.8%; 95%CI = [99.3%, 100%]) was equivalent to that of QMTA (96.8%; 95%CI = [92.9%, 100%]), and better than that of HVR (86.8%; 95%CI = [78.2%, 95.4%]) or HRV (90.3%; 95%CI = [83.0%, 97.6%]). While baseline AD-RAI was significantly higher in the AD group, it did not show detectable changes over 2 years. Baseline AD-RAI was negatively associated with MMSE scores and the rate of the change in MMSE scores over time. A negative longitudinal association was also found between AD-RAI values and the MMSE scores among AD patients.

Conclusions

The AD-RAI represents a potential biomarker that may support AD diagnosis and be used to predict the rate of future cognitive decline in AD patients.

Early diagnosis of Alzheimer's disease using machine learning: a multi-diagnostic, generalizable approach

BACKGROUND

Early and accurate diagnosis of Alzheimer's disease (AD) is essential for disease management and therapeutic choices that can delay disease progression. Machine learning (ML) approaches have been extensively used in attempts to develop algorithms for reliable early diagnosis of AD, although clinical usefulness, interpretability, and generalizability of the classifiers across datasets and MRI protocols remain limited.

METHODS

We report a multi-diagnostic and generalizable approach for mild cognitive impairment (MCI) and AD diagnosis using structural MRI and ML. Classifiers were trained and tested using subjects from the AD Neuroimaging Initiative (ADNI) database (n = 570) and the Open Access Series of Imaging Studies (OASIS) project database (n = 531). Several classifiers are compared and combined using voting for a decision. Additionally, we report tests of generalizability across datasets and protocols (IR-SPGR and MPRAGE), the impact of using graph theory measures on diagnostic classification performance, the relative importance of different brain regions on classification for better interpretability, and an evaluation of the potential for clinical applicability of the classifier.

RESULTS

Our "healthy controls (HC) vs. AD" classifier trained and tested on the combination of ADNI and OASIS datasets obtained a balanced accuracy (BAC) of 90.6% and a Matthew's correlation coefficient (MCC) of 0.811. Our "HC vs. MCI vs. AD" classifier trained and tested on the ADNI dataset obtained a 62.1% BAC (33.3% being the by-chance cut-off) and 0.438 MCC. Hippocampal features were the strongest contributors to the classification decisions (approx. 25-45%), followed by temporal (approx. 13%), cingulate, and frontal regions (approx. 8-13% each), which is consistent with our current understanding of AD and its progression. Classifiers generalized well across both datasets and protocols. Finally, using graph theory measures did not improve classification performance.

CONCLUSIONS

In sum, we present a diagnostic tool for MCI and AD trained using baseline scans and a follow-up diagnosis regardless of progression, which is multi-diagnostic, generalizable across independent data sources and acquisition protocols, and with transparently reported performance. Rated as potentially clinically applicable, our tool may be clinically useful to inform diagnostic decisions in dementia, if successful in real-world prospective clinical trials.