White matter hyperintensity (WMH) is a primary manifestation of small vessel disease (SVD), leading to vascular cognitive impairment and other disorders. Accurate WMH quantification is vital for diagnosis and prognosis, but current automatic segmentation methods often fall short, especially across different datasets. The aims of this study are to develop and validate a robust deep learning segmentation method for WMH of vascular-origin. In this study, we developed a transformer-based method for the automatic segmentation of vascular-origin WMH using both 3D T1 and 3D T2-FLAIR images. Our initial dataset comprised 126 participants with varying WMH burdens due to SVD, each with manually segmented WMH masks used for training and testing. External validation was performed on two independent datasets: the WMH Segmentation Challenge 2017 dataset (170 subjects) and an in-house vascular risk factor dataset (70 subjects), which included scans acquired on eight different MRI systems at field strengths of 1.5T, 3T, and 5T This approach enabled a comprehensive assessment of the method's generalizability across diverse imaging conditions. We further compared our method against LGA, LPA, BIANCA, UBO-detector and TrUE-Net in optimized settings. Our method consistently outperformed others, achieving a median Dice coefficient of 0.78 ± 0.09 in our primary dataset, 0.72 ± 0.15 in the external dataset 1, and 0.72 ± 0.14 in the external dataset 2. The relative volume errors were 0.15 ± 0.14, 0.50 ± 0.86, and 0.47 ± 1.02, respectively. The true positive rates were 0.81 ± 0.13, 0.92 ± 0.09, and 0.92 ± 0.12, while the false positive rates were 0.20 ± 0.09, 0.40 ± 0.18, and 0.40 ± 0.19. None of the external validation datasets were used for model training; instead, they comprise previously unseen MRI scans acquired from different scanners and protocols. This setup closely reflects real-world clinical scenarios and further demonstrates the robustness and generalizability of our model across diverse MRI systems and acquisition settings. As such, the proposed method provides a reliable solution for WMH segmentation in large-scale cohort studies.

Neuroimaging correlates of remote traumatic brain injury (TBI) are not well understood. Our objective was to examine associations of TBI with brain MRI markers of degeneration and vascular disease.

We performed a cross-sectional analysis using data from a subset of participants who underwent a 3T brain MRI during the fifth Atherosclerosis Risk in Communities Study visit in 2011-2013. Prior TBI and age at first TBI (<18 years, 18-65 years, >65 years) was defined using self-report and International Classification of Diseases, Ninth Edition code data. We examined the following brain MRI metrics: presence of infarcts and microhemorrhages, white matter hyperintensity (WMH) volume, and the distribution of the number of regions of interest (ROIs) below a z-score cut-point of -1.5 for volumetrics, cortical thickness, and fractional anisotropy (FA) and above +1.5 for mean diffusivity (MD).

A total of 1,642 participants were included (mean age 76.8 ± 5.32 years, 61.0% female, 28.3% self-reported Black race, and 25.5% with a history of TBI [median time between first TBI and MRI: 38.2 years]). There was no evidence of differences in vascular imaging findings by overall TBI status, but individuals who sustained their first TBI at age <18 years had higher WMH volume (adjusted β = 0.22 mm3, 95% CI 0.00-0.43) and individuals who sustained their first TBI at age >65 years were more likely to have subcortical microhemorrhages (adjusted OR 1.69, 95% CI 1.03-2.75) compared with individuals without TBI. Compared with individuals without TBI, individuals with a history of TBI had a greater number of ROIs beyond the z-score cut-point for all metrics (smaller volumes, lower cortical thickness, lower FA, and higher MD). These findings were consistent among participants with first TBI sustained at age >65 years old, whereas participants with first TBI sustained at age <18 years old had a greater number of regions beyond the z-score cut-point only for FA and MD.

In this community-dwelling cohort of older adults, TBI was associated with smaller brain volumes, lower cortical thickness, lower FA, and higher MD. Further work is needed in the chronic postinjury period to elucidate the mechanisms underlying the observed structural changes after TBI.

Tumour necrosis factor-α (TNF-α) is a proinflammatory cytokine implicated in the regulation of innate and adaptive immunity. Two receptors exist for TNF-α: TNF receptors 1 (TNFR-1) and 2 (TNFR-2). TNFR-1 and TNFR-2 have been reported to be involved in pleiotropic functions. Multiple lines of evidence implicate TNF-α and its receptors as potential risk factors for Alzheimer's disease (AD). Studies are warranted to assess the association of TNF-α, TNFR-1, and TNFR-2 with AD pathogenesis and whether they can serve as prognostic biomarkers indicative of AD.

In the present study, baseline levels of cerebrospinal fluid (CSF) TNF-α, TNFR-1, and TNFR-2 were explored, and their potential as biomarkers to differentiate between individuals who remain stable and those who experience disease progression over 10 years in the Alzheimer's Disease Neuroimaging Initiative (ADNI) was assessed. The study also examined the correlation between baseline CSF proteins with established AD biomarkers, neuroimaging measures, and cognition.

Whilst the present study shows associations between baseline CSF levels of TNFs with AD biomarkers, the nature of the relationship is ambiguous.

The present study concludes that CSF TNFs do not serve as reliable or robust disease biomarkers of AD.

The overall goal of the Alzheimer's Disease Neuroimaging Initiative (ADNI) is to optimize and validate biomarkers for clinical trials while sharing all data and biofluid samples with the global scientific community. ADNI has been instrumental in standardizing and validating amyloid beta (Aβ) and tau positron emission tomography (PET) imaging. ADNI data were used for the US Food and Drug Administration (FDA) approval of the Fujirebio and Roche Elecsys cerebrospinal fluid diagnostic tests. Additionally, ADNI provided data for the trials of the FDA-approved treatments aducanumab, lecanemab, and donanemab. More than 6000 scientific papers have been published using ADNI data, reflecting ADNI's promotion of open science and data sharing. Despite its enormous success, ADNI has some limitations, particularly in generalizing its data and findings to the entire US/Canadian population. This introduction provides a historical overview of ADNI and highlights its significant accomplishments and future vision to pioneer "the clinical trial of the future" focusing on demographic inclusivity. HIGHLIGHTS: The Alzheimer's Disease Neuroimaging Initiative (ADNI) introduced a novel model for public-private partnerships and data sharing. It successfully validated amyloid and Tau PET imaging, as well as CSF and plasma biomarkers, for diagnosing Alzheimer's disease. ADNI generated and disseminated vital data for designing AD clinical trials.

Pooling MRI data from multiple datasets requires harmonization to reduce undesired inter-site variabilities, while preserving effects of biological variables (or covariates). The popular harmonization approach ComBat uses a mixed effect regression framework that explicitly accounts for covariate distribution differences across datasets. There is also significant interest in developing harmonization approaches based on deep neural networks (DNNs), such as conditional variational autoencoder (cVAE). However, current DNN approaches do not explicitly account for covariate distribution differences across datasets. Here, we provide mathematical results, suggesting that not accounting for covariates can lead to suboptimal harmonization. We propose two DNN-based covariate-aware harmonization approaches: covariate VAE (coVAE) and DeepResBat. The coVAE approach is a natural extension of cVAE by concatenating covariates and site information with site- and covariate-invariant latent representations. DeepResBat adopts a residual framework inspired by ComBat. DeepResBat first removes the effects of covariates with nonlinear regression trees, followed by eliminating site differences with cVAE. Finally, covariate effects are added back to the harmonized residuals. Using three datasets from three continents with a total of 2787 participants and 10,085 anatomical T1 scans, we find that DeepResBat and coVAE outperformed ComBat, CovBat and cVAE in terms of removing dataset differences, while enhancing biological effects of interest. However, coVAE hallucinates spurious associations between anatomical MRI and covariates even when no association exists. Future studies proposing DNN-based harmonization approaches should be aware of this false positive pitfall. Overall, our results suggest that DeepResBat is an effective deep learning alternative to ComBat. Code for DeepResBat can be found here: https://github.com/ThomasYeoLab/CBIG/tree/master/stable_projects/harmonization/An2024_DeepResBat.

Alzheimer's disease (AD) is a neurodegenerative disorder that results in progressive cognitive decline but without any clinically validated cures so far. Understanding the progression of AD is critical for early detection and risk assessment for AD in aging individuals, thereby enabling initiation of timely intervention and improved chance of success in AD trials. Recent pseudotime approach turns cross-sectional data into "faux" longitudinal data to understand how a complex process evolves over time. This is critical for Alzheimer, which unfolds over the course of decades, but the collected data offers only a snapshot. In this study, we tested several state-of-the-art pseudotime approaches to model the full spectrum of AD progression. Subsequently, we evaluated and compared the pseudotime progression score derived from individual imaging modalities and multi-modalities in the ADNI cohort. Our results showed that most existing pseudotime analysis tools do not generalize well to the imaging data, with either flipped progression score or poor separation of diagnosis groups. This is likely due to the underlying assumptions that only stand for single cell data. From the only tool with promising results, it was observed that all pseudotime, derived from either single imaging modalities or multi-modalities, captures the progressiveness of diagnosis groups. Pseudotime from multi-modality, but not the single modalities, confirmed the hypothetical temporal order of imaging phenotypes. In addition, we found that multi-modal pseudotime is mostly driven by amyloid and tau imaging, suggesting their continuous changes along the full spectrum of AD progression.

Brain age is a powerful marker of general brain health. Furthermore, brain age models are trained on large datasets, thus giving them a potential advantage in predicting specific outcomes - much like the success of finetuning large language models for specific applications. However, it is also well-accepted in machine learning that models trained to directly predict specific outcomes (i.e., direct models) often perform better than those trained on surrogate outcomes. Therefore, despite their much larger training data, it is unclear whether brain age models outperform direct models in predicting specific brain health outcomes. Here, we compare large-scale brain age models and direct models for predicting specific health outcomes in the context of Alzheimer's Disease (AD) dementia. Using anatomical T1 scans from three continents (N = 1,848), we find that direct models outperform brain age models without finetuning. Finetuned brain age models yielded similar performance as direct models, but importantly, did not outperform direct models although the brain age models were pretrained on 1000 times more data than the direct models: N = 53,542 vs N = 50. Overall, our results do not discount brain age as a useful marker of general brain health. However, in this era of large-scale brain age models, our results suggest that small-scale, targeted approaches for extracting specific brain health markers still hold significant value.

Alzheimer's disease (AD) initiates years prior to symptoms, underscoring the importance of early detection. While amyloid accumulation starts early, individuals with substantial amyloid burden may remain cognitively normal, implying that amyloid alone is not sufficient for early risk assessment.

Given the genetic susceptibility of AD, a multi-factorial pseudotime approach was proposed to integrate amyloid imaging and genotype data for estimating a risk score. Validation involved association with cognitive decline and survival analysis across risk-stratified groups, focusing on patients with mild cognitive impairment (MCI).

Our risk score outperformed amyloid composite standardized uptake value ratio in correlation with cognitive scores. MCI subjects with lower pseudotime risk score showed substantial delayed onset of AD and slower cognitive decline. Moreover, pseudotime risk score demonstrated strong capability in risk stratification within traditionally defined subgroups such as early MCI, apolipoprotein E (APOE) ε4+ MCI, APOE ε4- MCI, and amyloid+ MCI.

Our risk score holds great potential to improve the precision of early risk assessment.

Accurate early risk assessment is critical for the success of clinical trials. A new risk score was built from integrating amyloid imaging and genetic data. Our risk score demonstrated improved capability in early risk stratification.

The grey matter of the brain develops and declines in coordinated patterns during the lifespan. Such covariation patterns of grey matter structure can be quantified as grey matter networks, which can be measured with magnetic resonance imaging. In Alzheimer's disease, the global organization of grey matter networks becomes more random, which is captured by a decline in the small-world coefficient. Such decline in the small-world value has been robustly associated with cognitive decline across clinical stages of Alzheimer's disease. The biological mechanisms causing this decline in small-world values remain unknown. Cerebrospinal fluid (CSF) protein biomarkers are available for studying diverse pathological mechanisms in humans and can provide insight into decline. We investigated the relationships between 10 CSF proteins and small-world coefficient in mutation carriers (N = 219) and non-carriers (N = 136) of the Dominantly Inherited Alzheimer Network Observational study. Abnormalities in Amyloid beta, Tau, synaptic (Synaptosome associated protein-25, Neurogranin) and neuronal calcium-sensor protein (Visinin-like protein-1) preceded loss of small-world coefficient by several years, while increased levels in CSF markers for inflammation (Chitinase-3-like protein 1) and axonal injury (Neurofilament light) co-occurred with decreasing small-world values. This suggests that axonal loss and inflammation play a role in structural grey matter network changes.

Cholinesterase inhibitors (ChEIs) are prescribed for Alzheimer's disease (AD) and sometimes for mild cognitive impairment (MCI) without knowing underlying pathologies and its effect on cognition. We investigated the frequency of ChEI prescriptions in amyloid-negative MCI and their association with cognitive changes in the Alzheimer's Disease Neuroimaging Initiative (ADNI) cohort.

We included participants with amyloid positron emission tomography (PET)-negative MCI from the ADNI. We analyzed the associations of ChEI use with cognitive changes, brain volume, and cerebrospinal fluid (CSF) total tau (t-tau), hyperphosphorylated tau181 (p-tau181), and p-tau181/t-tau ratio.

ChEIs were prescribed in 27.4% of amyloid PET-negative MCI and were associated with faster cognitive decline, reduced baseline hippocampal volume and entorhinal cortical thickness, and a longitudinal decrease in the frontal lobe cortical thickness.

The association between ChEI use and accelerated cognitive decline may stem from underlying pathologies involving reduced hippocampal volume, entorhinal cortical thickness and faster frontal lobe atrophy. We suggest that ChEI use in amyloid PET-negative MCI patients might need further consideration, and studies investigating the causality between ChEI use and cognitive decline are warranted in the future.

Associations between magnetic resonance imaging markers of cerebral small vessel disease (CSVD) and dementia risk in older adults have been established, but it remains unclear how lifestyle factors, including psychosocial health, may modify this association.

Social support and social isolation were assessed among participants of the community-based ARIC (Atherosclerosis Risk in Communities) Study, via self-reported questionnaires (1990-1992). Following categorization of both factors, participants were classified as having strong or poor mid-life social relationships. At visit 5 (2011-2013), participants underwent 3T brain magnetic resonance imaging quantifying CSVD measures: white matter hyperintensity volume, microbleeds (subcortical), infarcts (lacunar), and white matter integrity (diffusion tensor imaging). Incident dementia cases were identified from the time of imaging through December 31, 2020 with ongoing surveillance. Associations between CSVD magnetic resonance imaging markers and incident dementia were evaluated using Cox proportional-hazard regressions adjusted for demographic and additional risk factors (from visit 2). Effect modification by mid-life social relationships was evaluated.

Of the 1977 participants with magnetic resonance imaging, 1617 participants (60.7% women; 26.5% Black participants; mean age at visit 2, 55.4 years) were examined. In this sample, mid-life social relationships significantly modified the association between white matter hyperintensity volume and dementia risk (P interaction=0.001). Greater white matter hyperintensity volume was significantly associated with risk of dementia in all participants, yet, more substantially in those with poor (hazard ratio, 1.84 [95% CI, 1.49-2.27]) versus strong (hazard ratio, 1.26 [95% CI, 1.08-1.47]) mid-life social relationships. Although not statistically significant, subcortical microbleeds in participants with poor mid-life social relationships were associated with a greater risk of dementia, relative to those with strong social relationships, in whom subcortical microbleeds were no longer associated with elevated dementia risk.

The elevated risk of dementia associated with CSVD may be reduced in participants with strong mid-life social relationships. Future studies evaluating psychosocial health through the life course and the mechanisms by which they modify the relationship between CSVD and dementia are needed.

Well-chosen biomarkers have the potential to increase the efficiency of clinical trials and drug discovery and should show good precision as well as clinical validity.

We suggest measures that operationalize these criteria and describe a general approach that can be used for inference-based comparisons of biomarker performance. The methods are applied to measures obtained from structural magnetic resonance imaging (MRI) from individuals with mild dementia (n = 70) or mild cognitive impairment (MCI; n = 303) enrolled in the Alzheimer's Disease Neuroimaging Initiative.

Ventricular volume and hippocampal volume showed the best precision in detecting change over time in both individuals with MCI and with dementia. Differences in clinical validity varied by group.

The methodology presented provides a standardized framework for comparison of biomarkers across modalities and across different methods used to generate similar measures and will help in the search for the most promising biomarkers.

A framework for comparison of biomarkers on pre-defined criteria is presented. Criteria for comparison include precision in capturing change and clinical validity. Ventricular volume has high precision in change for both dementia and mild cognitive impairment (MCI) trials. Imaging measures' performance in clinical validity varies more for dementia than for MCI.

This systematic review examines the prevalence, underlying mechanisms, cohort characteristics, evaluation criteria, and cohort types in white matter hyperintensity (WMH) pipeline and implementation literature spanning the last two decades. Following Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) guidelines, we categorized WMH segmentation tools based on their methodologies from January 1, 2000, to November 18, 2022. Inclusion criteria involved articles using openly available techniques with detailed descriptions, focusing on WMH as a primary outcome. Our analysis identified 1007 visual rating scales, 118 pipeline development articles, and 509 implementation articles. These studies predominantly explored aging, dementia, psychiatric disorders, and small vessel disease, with aging and dementia being the most prevalent cohorts. Deep learning emerged as the most frequently developed segmentation technique, indicative of a heightened scrutiny in new technique development over the past two decades. We illustrate observed patterns and discrepancies between published and implemented WMH techniques. Despite increasingly sophisticated quantitative segmentation options, visual rating scales persist, with the SPM technique being the most utilized among quantitative methods and potentially serving as a reference standard for newer techniques. Our findings highlight the need for future standards in WMH segmentation, and we provide recommendations based on these observations.

Neuroanatomical normative modeling captures individual variability in Alzheimer's disease (AD). Here we used normative modeling to track individuals' disease progression in people with mild cognitive impairment (MCI) and patients with AD.

Cortical and subcortical normative models were generated using healthy controls (n ≈ 58k). These models were used to calculate regional z scores in 3233 T1-weighted magnetic resonance imaging time-series scans from 1181 participants. Regions with z scores < -1.96 were classified as outliers mapped on the brain and summarized by total outlier count (tOC).

tOC increased in AD and in people with MCI who converted to AD and also correlated with multiple non-imaging markers. Moreover, a higher annual rate of change in tOC increased the risk of progression from MCI to AD. Brain outlier maps identified the hippocampus as having the highest rate of change.

Individual patients' atrophy rates can be tracked by using regional outlier maps and tOC.

Neuroanatomical normative modeling was applied to serial Alzheimer's disease (AD) magnetic resonance imaging (MRI) data for the first time. Deviation from the norm (outliers) of cortical thickness or brain volume was computed in 3233 scans. The number of brain-structure outliers increased over time in people with AD. Patterns of change in outliers varied markedly between individual patients with AD. People with mild cognitive impairment whose outliers increased over time had a higher risk of progression from AD.

Population-based evidence in the interrelationships among hearing, brain structure, and cognition is limited. This study aims to investigate the cross-sectional associations of peripheral hearing, brain imaging measures, and cognitive function with speech-in-noise performance among older adults.

We studied 602 participants in the Aging and Cognitive Health Evaluation in Elders (ACHIEVE) brain magnetic resonance imaging (MRI) ancillary study, including 427 ACHIEVE baseline (2018-2020) participants with hearing loss and 175 Atherosclerosis Risk in Communities Neurocognitive Study Visit 6/7 (2016-2017/2018-2019) participants with normal hearing. Speech-in-noise performance, as outcome of interest, was assessed by the Quick Speech-in-Noise (QuickSIN) test (range: 0-30; higher = better). Predictors of interest included (a) peripheral hearing assessed by pure-tone audiometry; (b) brain imaging measures: structural MRI measures, white matter hyperintensities, and diffusion tensor imaging measures; and (c) cognitive performance assessed by a battery of 10 cognitive tests. All predictors were standardized to z scores. We estimated the differences in QuickSIN associated with every standard deviation (SD) worse in each predictor (peripheral hearing, brain imaging, and cognition) using multivariable-adjusted linear regression, adjusting for demographic variables, lifestyle, and disease factors (Model 1), and, additionally, for other predictors to assess independent associations (Model 2).

Participants were aged 70-84 years, 56% female, and 17% Black. Every SD worse in better-ear 4-frequency pure-tone average was associated with worse QuickSIN (-4.89, 95% confidence interval, CI [-5.57, -4.21]) when participants had peripheral hearing loss, independent of other predictors. Smaller temporal lobe volume was associated with worse QuickSIN, but the association was not independent of other predictors (-0.30, 95% CI [-0.86, 0.26]). Every SD worse in global cognitive performance was independently associated with worse QuickSIN (-0.90, 95% CI [-1.30, -0.50]).

Peripheral hearing and cognitive performance are independently associated with speech-in-noise performance among dementia-free older adults. The ongoing ACHIEVE trial will elucidate the effect of a hearing intervention that includes amplification and auditory rehabilitation on speech-in-noise understanding in older adults.

https://doi.org/10.23641/asha.25733679.

Pooling MRI data from multiple datasets requires harmonization to reduce undesired inter-site variabilities, while preserving effects of biological variables (or covariates). The popular harmonization approach ComBat uses a mixed effect regression framework that explicitly accounts for covariate distribution differences across datasets. There is also significant interest in developing harmonization approaches based on deep neural networks (DNNs), such as conditional variational autoencoder (cVAE). However, current DNN approaches do not explicitly account for covariate distribution differences across datasets. Here, we provide mathematical results, suggesting that not accounting for covariates can lead to suboptimal harmonization. We propose two DNN-based covariate-aware harmonization approaches: covariate VAE (coVAE) and DeepResBat. The coVAE approach is a natural extension of cVAE by concatenating covariates and site information with site- and covariate-invariant latent representations. DeepResBat adopts a residual framework inspired by ComBat. DeepResBat first removes the effects of covariates with nonlinear regression trees, followed by eliminating site differences with cVAE. Finally, covariate effects are added back to the harmonized residuals. Using three datasets from three continents with a total of 2787 participants and 10085 anatomical T1 scans, we find that DeepResBat and coVAE outperformed ComBat, CovBat and cVAE in terms of removing dataset differences, while enhancing biological effects of interest. However, coVAE hallucinates spurious associations between anatomical MRI and covariates even when no association exists. Future studies proposing DNN-based harmonization approaches should be aware of this false positive pitfall. Overall, our results suggest that DeepResBat is an effective deep learning alternative to ComBat. Code for DeepResBat can be found here: https://github.com/ThomasYeoLab/CBIG/tree/master/stable_projects/harmonization/An2024_DeepResBat.

Neuroimaging data acquired using multiple scanners or protocols are increasingly available. However, such data exhibit technical artifacts across batches which introduce confounding and decrease reproducibility. This is especially true when multi-batch data are analyzed using complex downstream models which are more likely to pick up on and implicitly incorporate batch-related information. Previously proposed image harmonization methods have sought to remove these batch effects; however, batch effects remain detectable in the data after applying these methods. We present DeepComBat, a deep learning harmonization method based on a conditional variational autoencoder and the ComBat method. DeepComBat combines the strengths of statistical and deep learning methods in order to account for the multivariate relationships between features while simultaneously relaxing strong assumptions made by previous deep learning harmonization methods. As a result, DeepComBat can perform multivariate harmonization while preserving data structure and avoiding the introduction of synthetic artifacts. We apply this method to cortical thickness measurements from a cognitive-aging cohort and show DeepComBat qualitatively and quantitatively outperforms existing methods in removing batch effects while preserving biological heterogeneity. Additionally, DeepComBat provides a new perspective for statistically motivated deep learning harmonization methods.

Mild traumatic brain injury (mTBI) can result in lasting brain damage that is often too subtle to detect by qualitative visual inspection on conventional MR imaging. Although a number of FDA-cleared MR neuroimaging tools have demonstrated changes associated with mTBI, they are still under-utilized in clinical practice.

We investigated a group of 65 individuals with predominantly mTBI (60 mTBI, 48 due to motor-vehicle collision, mean age 47 ± 13 years, 27 men and 38 women) with MR neuroimaging performed in a median of 37 months post-injury. We evaluated abnormalities in brain volumetry including analysis of left-right asymmetry by quantitative volumetric analysis, cerebral perfusion by pseudo-continuous arterial spin labeling (PCASL), white matter microstructure by diffusion tensor imaging (DTI), and neurometabolites via magnetic resonance spectroscopy (MRS).

All participants demonstrated atrophy in at least one lobar structure or increased lateral ventricular volume. The globus pallidi and cerebellar grey matter were most likely to demonstrate atrophy and asymmetry. Perfusion imaging revealed significant reductions of cerebral blood flow in both occipital and right frontoparietal regions. Diffusion abnormalities were relatively less common though a subset analysis of participants with higher resolution DTI demonstrated additional abnormalities. All participants showed abnormal levels on at least one brain metabolite, most commonly in choline and N-acetylaspartate.

We demonstrate the presence of coup-contrecoup perfusion injury patterns, widespread atrophy, regional brain volume asymmetry, and metabolic aberrations as sensitive markers of chronic mTBI sequelae. Our findings expand the historic focus on quantitative imaging of mTBI with DTI by highlighting the complementary importance of volumetry, arterial spin labeling perfusion and magnetic resonance spectroscopy neurometabolite analyses in the evaluation of chronic mTBI.

Alzheimer's disease (AD) pathology is featured by the extracellular accumulation of amyloid-β (Aβ) plaques and intracellular tau neurofibrillary tangles in the brain. We studied whether Aβ and tau accumulation are independently associated with future cognitive decline in the AD continuum.

Data were acquired from the Alzheimer's Disease Neuroimaging Initiative (ADNI) public database. A total of 1272 participants were selected based on the availability of Aβ-PET and CSF tau at baseline and of those 777 participants with follow-up visits.

We found that Aβ-PET and CSF tau pathology were related to cognitive decline across the AD clinical spectrum, both as potential predictors for dementia progression. Among them, Aβ-PET (A + T- subjects) is an independent reliable predictor of longitudinal cognitive decline in terms of ADAS-13, ADNI-MEM, and MMSE scores rather than tau pathology (A - T+ subjects), indicating tau accumulation is not closely correlated with future cognitive impairment without being driven by Aβ deposition. Of note, a high percentage of APOE ε4 carriers with Aβ pathology (A+) develop poor memory and learning capacity. Interestingly, this condition is not recurrence in terms of the ADNI-MEM domain when adding APOE ε4 status. Finally, the levels of Aβ-PET SUVR related to glucose hypometabolism more strongly in subjects with A + T- than A - T+ both happen at baseline and longitudinal changes.

In conclusion, Aβ-PET alone without tau pathology (A + T-) measure is an independent reliable predictor of longitudinal cognitive decline but may nonetheless forecast different status of dementia progression. However, tau accumulation alone without Aβ pathology background (A - T+) was not enough to be an independent predictor of cognitive worsening.

Electronic health records (EHRs) represent a comprehensive resource of a patient's medical history. EHRs are essential for utilizing advanced technologies such as deep learning (DL), enabling healthcare providers to analyze extensive data, extract valuable insights, and make precise and data-driven clinical decisions. DL methods such as recurrent neural networks (RNN) have been utilized to analyze EHR to model disease progression and predict diagnosis. However, these methods do not address some inherent irregularities in EHR data such as irregular time intervals between clinical visits. Furthermore, most DL models are not interpretable. In this study, we propose two interpretable DL architectures based on RNN, namely time-aware RNN (TA-RNN) and TA-RNN-autoencoder (TA-RNN-AE) to predict patient's clinical outcome in EHR at the next visit and multiple visits ahead, respectively. To mitigate the impact of irregular time intervals, we propose incorporating time embedding of the elapsed times between visits. For interpretability, we propose employing a dual-level attention mechanism that operates between visits and features within each visit.

The results of the experiments conducted on Alzheimer's Disease Neuroimaging Initiative (ADNI) and National Alzheimer's Coordinating Center (NACC) datasets indicated the superior performance of proposed models for predicting Alzheimer's Disease (AD) compared to state-of-the-art and baseline approaches based on F2 and sensitivity. Additionally, TA-RNN showed superior performance on the Medical Information Mart for Intensive Care (MIMIC-III) dataset for mortality prediction. In our ablation study, we observed enhanced predictive performance by incorporating time embedding and attention mechanisms. Finally, investigating attention weights helped identify influential visits and features in predictions.

https://github.com/bozdaglab/TA-RNN.

Accurate prediction of the progression from mild cognitive impairment (MCI) to Alzheimer's disease (AD) is crucial for disease management. Machine learning techniques have demonstrated success in classifying AD and MCI cases, particularly with the use of resting-state functional magnetic resonance imaging (rs-fMRI) data.This study utilized three years of rs-fMRI data from the ADNI, involving 142 patients with stable MCI (sMCI) and 136 with progressive MCI (pMCI). Graph signal processing was applied to filter rs-fMRI data into low, middle, and high frequency bands. Connectivity-based features were derived from both filtered and unfiltered data, resulting in a comprehensive set of 100 features, including global graph metrics, minimum spanning tree (MST) metrics, triadic interaction metrics, hub tendency metrics, and the number of links. Feature selection was enhanced using particle swarm optimization (PSO) and simulated annealing (SA). A support vector machine (SVM) with a radial basis function (RBF) kernel and a 10-fold cross-validation setup were employed for classification. The proposed approach demonstrated superior performance, achieving optimal accuracy with minimal feature utilization. When PSO selected five features, SVM exhibited accuracy, specificity, and sensitivity rates of 77%, 70%, and 83%, respectively. The identified features were as follows: (Mean of clustering coefficient, Mean of strength)/Radius/(Mean Eccentricity, and Modularity) from low/middle/high frequency bands of graph. The study highlights the efficacy of the proposed framework in identifying individuals at risk of AD development using a parsimonious feature set. This approach holds promise for advancing the precision of MCI to AD progression prediction, aiding in early diagnosis and intervention strategies.

Mild cognitive impairment (MCI) is an important stage in Alzheimer's disease (AD) research, focusing on early pathogenic factors and mechanisms. Examining MCI patient subtypes and identifying their cognitive and neuropathological patterns as the disease progresses can enhance our understanding of the heterogeneous disease progression in the early stages of AD. However, few studies have thoroughly analyzed the subtypes of MCI, such as the cortical atrophy, and disease development characteristics of each subtype.

In this study, 396 individuals with MCI, 228 cognitive normal (CN) participants, and 192 AD patients were selected from ADNI database, and a semi-supervised mixture expert algorithm (MOE) with multiple classification boundaries was constructed to define AD subtypes. Moreover, the subtypes of MCI were obtained by using the multivariate linear boundary mapping of support vector machine (SVM). Then, the gray matter atrophy regions and severity of each MCI subtype were analyzed and the features of each subtype in demography, pathology, cognition, and disease progression were explored combining the longitudinal data collected for 2 years and analyzed important factors that cause conversion of MCI were analyzed.

Three MCI subtypes were defined by MOE algorithm, and the three subtypes exhibited their own features in cortical atrophy. Nearly one-third of patients diagnosed with MCI have almost no significant difference in cerebral cortex from the normal aging population, and their conversion rate to AD are the lowest. The subtype characterized by severe atrophy in temporal lobe and frontal lobe have a faster decline rate in many cognitive manifestations than the subtype featured with diffuse atrophy in the whole cortex. APOE ε4 is an important factor that cause the conversion of MCI to AD.

It was proved through the data-driven method that MCI collected by ADNI baseline presented different subtype features. The characteristics and disease development trajectories among subtypes can help to improve the prediction of clinical progress in the future and also provide necessary clues to solve the classification accuracy of MCI.

Fairness is crucial in machine learning to prevent bias based on sensitive attributes in classifier predictions. However, the pursuit of strict fairness often sacrifices accuracy, particularly when significant prevalence disparities exist among groups, making classifiers less practical. For example, Alzheimer's disease (AD) is more prevalent in women than men, making equal treatment inequitable for females. Accounting for prevalence ratios among groups is essential for fair decision-making. In this paper, we introduce prior knowledge for fairness, which incorporates prevalence ratio information into the fairness constraint within the Empirical Risk Minimization (ERM) framework. We develop the Prior-knowledge-guided Fair ERM (PFERM) framework, aiming to minimize expected risk within a specified function class while adhering to a prior-knowledge-guided fairness constraint. This approach strikes a flexible balance between accuracy and fairness. Empirical results confirm its effectiveness in preserving fairness without compromising accuracy.

Longitudinal hippocampal atrophy is commonly used as progressive marker assisting clinical diagnose of dementia. However, precise quantification of the atrophy is limited by longitudinal segmentation errors resulting from MRI artifacts across multiple independent scans. To accurately segment the hippocampal morphology from longitudinal 3T T1-weighted MR images, we propose a diffeomorphic geodesic guided deep learning method called the GeoLongSeg to mitigate the longitudinal variabilities that unrelated to diseases by enhancing intra-individual morphological consistency. Specifically, we integrate geodesic shape regression, an evolutional model that estimates smooth deformation process of anatomical shapes, into a two-stage segmentation network. We adopt a 3D U-Net in the first-stage network with an enhanced attention mechanism for independent segmentation. Then, a hippocampal shape evolutional trajectory is estimated by geodesic shape regression and fed into the second network to refine the independent segmentation. We verify that GeoLongSeg outperforms other four state-of-the-art segmentation pipelines in longitudinal morphological consistency evaluated by test-retest reliability, variance ratio and atrophy trajectories. When assessing hippocampal atrophy in longitudinal data from the Alzheimer's Disease Neuroimaging Initiative (ADNI), results based on GeoLongSeg exhibit spatial and temporal local atrophy in bilateral hippocampi of dementia patients. These features derived from GeoLongSeg segmentation exhibit the greatest discriminatory capability compared to the outcomes of other methods in distinguishing between patients and normal controls. Overall, GeoLongSeg provides an accurate and efficient segmentation network for extracting hippocampal morphology from longitudinal MR images, which assist precise atrophy measurement of the hippocampus in early stage of dementia.

Whole brain delineation (WBD) is utilized in neuroimaging analysis for data preprocessing and deriving whole brain image metrics. Current automated WBD techniques for analysis of preclinical brain MRI data show limited accuracy when images present with significant neuropathology and anatomical deformations, such as that resulting from organophosphate intoxication (OPI) and Alzheimer's Disease (AD), and inadequate generalizability.

A modified 2D U-Net framework was employed for WBD of MRI rodent brains, consisting of 27 convolutional layers, batch normalization, two dropout layers and data augmentation, after training parameter optimization. A total of 265 T2-weighted 7.0 T MRI scans were utilized for the study, including 125 scans of an OPI rat model for neural network training. For testing and validation, 20 OPI rat scans and 120 scans of an AD rat model were utilized. U-Net performance was evaluated using Dice coefficients (DC) and Hausdorff distances (HD) between the U-Net-generated and manually segmented WBDs.

The U-Net achieved a DC (median[range]) of 0.984[0.936-0.990] and HD of 1.69[1.01-6.78] mm for OPI rat model scans, and a DC (mean[range]) of 0.975[0.898-0.991] and HD of 1.49[0.86-3.89] for the AD rat model scans.

The proposed approach is fully automated and robust across two rat strains and longitudinal brain changes with a computational speed of 8 seconds/scan, overcoming limitations of manual segmentation.

The modified 2D U-Net provided a fully automated, efficient, and generalizable segmentation approach that achieved high accuracy across two disparate rat models of neurological diseases.

We investigated a potential sex difference in the relationship between alcohol consumption, brain age gap and cognitive function in older adults without cognitive impairment from the population-based Mayo Clinic Study of Aging.

Self-reported alcohol consumption was collected using the food-frequency questionnaire. A battery of cognitive testing assessed performance in four different domains: attention, memory, language, and visuospatial. Brain magnetic resonance imaging (MRI) was conducted using 3-T scanners (Signa; GE Healthcare). Brain age was estimated using the Brain-Age Regression Analysis and Computational Utility Software (BARACUS). We calculated the brain age gap as the difference between predicted brain age and chronological age.

The sample consisted of 269 participants [55% men (n=148) and 45% women (n=121) with a mean age of 79.2 ± 4.6 and 79.5 ± 4.7 years respectively]. Women had significantly better performance compared to men in memory, (1.12 ± 0.87 vs 0.57 ± 0.89, P<0.0001) language (0.66 ± 0.8 vs 0.33 ± 0.72, P=0.0006) and attention (0.79 ± 0.87 vs 0.39 ± 0.83, P=0.0002) z-scores. Men scored higher in visuospatial skills (0.71 ± 0.91 vs 0.44 ± 0.90, P=0.016). Compared to participants who reported zero alcohol drinking (n=121), those who reported alcohol consumption over the year prior to study enrollment (n=148) scored significantly higher in all four cognitive domains [memory: F3,268 = 5.257, P=0.002, Language: F3,258 = 12.047, P<0.001, Attention: F3,260 = 22.036, P<0.001, and Visuospatial: F3,261 = 9.326, P<0.001] after correcting for age and years of education. In addition, we found a significant positive correlation between alcohol consumption and the brain age gap (P=0.03). Post hoc regression analysis for each sex with language z-score revealed a significant negative correlation between brain age gap and language z-scores in women only (P=0.008).

Among older adults who report alcohol drinking, there is a positive association between higher average daily alcohol consumption and accelerated brain aging despite the fact that drinkers had better cognitive performance compared to zero drinkers. In women only, accelerated brain aging is associated with worse performance in language cognitive domain. Older adult women seem to be vulnerable to the negative effects of alcohol on brain structure and on certain cognitive functions.

Neuropsychiatric symptoms (NPS) are prevalent in cognitively impaired individuals including Alzheimer's disease (AD) dementia and mild cognitive impairment (MCI). Whereas several studies have reported the associations between NPS with AD pathologic biomarkers and cerebral small vessel disease (SVD), but it remains unknown whether AD pathology and SVD contribute to different sub-syndromes independently or aggravate same symptoms synergistically.

We included 445 cognitively impaired individuals (including 316 MCI and 129 AD) with neuropsychiatric, cerebrospinal fluid (CSF) biomarkers (Aβ42, p-tau, and t-tau) and multi-model MRI data. Psychiatric symptoms were accessed by using the Neuropsychiatric Inventory (NPI). Visual assessment of SVD (white matter hyperintensity, microbleed, perivascular space, lacune) on MRI images was performed by experienced radiologist. Linear regression analyses were conducted to test the association between neuropsychiatric symptoms with AD pathology and CSVD burden after adjustment for age, sex, education, apolipoprotein E (APOE) ε4 carrier status, and clinical diagnosis.

The NPI total scores were related to microbleed (estimate 2.424; 95% CI [0.749, 4.099]; P =0.005). Considering the sub-syndromes, the hyperactivity was associated with microbleed (estimate 0.925; 95% CI [0.115, 1.735]; P =0.025), whereas the affective symptoms were correlated to CSF level of Aβ42 (estimate -0.006; 95% CI [-0.011, -0.002]; P =0.005). Furthermore, we found the apathy sub-syndrome was associated with CSF t-tau/Aβ42 (estimate 0.636; 95% CI [0.078, 1.194]; P =0.041) and microbleed (estimate 0.693; 95% CI [0.046, 1.340]; P =0.036). In addition, we found a significant interactive effect between CSF t-tau/Aβ42 and microbleed (estimate 0.993; 95% CI [0.360, 1.626]; P =0.019) on severity of apathy sub-syndrome.

Our study showed that CSF Aβ42 was associated with affective symptoms, but microbleed was correlated with hyperactivity and apathy, suggesting the effect of AD pathology and SVD on different neuropsychiatric sub-syndromes.