Interpretable deep learning of myelin histopathology in age-related cognitive impairment

The AI Future of Emergency Medicine

In the coming years, artificial intelligence (AI) and machine learning will likely give rise to profound changes in the field of emergency medicine, and medicine more broadly. This article discusses these anticipated changes in terms of 3 overlapping yet distinct stages of AI development. It reviews some fundamental concepts in AI and explores their relation to clinical practice, with a focus on emergency medicine. In addition, it describes some of the applications of AI in disease diagnosis, prognosis, and treatment, as well as some of the practical issues that they raise, the barriers to their implementation, and some of the legal and regulatory challenges they create.

Cognitive and neuroscientific perspectives of healthy ageing

With dementia incidence projected to escalate significantly within the next 25 years, the United Nations declared 2021-2030 the Decade of Healthy Ageing, emphasising cognition as a crucial element. As a leading discipline in cognition and ageing research, psychology is well-equipped to offer insights for translational research, clinical practice, and policy-making. In this comprehensive review, we discuss the current state of knowledge on age-related changes in cognition and psychological health. We discuss cognitive changes during ageing, including (a) heterogeneity in the rate, trajectory, and characteristics of decline experienced by older adults, (b) the role of cognitive reserve in age-related cognitive decline, and (c) the potential for cognitive training to slow this decline. We also examine ageing and cognition through multiple theoretical perspectives. We highlight critical unresolved issues, such as the disparate implications of subjective versus objective measures of cognitive decline and the insufficient evaluation of cognitive training programs. We suggest future research directions, and emphasise interdisciplinary collaboration to create a more comprehensive understanding of the factors that modulate cognitive ageing.

Disentangling and quantifying the relative cognitive impact of concurrent mixed neurodegenerative pathologies

Neurodegenerative pathologies such as Alzheimer disease neuropathologic change (ADNC), Lewy body disease (LBD), limbic-predominant age-related TDP-43 encephalopathy neuropathologic change (LATE-NC), and cerebrovascular disease (CVD) frequently coexist, but little is known about the exact contribution of each pathology to cognitive decline and dementia in subjects with mixed pathologies. We explored the relative cognitive impact of concurrent common and rare neurodegenerative pathologies employing multivariate logistic regression analysis adjusted for age, gender, and level of education. We analyzed a cohort of 6,262 subjects from the National Alzheimer's Coordinating Center database, ranging from 0 to 6 comorbid neuropathologic findings per individual, where 95.7% of individuals had at least 1 neurodegenerative finding at autopsy and 75.5% had at least 2 neurodegenerative findings. We identified which neuropathologic entities correlate most frequently with one another and demonstrated that the total number of pathologies per individual was directly correlated with cognitive performance as assessed by Clinical Dementia Rating (CDR®) and Mini-Mental State Examination (MMSE). We show that ADNC, LBD, LATE-NC, CVD, hippocampal sclerosis, Pick disease, and FTLD-TDP significantly impact overall cognition as independent variables. More specifically, ADNC significantly affected all assessed cognitive domains, LBD affected attention, processing speed, and language, LATE-NC primarily affected tests related to logical memory and language, while CVD and other less common pathologies (including Pick disease, progressive supranuclear palsy, and corticobasal degeneration) had more variable neurocognitive effects. Additionally, ADNC, LBD, and higher numbers of comorbid neuropathologies were associated with the presence of at least one APOE ε4 allele, and ADNC and higher numbers of neuropathologies were inversely correlated with APOE ε2 alleles. Understanding the mechanisms by which individual and concomitant neuropathologies affect cognition and the degree to which each contributes is an imperative step in the development of biomarkers and disease-modifying therapeutics, particularly as these medical interventions become more targeted and personalized.

Spatial proteomics of hippocampal subfield-specific pathology in Alzheimer's disease and primary age-related tauopathy

INTRODUCTION

Alzheimer's disease (AD) and primary age-related tauopathy (PART) both harbor 3R/4R hyperphosphorylated-tau (p-tau)-positive neurofibrillary tangles (NFTs) but differ in the spatial p-tau development in the hippocampus.

METHODS

Using Nanostring GeoMx Digital Spatial Profiling, we compared protein expression within hippocampal subregions in NFT-bearing and non-NFT-bearing neurons in AD (n = 7) and PART (n = 7) subjects.

RESULTS

Proteomic measures of synaptic health were inversely correlated with the subregional p-tau burden in AD and PART, and there were numerous differences in proteins involved in proteostasis, amyloid beta (Aβ) processing, inflammation, microglia, oxidative stress, and neuronal/synaptic health between AD and PART and between definite PART and possible PART.

DISCUSSION

These results suggest subfield-specific proteome differences that may explain some of the differences in Aβ and p-tau distribution and apparent pathogenicity. In addition, hippocampal neurons in possible PART may have more in common with AD than with definite PART, highlighting the importance of Aβ in the pathologic process.

HIGHLIGHTS

Synaptic health is inversely correlated with local p-tau burden. The proteome of NFT- and non-NFT-bearing neurons is influenced by the presence of Aβ in the hippocampus. Neurons in possible PART cases share more proteomic similarities with neurons in ADNC than they do with neurons in definite PART cases.

Histopathologic brain age estimation via multiple instance learning

Understanding age acceleration, the discordance between biological and chronological age, in the brain can reveal mechanistic insights into normal physiology as well as elucidate pathological determinants of age-related functional decline and identify early disease changes in the context of Alzheimer's and other disorders. Histopathological whole slide images provide a wealth of pathologic data on the cellular level that can be leveraged to build deep learning models to assess age acceleration. Here, we used a collection of digitized human post-mortem hippocampal sections to develop a histological brain age estimation model. Our model predicted brain age within a mean absolute error of 5.45 ± 0.22 years, with attention weights corresponding to neuroanatomical regions vulnerable to age-related changes. We found that histopathologic brain age acceleration had significant associations with clinical and pathologic outcomes that were not found with epigenetic based measures. Our results indicate that histopathologic brain age is a powerful, independent metric for understanding factors that contribute to brain aging.

High-throughput digital quantification of Alzheimer disease pathology and associated infrastructure in large autopsy studies

High-throughput digital pathology offers considerable advantages over traditional semiquantitative and manual methods of counting pathology. We used brain tissue from 5 clinical-pathologic cohort studies of aging; the Religious Orders Study, the Rush Memory and Aging Project, the Minority Aging Research Study, the African American Clinical Core, and the Latino Core to (1) develop a workflow management system for digital pathology processes, (2) optimize digital algorithms to quantify Alzheimer disease (AD) pathology, and (3) harmonize data statistically. Data from digital algorithms for the quantification of β-amyloid (Aβ, n = 413) whole slide images and tau-tangles (n = 639) were highly correlated with manual pathology data (r = 0.83 to 0.94). Measures were robust and reproducible across different magnifications and repeated scans. Digital measures for Aβ and tau-tangles across multiple brain regions reproduced established patterns of correlations, even when samples were stratified by clinical diagnosis. Finally, we harmonized newly generated digital measures with historical measures across multiple large autopsy-based studies. We describe a multidisciplinary approach to develop a digital pathology pipeline that reproducibly identifies AD neuropathologies, Aβ load, and tau-tangles. Digital pathology is a powerful tool that can overcome critical challenges associated with traditional microscopy methods.

Artificial Intelligence: Deciphering the Links between Psychiatric Disorders and Neurodegenerative Disease

Artificial Intelligence (AI), which is the general term used to describe technology that simulates human cognition [...].

The relationship between hippocampal amyloid beta burden and spatial distribution of neurofibrillary degeneration

INTRODUCTION

Neurofibrillary degeneration in Alzheimer's disease (AD) typically involves the entorhinal cortex and CA1 subregion of the hippocampus early in the disease process, whereas in primary age-related tauopathy (PART), there is an early selective vulnerability of the CA2 subregion.

METHODS

Image analysis-based quantitative pixel assessments were used to objectively evaluate amyloid beta (Aβ) burden in the medial temporal lobe in relation to the distribution of hyperphosphorylated-tau (p-tau) in 142 cases of PART and AD.

RESULTS

Entorhinal, CA1, CA3, and CA4 p-tau deposition levels are significantly correlated with Aβ burden, while CA2 p-tau is not. Furthermore, the CA2/CA1 p-tau ratio is inversely correlated with Aβ burden and distribution. In addition, cognitive impairment is correlated with overall p-tau burden.

DISCUSSION

These data indicate that the presence and extent of medial temporal lobe Aβ may determine the distribution and spread of neurofibrillary degeneration. The resulting p-tau distribution patterns may discriminate between PART and AD.

HIGHLIGHTS

Subregional hyperphosphorylated-tau (p-tau) distribution is influenced by hippocampal amyloid beta burden. Higher CA2/CA1 p-tau ratio is predictive of primary age-related tauopathy-like neuropathology. Cognitive function is correlated with the overall hippocampal p-tau burden.

Artificial Intelligence as a Tool for the Diagnosis and Treatment of Neurodegenerative Diseases

While humans have much in common biologically with other mammalian species, they are largely distinguished by their innate intelligence, specifically, their ability to generate complex and sophisticated tools [...].

Diagnosis of Alzheimer Disease and Tauopathies on Whole-Slide Histopathology Images Using a Weakly Supervised Deep Learning Algorithm

Neuropathologic assessment during autopsy is the gold standard for diagnosing neurodegenerative disorders. Neurodegenerative conditions, such as Alzheimer disease (AD) neuropathological change, are a continuous process from normal aging rather than categorical; therefore, diagnosing neurodegenerative disorders is a complicated task. We aimed to develop a pipeline for diagnosing AD and other tauopathies, including corticobasal degeneration (CBD), globular glial tauopathy, Pick disease, and progressive supranuclear palsy. We used a weakly supervised deep learning-based approach called clustering-constrained-attention multiple-instance learning (CLAM) on the whole-slide images (WSIs) of patients with AD (n = 30), CBD (n = 20), globular glial tauopathy (n = 10), Pick disease (n = 20), and progressive supranuclear palsy (n = 20), as well as nontauopathy controls (n = 21). Three sections (A: motor cortex; B: cingulate gyrus and superior frontal gyrus; and C: corpus striatum) that had been immunostained for phosphorylated tau were scanned and converted to WSIs. We evaluated 3 models (classic multiple-instance learning, single-attention-branch CLAM, and multiattention-branch CLAM) using 5-fold cross-validation. Attention-based interpretation analysis was performed to identify the morphologic features contributing to the classification. Within highly attended regions, we also augmented gradient-weighted class activation mapping to the model to visualize cellular-level evidence of the model's decisions. The multiattention-branch CLAM model using section B achieved the highest area under the curve (0.970 ± 0.037) and diagnostic accuracy (0.873 ± 0.087). A heatmap showed the highest attention in the gray matter of the superior frontal gyrus in patients with AD and the white matter of the cingulate gyrus in patients with CBD. Gradient-weighted class activation mapping showed the highest attention in characteristic tau lesions for each disease (eg, numerous tau-positive threads in the white matter inclusions for CBD). Our findings support the feasibility of deep learning-based approaches for the classification of neurodegenerative disorders on WSIs. Further investigation of this method, focusing on clinicopathologic correlations, is warranted.

Preanalytic variable effects on segmentation and quantification machine learning algorithms for amyloid-β analyses on digitized human brain slides

Computational machine learning (ML)-based frameworks could be advantageous for scalable analyses in neuropathology. A recent deep learning (DL) framework has shown promise in automating the processes of visualizing and quantifying different types of amyloid-β deposits as well as segmenting white matter (WM) from gray matter (GM) on digitized immunohistochemically stained slides. However, this framework has only been trained and evaluated on amyloid-β-stained slides with minimal changes in preanalytic variables. In this study, we evaluated select preanalytical variables including magnification, compression rate, and storage format using three digital slides scanners (Zeiss Axioscan Z1, Leica Aperio AT2, and Leica Aperio GT 450), on over 60 whole slide images, in a cohort of 14 cases having a spectrum of amyloid-β deposits. We conducted statistical comparisons of preanalytic variables with repeated measures analysis of variance evaluating the outputs of two DL frameworks for segmentation and object classification tasks. For both WM/GM segmentation and amyloid-β plaque classification tasks, there were statistical differences with respect to scanner types (p < 0.05) and magnifications (p < 0.05). Although small numbers of cases were analyzed, this pilot study highlights the significance of preanalytic variables that may alter the performance of ML algorithms.

The status of digital pathology and associated infrastructure within Alzheimer's Disease Centers

Digital pathology (DP) has transformative potential, especially for Alzheimer disease and related disorders. However, infrastructure barriers may limit adoption. To provide benchmarks and insights into implementation barriers, a survey was conducted in 2019 within National Institutes of Health's Alzheimer's Disease Centers (ADCs). Questions covered infrastructure, funding sources, and data management related to digital pathology. Of the 35 ADCs to which the survey was sent, 33 responded. Most respondents (81%) stated that their ADC had digital slide scanner access, with the most frequent brand being Aperio/Leica (62.9%). Approximately a third of respondents stated there were fees to utilize the scanner. For DP and machine learning (ML) resources, 41% of respondents stated none was supported by their ADC. For scanner purchasing and operations, 50% of respondents stated they received institutional support. Some were unsure of the file size of scanned digital images (37%) and total amount of storage space files occupied (50%). Most (76%) were aware of other departments at their institution working with ML; a similar (76%) percentage were unaware of multiuniversity or industry partnerships. These results demonstrate many ADCs have access to a digital slide scanner; additional investigations are needed to further understand hurdles to implement DP and ML workflows.

Neurodegeneration: 2023 update

This paper reviews ten highly impactful studies published in the previous year selected by the author from the neurodegenerative neuropathology literature. As in previous years, the focus is to highlight human tissue-based experimentation most relevant to neuropathologists. A concerted effort was made to balance the selected studies across disease categories, approaches, and methodologies to capture the breadth of the research landscape. Studies include an integrated proteomic and transcriptomic study of Alzheimer disease (AD) and new consensus diagnostic neuropathological criteria for progressive supranuclear palsy. A number of studies looking at TAR DNA-binding protein 43 (TDP-43) are highlighted. One examined interaction between AD and limbic age-related TDP-43 encephalopathy (LATE) and yet another demonstrated how TDP-43 represses cryptic exon inclusion in UNC13A, suggesting a novel pathogenic mechanism. Most surprisingly, three cryogenic electron microscopy (cryo-EM) studies showed that TMEM106B filaments form the core of TDP-43-positive inclusions. Cryo-EM revealed a prion protein amyloid structure from aggregates in Gerstmann-Sträussler-Scheinker disease. There was an elegant functional genomic study cataloging microglial gene expression in the human brain. A study shed light on how APOE influences chronic traumatic encephalopathy. A pathoanatomical study tested the dual hit hypothesis of Lewy body progression throughout the nervous system. And finally, deep learning continues to show its promise with application of a weakly supervised multiple instance learning paradigm to assess aging post-mortem brains.

Postmortem changes in brain cell structure: a review

Brain cell structure is a key determinant of neural function that is frequently altered in neurobiological disorders. Following the global loss of blood flow to the brain that initiates the postmortem interval (PMI), cells rapidly become depleted of energy and begin to decompose. To ensure that our methods for studying the brain using autopsy tissue are robust and reproducible, there is a critical need to delineate the expected changes in brain cell morphometry during the PMI. We searched multiple databases to identify studies measuring the effects of PMI on the morphometry (i.e. external dimensions) of brain cells. We screened 2119 abstracts, 361 full texts, and included 172 studies. Mechanistically, fluid shifts causing cell volume alterations and vacuolization are an early event in the PMI, while the loss of the ability to visualize cell membranes altogether is a later event. Decomposition rates are highly heterogenous and depend on the methods for visualization, the structural feature of interest, and modifying variables such as the storage temperature or the species. Geometrically, deformations of cell membranes are common early events that initiate within minutes. On the other hand, topological relationships between cellular features appear to remain intact for more extended periods. Taken together, there is an uncertain period of time, usually ranging from several hours to several days, over which cell membrane structure is progressively lost. This review may be helpful for investigators studying human postmortem brain tissue, wherein the PMI is an unavoidable aspect of the research.

Artificial intelligence-derived neurofibrillary tangle burden is associated with antemortem cognitive impairment

Tauopathies are a category of neurodegenerative diseases characterized by the presence of abnormal tau protein-containing neurofibrillary tangles (NFTs). NFTs are universally observed in aging, occurring with or without the concomitant accumulation of amyloid-beta peptide (Aβ) in plaques that typifies Alzheimer disease (AD), the most common tauopathy. Primary age-related tauopathy (PART) is an Aβ-independent process that affects the medial temporal lobe in both cognitively normal and impaired subjects. Determinants of symptomology in subjects with PART are poorly understood and require clinicopathologic correlation; however, classical approaches to staging tau pathology have limited quantitative reproducibility. As such, there is a critical need for unbiased methods to quantitatively analyze tau pathology on the histological level. Artificial intelligence (AI)-based convolutional neural networks (CNNs) generate highly accurate and precise computer vision assessments of digitized pathology slides, yielding novel histology metrics at scale. Here, we performed a retrospective autopsy study of a large cohort (n = 706) of human post-mortem brain tissues from normal and cognitively impaired elderly individuals with mild or no Aβ plaques (average age of death of 83.1 yr, range 55-110). We utilized a CNN trained to segment NFTs on hippocampus sections immunohistochemically stained with antisera recognizing abnormal hyperphosphorylated tau (p-tau), which yielded metrics of regional NFT counts, NFT positive pixel density, as well as a novel graph-theory based metric measuring the spatial distribution of NFTs. We found that several AI-derived NFT metrics significantly predicted the presence of cognitive impairment in both the hippocampus proper and entorhinal cortex (p < 0.0001). When controlling for age, AI-derived NFT counts still significantly predicted the presence of cognitive impairment (p = 0.04 in the entorhinal cortex; p = 0.04 overall). In contrast, Braak stage did not predict cognitive impairment in either age-adjusted or unadjusted models. These findings support the hypothesis that NFT burden correlates with cognitive impairment in PART. Furthermore, our analysis strongly suggests that AI-derived metrics of tau pathology provide a powerful tool that can deepen our understanding of the role of neurofibrillary degeneration in cognitive impairment.