Automated text-level semantic markers of Alzheimer's disease.
Ontology highlight
ABSTRACT: Introduction
Automated speech analysis has emerged as a scalable, cost-effective tool to identify persons with Alzheimer's disease dementia (ADD). Yet, most research is undermined by low interpretability and specificity.Methods
Combining statistical and machine learning analyses of natural speech data, we aimed to discriminate ADD patients from healthy controls (HCs) based on automated measures of domains typically affected in ADD: semantic granularity (coarseness of concepts) and ongoing semantic variability (conceptual closeness of successive words). To test for specificity, we replicated the analyses on Parkinson's disease (PD) patients.Results
Relative to controls, ADD (but not PD) patients exhibited significant differences in both measures. Also, these features robustly discriminated between ADD patients and HC, while yielding near-chance classification between PD patients and HCs.Discussion
Automated discourse-level semantic analyses can reveal objective, interpretable, and specific markers of ADD, bridging well-established neuropsychological targets with digital assessment tools.
SUBMITTER: Sanz C
PROVIDER: S-EPMC8759093 | biostudies-literature |
REPOSITORIES: biostudies-literature
Similar Datasets
Project description:Patients with the semantic variant of primary progressive aphasia, also known as semantic dementia, and Alzheimer's disease have deficits in semantic memory. However, few comparative studies have been performed to determine whether these patient groups have distinct semantic memory impairments. We asked 15 patients with semantic variant primary progressive aphasia and 57 patients with Alzheimer's disease to judge semantic category membership of coloured photos and printed words that are members of familiar natural and manufactured categories, and we related performance to grey matter atrophy. We found that both semantic variant primary progressive aphasia and Alzheimer's disease are significantly impaired on this task. Moreover, patients with semantic variant primary progressive aphasia had a significantly more prominent deficit for natural objects than their own deficit judging manufactured objects. Both semantic variant primary progressive aphasia and Alzheimer's disease had atrophy that included portions of the left temporal lobe. Regression analyses related performance in semantic variant primary progressive aphasia to ventral and medial portions of the left temporal lobe, while regression analyses in Alzheimer's disease related performance to these ventral and medial temporal areas as well as lateral temporal-parietal regions in the left hemisphere. We conclude that both semantic variant primary progressive aphasia and Alzheimer's disease are significantly impaired in a simple category membership judgement task and the selective impairment for natural kinds in semantic variant primary progressive aphasia is related in part to disease in visual association cortex in ventral-medial portions of the left temporal lobe. We discuss factors that may contribute to the semantic memory deficit in semantic variant primary progressive aphasia.
| S-EPMC3722350 | biostudies-literature
Project description:Concept-centered semantic maps were created based on a text-mining analysis of PubMed using the BiblioEngine_v2018 software. The objects ("concepts") of a semantic map can be MeSH-terms or other terms (names of proteins, diseases, chemical compounds, etc.) structured in the form of controlled vocabularies. The edges between the two objects were automatically calculated based on the index of semantic similarity, which is proportional to the number of publications related to both objects simultaneously. On the one hand, an individual semantic map created based on the already published papers allows us to trace scientific inquiry. On the other hand, a prospective analysis based on the study of PubMed search history enables us to determine the possible directions for future research.
| S-EPMC6083525 | biostudies-literature
Project description:Explicit semantic analysis (ESA) utilizes an immense Wikipedia index matrix in its interpreter part. This part of the analysis multiplies a large matrix by a term vector to produce a high-dimensional concept vector. A similarity measurement between two texts is performed between two concept vectors with numerous dimensions. The cost is expensive in both interpretation and similarity measurement steps. This paper proposes an economic scheme of ESA, named econo-ESA. We investigate two aspects of this proposal: dimensional reduction and experiments with various data. We use eight recycling test collections in semantic text similarity. The experimental results show that both the dimensional reduction and test collection characteristics can influence the results. They also show that an appropriate concept reduction of econo-ESA can decrease the cost with minor differences in the results from the original ESA.
| S-EPMC4003000 | biostudies-literature
Project description:ObjectiveSemantic role labeling (SRL), which extracts a shallow semantic relation representation from different surface textual forms of free text sentences, is important for understanding natural language. Few studies in SRL have been conducted in the medical domain, primarily due to lack of annotated clinical SRL corpora, which are time-consuming and costly to build. The goal of this study is to investigate domain adaptation techniques for clinical SRL leveraging resources built from newswire and biomedical literature to improve performance and save annotation costs.Materials and methodsMultisource Integrated Platform for Answering Clinical Questions (MiPACQ), a manually annotated SRL clinical corpus, was used as the target domain dataset. PropBank and NomBank from newswire and BioProp from biomedical literature were used as source domain datasets. Three state-of-the-art domain adaptation algorithms were employed: instance pruning, transfer self-training, and feature augmentation. The SRL performance using different domain adaptation algorithms was evaluated by using 10-fold cross-validation on the MiPACQ corpus. Learning curves for the different methods were generated to assess the effect of sample size.Results and conclusionWhen all three source domain corpora were used, the feature augmentation algorithm achieved statistically significant higher F-measure (83.18%), compared to the baseline with MiPACQ dataset alone (F-measure, 81.53%), indicating that domain adaptation algorithms may improve SRL performance on clinical text. To achieve a comparable performance to the baseline method that used 90% of MiPACQ training samples, the feature augmentation algorithm required <50% of training samples in MiPACQ, demonstrating that annotation costs of clinical SRL can be reduced significantly by leveraging existing SRL resources from other domains.
| S-EPMC4986662 | biostudies-literature
Project description:BackgroundThe aim of this study is to use classification methods to predict future onset of Alzheimer's disease in cognitively normal subjects through automated linguistic analysis.MethodsTo study linguistic performance as an early biomarker of AD, we performed predictive modeling of future diagnosis of AD from a cognitively normal baseline of Framingham Heart Study participants. The linguistic variables were derived from written responses to the cookie-theft picture-description task. We compared the predictive performance of linguistic variables with clinical and neuropsychological variables. The study included 703 samples from 270 participants out of which a dataset consisting of a single sample from 80 participants was held out for testing. Half of the participants in the test set developed AD symptoms before 85 years old, while the other half did not. All samples in the test set were collected during the cognitively normal period (before MCI). The mean time to diagnosis of mild AD was 7.59 years.FindingsSignificant predictive power was obtained, with AUC of 0.74 and accuracy of 0.70 when using linguistic variables. The linguistic variables most relevant for predicting onset of AD have been identified in the literature as associated with cognitive decline in dementia.InterpretationThe results suggest that language performance in naturalistic probes expose subtle early signs of progression to AD in advance of clinical diagnosis of impairment.FundingPfizer, Inc. provided funding to obtain data from the Framingham Heart Study Consortium, and to support the involvement of IBM Research in the initial phase of the study. The data used in this study was supported by Framingham Heart Study's National Heart, Lung, and Blood Institute contract (N01-HC-25195), and by grants from the National Institute on Aging grants (R01-AG016495, R01-AG008122) and the National Institute of Neurological Disorders and Stroke (R01-NS017950).
| S-EPMC7700896 | biostudies-literature
Project description:The progression of cognitive deficits in Alzheimer's disease and semantic dementia is accompanied by grey matter atrophy and white matter deterioration. The impact of neuronal loss on the structural network connectivity in these dementia subtypes is, however, not well understood. In order to gain a more refined knowledge of the topological organization of white matter alterations in dementia, we used a network-based approach to analyze the brain's structural connectivity network. Diffusion-weighted and anatomical MRI images of groups with eighteen Alzheimer's disease and six semantic dementia patients, as well as twenty-one healthy controls were recorded to reconstruct individual connectivity networks. Additionally, voxel-based morphometry, using grey and white matter volume, served to relate atrophy to altered structural connectivity. The analyses showed that Alzheimer's disease is characterized by decreased connectivity strength in various cortical regions. An overlap with grey matter loss was found only in the inferior frontal and superior temporal areas. In semantic dementia, significantly reduced network strength was found in the temporal lobes, which converged with grey and white matter atrophy. Therefore, this study demonstrated that the structural disconnection in early Alzheimer's disease goes beyond grey matter atrophy and is independent of white matter volume loss, an observation that was not found in semantic dementia.
| S-EPMC5147505 | biostudies-other
Project description:Hippocampal atrophy is a well-known feature of Alzheimer's disease (AD), but sensitivity and specificity of hippocampal volumetry are limited. Neuropathological studies have shown that hippocampal subfields are differentially vulnerable to AD; hippocampal subfield volumetry may thus prove to be more accurate than global hippocampal volumetry to detect AD.CA1, subiculum and other subfields were manually delineated from 40 healthy controls, 18 AD, 17 amnestic Mild Cognitive Impairment (aMCI), and 8 semantic dementia (SD) patients using a previously developed high resolution MRI procedure. Non-parametric group comparisons and receiver operating characteristic (ROC) analyses were conducted. Complementary analyses were conducted to evaluate differences of hemispheric asymmetry and anterior-predominance between AD and SD patients and to distinguish aMCI patients with or without ?-amyloid deposition as assessed by Florbetapir-TEP.Global hippocampi were atrophied in all three patient groups and volume decreases were maximal in the CA1 subfield (22% loss in aMCI, 27% in both AD and SD; all p < 0.001). In aMCI, CA1 volumetry was more accurate than global hippocampal measurement to distinguish patients from controls (areas under the ROC curve = 0.88 and 0.76, respectively; p = 0.05) and preliminary analyses suggest that it was independent from the presence of ?-amyloid deposition. In patients with SD, whereas the degree of CA1 and subiculum atrophy was similar to that found in AD patients, hemispheric and anterior-posterior asymmetry were significantly more marked than in AD with greater involvement of the left and anterior hippocampal subfields.The findings suggest that CA1 measurement is more sensitive than global hippocampal volumetry to detect structural changes at the pre-dementia stage, although the predominance of CA1 atrophy does not appear to be specific to AD pathophysiological processes.
| S-EPMC3791274 | biostudies-literature
Project description:PurposeTo evaluate the performance of machine learning algorithms on organ-level classification of semistructured pathology reports, to incorporate surgical pathology monitoring into an automated imaging recommendation follow-up engine.Materials and methodsThis retrospective study included 2013 pathology reports from patients who underwent abdominal imaging at a large tertiary care center between 2012 and 2018. The reports were labeled by two annotators as relevant to four abdominal organs: liver, kidneys, pancreas and/or adrenal glands, or none. Automated classification methods were compared: simple string matching, random forests, extreme gradient boosting, support vector machines, and two neural network architectures-convolutional neural networks and long short-term memory networks. Three methods from the literature were used to provide interpretability and qualitative validation of the learned network features.ResultsThe neural networks performed well on the four-organ classification task (F1 score: 96.3% for convolutional neural network and 96.7% for long short-term memory vs 89.9% for support vector machines, 93.9% for extreme gradient boosting, 82.8% for random forests, and 75.2% for simple string matching). Multiple methods were used to visualize the decision-making process of the network, verifying that the networks used similar heuristics to a human annotator. The neural networks were able to classify, with a high degree of accuracy, pathology reports written in unseen formats, suggesting the networks had learned a generalizable encoding of the salient features.ConclusionNeural network-based approaches achieve high performance on organ-level pathology report classification, suggesting that it is feasible to use them within automated tracking systems.© RSNA, 2019Supplemental material is available for this article.See also the commentary by Liu in this issue.
| S-EPMC8017395 | biostudies-literature
Project description:Overexpression of pro-inflammatory cytokines and cellular energy failure are associated with neuroinflammatory disorders, such as Alzheimer's disease. Transgenic mice homozygous for human ApoE4 gene, a well known AD and atherosclerosis animal model, show decreased levels of ATP, increased inflammatory cytokines level and accumulation of beta amyloid in the brain. All these findings are considered responsible for triggering cognitive decline. We have demonstrated that a single administration of the bacterial E. coli protein toxin CNF1 to aged apoE4 mice, beside inducing a strong amelioration of both spatial and emotional memory deficits, favored the cell energy restore through an increment of ATP content. This was accompanied by a modulation of cerebral Rho and Rac1 activity. Furthermore, CNF1 decreased the levels of beta amyloid accumulation and interleukin-1β expression in the hippocampus. Altogether, these data suggest that the pharmacological modulation of Rho GTPases by CNF1 can improve memory performances in an animal model of Alzheimer's disease via a control of neuroinflammation and a rescue of systemic energy homeostasis.
| S-EPMC3667817 | biostudies-literature