Dataset Information

Bayesian model reveals latent atrophy factors with dissociable cognitive trajectories in Alzheimer's disease.

ABSTRACT: We used a data-driven Bayesian model to automatically identify distinct latent factors of overlapping atrophy patterns from voxelwise structural MRIs of late-onset Alzheimer's disease (AD) dementia patients. Our approach estimated the extent to which multiple distinct atrophy patterns were expressed within each participant rather than assuming that each participant expressed a single atrophy factor. The model revealed a temporal atrophy factor (medial temporal cortex, hippocampus, and amygdala), a subcortical atrophy factor (striatum, thalamus, and cerebellum), and a cortical atrophy factor (frontal, parietal, lateral temporal, and lateral occipital cortices). To explore the influence of each factor in early AD, atrophy factor compositions were inferred in beta-amyloid-positive (A?+) mild cognitively impaired (MCI) and cognitively normal (CN) participants. All three factors were associated with memory decline across the entire clinical spectrum, whereas the cortical factor was associated with executive function decline in A?+ MCI participants and AD dementia patients. Direct comparison between factors revealed that the temporal factor showed the strongest association with memory, whereas the cortical factor showed the strongest association with executive function. The subcortical factor was associated with the slowest decline for both memory and executive function compared with temporal and cortical factors. These results suggest that distinct patterns of atrophy influence decline across different cognitive domains. Quantification of this heterogeneity may enable the computation of individual-level predictions relevant for disease monitoring and customized therapies. Factor compositions of participants and code used in this article are publicly available for future research.

SUBMITTER: Zhang X

PROVIDER: S-EPMC5081632 | biostudies-literature | 2016 Oct

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Bayesian model reveals latent atrophy factors with dissociable cognitive trajectories in Alzheimer's disease.

Zhang Xiuming X Mormino Elizabeth C EC Sun Nanbo N Sperling Reisa A RA Sabuncu Mert R MR Yeo B T Thomas BT

Proceedings of the National Academy of Sciences of the United States of America 20161004 42

We used a data-driven Bayesian model to automatically identify distinct latent factors of overlapping atrophy patterns from voxelwise structural MRIs of late-onset Alzheimer's disease (AD) dementia patients. Our approach estimated the extent to which multiple distinct atrophy patterns were expressed within each participant rather than assuming that each participant expressed a single atrophy factor. The model revealed a temporal atrophy factor (medial temporal cortex, hippocampus, and amygdala), ...[more]

PMID: 27702899

Similar Datasets

Project description:ObjectiveTo determine whether atrophy relates to phenotypical variants of posterior cortical atrophy (PCA) recently proposed in clinical criteria (i.e., dorsal, ventral, dominant-parietal, and caudal) we assessed associations between latent atrophy factors and cognition.MethodsWe employed a data-driven Bayesian modeling framework based on latent Dirichlet allocation to identify latent atrophy factors in a multicenter cohort of 119 individuals with PCA (age 64 ± 7 years, 38% male, Mini-Mental State Examination 21 ± 5, 71% β-amyloid positive, 29% β-amyloid status unknown). The model uses standardized gray matter density images as input (adjusted for age, sex, intracranial volume, MRI scanner field strength, and whole-brain gray matter volume) and provides voxelwise probabilistic maps for a predetermined number of atrophy factors, allowing every individual to express each factor to a degree without a priori classification. Individual factor expressions were correlated to 4 PCA-specific cognitive domains (object perception, space perception, nonvisual/parietal functions, and primary visual processing) using general linear models.ResultsThe model revealed 4 distinct yet partially overlapping atrophy factors: right-dorsal, right-ventral, left-ventral, and limbic. We found that object perception and primary visual processing were associated with atrophy that predominantly reflects the right-ventral factor. Furthermore, space perception was associated with atrophy that predominantly represents the right-dorsal and right-ventral factors. However, individual participant profiles revealed that the large majority expressed multiple atrophy factors and had mixed clinical profiles with impairments across multiple domains, rather than displaying a discrete clinical-radiologic phenotype.ConclusionOur results indicate that specific brain behavior networks are vulnerable in PCA, but most individuals display a constellation of affected brain regions and symptoms, indicating that classification into 4 mutually exclusive variants is unlikely to be clinically useful.

Project description:See O'Sullivan and Vann (doi:10.1093/aww166) for a scientific commentary on this article.About 15% of patients clinically diagnosed with Alzheimer's disease do not show high tracer retention on amyloid positon emission tomography imaging. The present study investigates clinical and demographic features, patterns of brain atrophy and hypometabolism and longitudinal clinical trajectories of these patients. Forty amyloid-negative patients carrying a pre-scan diagnosis of Alzheimer's disease dementia from four centres were included (11/29 females/males; mean age = 67 ± 9). Detailed clinical histories, including the clinical diagnoses before and after the amyloid scan and at follow-up, were collected. Patients were classified according to their pre-scan clinical phenotype as amnestic (memory predominant), non-amnestic (predominant language, visuospatial or frontal symptoms), or non-specific (diffuse cognitive deficits). Demographic, clinical, neuropsychological, magnetic resonance imaging and (18)F-fluorodeoxyglucose positon emission tomography data were compared to 27 amyloid-positive typical Alzheimer's disease cases (14/13 females/males; mean age = 71 ± 10) and 29 amyloid-negative controls (15/14 females/males; mean age = 69 ± 12) matched for age, gender and education. There were 21 amnestic, 12 non-amnestic, and seven non-specific amyloid-negative Alzheimer's disease cases. Amyloid-negative subgroups did not differ in age, gender or education. After the amyloid scan, clinicians altered the diagnosis in 68% of amyloid-negative patients including 48% of amnestic versus 94% of non-amnestic and non-specific cases. Amnestic amyloid-negative cases were most often reclassified as frontotemporal dementia, non-amnestic as frontotemporal dementia or corticobasal degeneration, and non-specific as dementia with Lewy bodies or unknown diagnosis. The longer-term clinical follow-up was consistent with the post-scan diagnosis in most cases (90%), including in amnestic amyloid-negative cases whose post-positon emission tomography diagnosis remained Alzheimer's disease. While the non-amnestic and non-specific amyloid-negative cases usually showed patterns of atrophy and hypometabolism suggestive of another degenerative disorder, the amnestic amyloid-negative cases had subtle atrophy and hypometabolism, restricted to the retrosplenial/posterior cingulate cortex. Patients with a negative amyloid positon emission tomography scan following an initial clinical diagnosis of Alzheimer's disease have heterogeneous clinical presentations and neuroimaging profiles; a majority showed a clinical progression that was consistent with a neurodegenerative condition. In contrast, in the subgroup of amnestic amyloid-negative cases, the clinical presentation and follow-up usually remained consistent with Alzheimer's disease. An alternative diagnosis was not made in about half of the amnestic amyloid-negative cases, highlighting the need for a clinical framework and terminology to define these patients, who may have underlying limbic-predominant, non-amyloid-related pathologies.

Project description:Alzheimer's disease (AD) is characterised by a dynamic process of neurocognitive changes from normal cognition to mild cognitive impairment (MCI) and progression to dementia. However, not all individuals with MCI develop dementia. Predicting whether individuals with MCI will decline (i.e. progressive MCI) or remain stable (i.e. stable MCI) is impeded by patient heterogeneity due to comorbidities that may lead to MCI diagnosis without progression to AD. Despite the importance of early diagnosis of AD for prognosis and personalised interventions, we still lack robust tools for predicting individual progression to dementia. Here, we propose a novel trajectory modelling approach based on metric learning (Generalised Metric Learning Vector Quantization) that mines multimodal data from MCI patients in the Alzheimer's disease Neuroimaging Initiative (ADNI) cohort to derive individualised prognostic scores of cognitive decline due to AD. We develop an integrated biomarker generation- using partial least squares regression- and classification methodology that extends beyond binary patient classification into discrete subgroups (i.e. stable vs. progressive MCI), determines individual profiles from baseline (i.e. cognitive or biological) data and predicts individual cognitive trajectories (i.e. change in memory scores from baseline). We demonstrate that a metric learning model trained on baseline cognitive data (memory, executive function, affective measurements) discriminates stable vs. progressive MCI individuals with high accuracy (81.4%), revealing an interaction between cognitive (memory, executive functions) and affective scores that may relate to MCI comorbidity (e.g. affective disturbance). Training the model to perform the same binary classification on biological data (mean cortical ?-amyloid burden, grey matter density, APOE 4) results in similar prediction accuracy (81.9%). Extending beyond binary classifications, we develop and implement a trajectory modelling approach that shows significantly better performance in predicting individualised rate of future cognitive decline (i.e. change in memory scores from baseline), when the metric learning model is trained with biological (r = -0.68) compared to cognitive (r = -0.4) data. Our trajectory modelling approach reveals interpretable and interoperable markers of progression to AD and has strong potential to guide effective stratification of individuals based on prognostic disease trajectories, reducing MCI patient misclassification, that is critical for clinical practice and discovery of personalised interventions.

Dataset Information

Bayesian model reveals latent atrophy factors with dissociable cognitive trajectories in Alzheimer's disease.

Publications

Bayesian model reveals latent atrophy factors with dissociable cognitive trajectories in Alzheimer's disease.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets