Unsupervised data-driven stratification of mentalizing heterogeneity in autism.
Ontology highlight
ABSTRACT:
SUBMITTER: Lombardo MV
PROVIDER: S-EPMC5067562 | biostudies-other | 2016 Oct
REPOSITORIES: biostudies-other
Publications
Unsupervised data-driven stratification of mentalizing heterogeneity in autism.
Lombardo Michael V MV Lai Meng-Chuan MC Auyeung Bonnie B Holt Rosemary J RJ Allison Carrie C Smith Paula P Chakrabarti Bhismadev B Ruigrok Amber N V AN Suckling John J Bullmore Edward T ET Ecker Christine C Craig Michael C MC Murphy Declan G M DG Happé Francesca F Baron-Cohen Simon S
Scientific reports 20161018
Individuals affected by autism spectrum conditions (ASC) are considerably heterogeneous. Novel approaches are needed to parse this heterogeneity to enhance precision in clinical and translational research. Applying a clustering approach taken from genomics and systems biology on two large independent cognitive datasets of adults with and without ASC (n = 694; n = 249), we find replicable evidence for 5 discrete ASC subgroups that are highly differentiated in item-level performance on an explicit ...[more]
Similar Datasets
Project description:The widespread adoption of online courses opens opportunities for analysing learner behaviour and optimising web-based learning adapted to observed usage. Here, we introduce a mathematical framework for the analysis of time-series of online learner engagement, which allows the identification of clusters of learners with similar online temporal behaviour directly from the raw data without prescribing a priori subjective reference behaviours. The method uses a dynamic time warping kernel to create a pair-wise similarity between time-series of learner actions, and combines it with an unsupervised multiscale graph clustering algorithm to identify groups of learners with similar temporal behaviour. To showcase our approach, we analyse task completion data from a cohort of learners taking an online post-graduate degree at Imperial Business School. Our analysis reveals clusters of learners with statistically distinct patterns of engagement, from distributed to massed learning, with different levels of regularity, adherence to pre-planned course structure and task completion. The approach also reveals outlier learners with highly sporadic behaviour. A posteriori comparison against student performance shows that, whereas high-performing learners are spread across clusters with diverse temporal engagement, low performers are located significantly in the massed learning cluster, and our unsupervised clustering identifies low performers more accurately than common machine learning classification methods trained on temporal statistics of the data. Finally, we test the applicability of the method by analysing two additional data sets: a different cohort of the same course, and time-series of different format from another university.
| S-EPMC6722089 | biostudies-literature
Project description:ObjectivesTo stratify patients with multiple sclerosis (pwMS) based on brain MRI-derived volumetric features using unsupervised machine learning.MethodsThe 3-T brain MRIs of relapsing-remitting pwMS including 3D-T1w and FLAIR-T2w sequences were retrospectively collected, along with Expanded Disability Status Scale (EDSS) scores and long-term (10 ± 2 years) clinical outcomes (EDSS, cognition, and progressive course). From the MRIs, volumes of demyelinating lesions and 116 atlas-defined gray matter regions were automatically segmented and expressed as z-scores referenced to external populations. Following feature selection, baseline MRI-derived biomarkers entered the Subtype and Stage Inference (SuStaIn) algorithm, which estimates subgroups characterized by distinct patterns of biomarker evolution and stages within subgroups. The trained model was then applied to longitudinal MRIs. Stability of subtypes and stage change over time were assessed via Krippendorf's α and multilevel linear regression models, respectively. The prognostic relevance of SuStaIn classification was assessed with ordinal/logistic regression analyses.ResultsWe selected 425 pwMS (35.9 ± 9.9 years; F/M: 301/124), corresponding to 1129 MRI scans, along with healthy controls (N = 148; 35.9 ± 13.0 years; F/M: 77/71) and external pwMS (N = 80; 40.4 ± 11.9 years; F/M: 56/24) as reference populations. Based on 11 biomarkers surviving feature selection, two subtypes were identified, designated as "deep gray matter (DGM)-first" subtype (N = 238) and "cortex-first" subtype (N = 187) according to the atrophy pattern. Subtypes were consistent over time (α = 0.806), with significant annual stage increase (b = 0.20; p < 0.001). EDSS was associated with stage and DGM-first subtype (p ≤ 0.02). Baseline stage predicted long-term disability, transition to progressive course, and cognitive impairment (p ≤ 0.03), with the latter also associated with DGM-first subtype (p = 0.005).ConclusionsUnsupervised learning modelling of brain MRI-derived volumetric features provides a biologically reliable and prognostically meaningful stratification of pwMS.Key points• The unsupervised modelling of brain MRI-derived volumetric features can provide a single-visit stratification of multiple sclerosis patients. • The so-obtained classification tends to be consistent over time and captures disease-related brain damage progression, supporting the biological reliability of the model. • Baseline stratification predicts long-term clinical disability, cognition, and transition to secondary progressive course.
| S-EPMC9279232 | biostudies-literature
Project description:BackgroundPhysiological motion, such as respiratory motion, has become a limiting factor in the spatial resolution of positron emission tomography (PET) imaging as the resolution of PET detectors continue to improve. Motion-induced misregistration between PET and CT images can also cause attenuation correction artifacts. Respiratory gating can be used to freeze the motion and to reduce motion induced artifacts.PurposeIn this study, we propose a robust data-driven approach using an unsupervised deep clustering network that employs an autoencoder (AE) to extract latent features for respiratory gating.MethodsWe first divide list-mode PET data into short-time frames. The short-time frame images are reconstructed without attenuation, scatter, or randoms correction to avoid attenuation mismatch artifacts and to reduce image reconstruction time. The deep AE is then trained using reconstructed short-time frame images to extract latent features for respiratory gating. No additional data are required for the AE training. K-means clustering is subsequently used to perform respiratory gating based on the latent features extracted by the deep AE. The effectiveness of our proposed Deep Clustering method was evaluated using physical phantom and real patient datasets. The performance was compared against phase gating based on an external signal (External) and image based principal component analysis (PCA) with K-means clustering (Image PCA).ResultsThe proposed method produced gated images with higher contrast and sharper myocardium boundaries than those obtained using the External gating method and Image PCA. Quantitatively, the gated images generated by the proposed Deep Clustering method showed larger center of mass (COM) displacement and higher lesion contrast than those obtained using the other two methods.ConclusionsThe effectiveness of our proposed method was validated using physical phantom and real patient data. The results showed our proposed framework could provide superior gating than the conventional External method and Image PCA.
| S-EPMC10592231 | biostudies-literature
Project description:BACKGROUND: Although many risk factors are well known, Clostridium difficile infection (CDI) continues to be a significant problem throughout the world. The purpose of this study was to develop and validate a data-driven, hospital-specific risk stratification procedure for estimating the probability that an inpatient will test positive for C difficile. METHODS: We consider electronic medical record (EMR) data from patients admitted for ?24 hours to a large urban hospital in the U.S. between April 2011 and April 2013. Predictive models were constructed using L2-regularized logistic regression and data from the first year. The number of observational variables considered varied from a small set of well known risk factors readily available to a physician to over 10 000 variables automatically extracted from the EMR. Each model was evaluated on holdout admission data from the following year. A total of 34 846 admissions with 372 cases of CDI was used to train the model. RESULTS: Applied to the separate validation set of 34 722 admissions with 355 cases of CDI, the model that made use of the additional EMR data yielded an area under the receiver operating characteristic curve (AUROC) of 0.81 (95% confidence interval [CI], .79-.83), and it significantly outperformed the model that considered only the small set of known clinical risk factors, AUROC of 0.71 (95% CI, .69-.75). CONCLUSIONS: Automated risk stratification of patients based on the contents of their EMRs can be used to accurately identify a high-risk population of patients. The proposed method holds promise for enabling the selective allocation of interventions aimed at reducing the rate of CDI.
| S-EPMC4281796 | biostudies-literature
Project description:BackgroundFew studies have answered the guiding significance of individual components of the Framingham risk score (FRS) to the risk of cardiovascular disease (CVD) after antihypertensive treatment. This study on the systolic blood pressure intervention trial (SPRINT) and the Action to Control Cardiovascular Risk in Diabetes blood pressure trial (ACCORD-BP) aimed to reveal previously undetected association patterns between individual components of the FRS and heterogeneity of treatment effects (HTEs) of intensive blood pressure control.MethodsA self-organizing map (SOM) methodology was applied to identify CVD-risk-specific subgroups in the SPRINT (n = 8,773), and the trained SOM was utilized directly in 4,495 patients from the ACCORD. The primary endpoints were myocardial infarction (MI), non-myocardial infarction acute coronary syndrome (non-MI ACS), stroke, heart failure (HF), death from CVD causes, and a primary composite cardiovascular outcome. Cox proportional hazards models were then used to explore the potential heterogeneous response to intensive SBP control.ResultsWe identified four SOM-based subgroups with distinct individual components of FRS profiles and the CVD risk. For individuals with type 2 diabetes mellitus (T2DM) in the ACCORD or without diabetes in the SPRINT, subgroup I characterized by male with the lowest concentrations for total cholesterol (TC) and high-density lipoprotein (HDL) cholesterol measures, experienced the highest risk for major CVD. Conversely, subgroup III characterized by a female with the highest values for these measures represented as the lowest CVD risk. Furthermore, subgroup II, with the highest systolic blood pressure (SBP) and no antihypertensive agent use at baseline, had a significantly greater frequency of non-MI ACS under intensive BP control, the number needed to harm (NNH) was 84.24 to cause 1 non-MI ACS [absolute risk reduction (ARR) = -1.19%; 95% CI: -2.08, -0.29%] in the SPRINT [hazard ratio (HR) = 3.62; 95% CI: 1.33, 9.81; P = 0.012], and the NNH of was 43.19 to cause 1 non-MI ACS (ARR = -2.32%; 95% CI: -4.63, 0.00%) in the ACCORD (HR = 1.81; 95% CI: 1.01-3.25; P = 0.046). Finally, subgroup IV characterized by mostly younger patients with antihypertensive medication use and smoking history represented the lowest risk for stroke, HF, and relatively low risk for death from CVD causes and primary composite CVD outcome in SPRINT, however, except stroke, a low risk for others were not observed in ACCORD.ConclusionSimilar findings in patients with hypertensive with T2DM or without diabetes by multivariate subgrouping suggested that the individual components of the FRS could enrich or improve CVD risk assessment. Further research was required to clarify the potential mechanism.
| S-EPMC8850629 | biostudies-literature
Project description:The reshuffling mobility of molecular building blocks in self-assembled micelles is a key determinant of many their interesting properties, from emerging morphologies and surface compartmentalization, to dynamic reconfigurability and stimuli-responsiveness. However, the microscopic details of such complex structural dynamics are typically nontrivial to elucidate, especially in multicomponent assemblies. Here we show a machine-learning approach that allows us to reconstruct the structural and dynamic complexity of mono- and bicomponent surfactant micelles from high-dimensional data extracted from equilibrium molecular dynamics simulations. Unsupervised clustering of smooth overlap of atomic position (SOAP) data enables us to identify, in a set of multicomponent surfactant micelles, the dominant local molecular environments that emerge within them and to retrace their dynamics, in terms of exchange probabilities and transition pathways of the constituent building blocks. Tested on a variety of micelles differing in size and in the chemical nature of the constitutive self-assembling units, this approach effectively recognizes the molecular motifs populating them in an exquisitely agnostic and unsupervised way, and allows correlating them to their composition in terms of constitutive surfactant species.
| S-EPMC10041528 | biostudies-literature
Project description:Most genome linkage scans for autism spectrum disorders (ASDs) have failed to be replicated. Recently, a new ASD phenotypic sub-classification method was developed which employed cluster analyses of severity scores from the Autism Diagnostic Interview-Revised (ADI-R). Here, we performed linkage analysis for each of the four identified ADI-R stratified subgroups. Additional stratification was also applied to reduce intra-family heterogeneity and to investigate the impact of gender. For the purpose of replication, two independent sets of single nucleotide polymorphism markers for 392 families were used in our study. This deep subject stratification protocol resulted in 16 distinct group-specific datasets for linkage analysis. No locus reached significance for the combined non-stratified cohort. However, study-wide significant (P = 0.02) linkage scores were reached for chromosomes 22q11 (LOD = 4.43) and 13q21 (LOD = 4.37) for two subsets representing the most severely language impaired individuals with ASD. Notably, 13q21 has been previously linked to autism with language impairment, and 22q11 has been separately associated with either autism or language disorders. Linkage analysis on chromosome 5p15 for a combination of two stratified female-containing subgroups demonstrated suggestive linkage (LOD = 3.5), which replicates previous linkage result for female-containing pedigrees. A trend was also found for the association of previously reported 5p14-p15 SNPs in the same female-containing cohort. This study demonstrates a novel and effective method to address the heterogeneity in genetic studies of ASD. Moreover, the linkage results for the stratified subgroups provide evidence at the gene scan level for both inter- and intra-family heterogeneity as well as for gender-specific loci.
| S-EPMC3694043 | biostudies-literature
Project description:BackgroundAutism spectrum disorder (ASD) is a neurodevelopmental condition with key deficits in social functioning. It is widely assumed that the biological underpinnings of social impairment are neurofunctional alterations in the "social brain," a neural circuitry involved in inferring the mental state of a social partner. However, previous evidence comes from small-scale studies and findings have been mixed. We therefore carried out the to-date largest study on neural correlates of mentalizing in ASD.MethodsAs part of the Longitudinal European Autism Project, we performed functional magnetic resonance imaging at six European sites in a large, well-powered, and deeply phenotyped sample of individuals with ASD (N = 205) and typically developing (TD) individuals (N = 189) aged 6 to 30 years. We presented an animated shapes task to assess and comprehensively characterize social brain activation during mentalizing. We tested for effects of age, diagnosis, and their association with symptom measures, including a continuous measure of autistic traits.ResultsWe observed robust effects of task. Within the ASD sample, autistic traits were moderately associated with functional activation in one of the key regions of the social brain, the dorsomedial prefrontal cortex. However, there were no significant effects of diagnosis on task performance and no effects of age and diagnosis on social brain responses. Besides a lack of mean group differences, our data provide no evidence for meaningful differences in the distribution of brain response measures. Extensive control analyses suggest that the lack of case-control differences was not due to a variety of potential confounders.ConclusionsContrary to prior reports, this large-scale study does not support the assumption that altered social brain activation during mentalizing forms a common neural marker of ASD, at least with the paradigm we employed. Yet, autistic individuals show socio-behavioral deficits. Our work therefore highlights the need to interrogate social brain function with other brain measures, such as connectivity and network-based approaches, using other paradigms, or applying complementary analysis approaches to assess individual differences in this heterogeneous condition.
| S-EPMC7036196 | biostudies-literature
Project description:MotivationIn the realm of precision medicine, effective patient stratification and disease subtyping demand innovative methodologies tailored for multi-omics data. Clustering techniques applied to multi-omics data have become instrumental in identifying distinct subgroups of patients, enabling a finer-grained understanding of disease variability. Meanwhile, clinical datasets are often small and must be aggregated from multiple hospitals. Online data sharing, however, is seen as a significant challenge due to privacy concerns, potentially impeding big data's role in medical advancements using machine learning. This work establishes a powerful framework for advancing precision medicine through unsupervised random forest-based clustering in combination with federated computing.ResultsWe introduce a novel multi-omics clustering approach utilizing unsupervised random forests. The unsupervised nature of the random forest enables the determination of cluster-specific feature importance, unraveling key molecular contributors to distinct patient groups. Our methodology is designed for federated execution, a crucial aspect in the medical domain where privacy concerns are paramount. We have validated our approach on machine learning benchmark datasets as well as on cancer data from The Cancer Genome Atlas. Our method is competitive with the state-of-the-art in terms of disease subtyping, but at the same time substantially improves the cluster interpretability. Experiments indicate that local clustering performance can be improved through federated computing.Availability and implementationThe proposed methods are available as an R-package (https://github.com/pievos101/uRF).
| S-EPMC11373406 | biostudies-literature
Project description:Ankle and foot orthoses are commonly prescribed to children with cerebral palsy (CP). It is unclear whether 3D gait analysis (3DGA) provides sufficient and reliable information for clinicians to be consistent when prescribing orthoses. Data-driven modeling can probe such questions by revealing non-intuitive relationships between variables such as 3DGA parameters and gait outcomes of orthoses use. The purpose of this study was to (1) develop a data-driven model to classify children with CP according to their gait biomechanics and (2) identify relationships between orthotics types and gait patterns. 3DGA data were acquired from walking trials of 25 typically developed children and 98 children with CP with additional prescribed orthoses. An unsupervised self-organizing map followed by k-means clustering was developed to group different gait patterns based on children's 3DGA. Model inputs were gait variable scores (GVSs) extracted from the gait profile score, measuring root mean square differences from TD children's gait cycle. The model identified five pathological gait patterns with statistical differences in GVSs. Only 43% of children improved their gait pattern when wearing an orthosis. Orthotics prescriptions were variable even in children with similar gait patterns. This study suggests that quantitative data-driven approaches may provide more clarity and specificity to support orthotics prescription.
| S-EPMC7290444 | biostudies-literature