A5074984657- Autism Spectrum Disorder Research
- Genetics and Neurodevelopmental Disorders
- Machine Learning in Healthcare
- Genetic Associations and Epidemiology
- Mental Health Treatment and Access
- Artificial Intelligence in Healthcare and Education
- ECG Monitoring and Analysis
- Mental Health via Writing
- Topic Modeling
- Bayesian Methods and Mixture Models
- Neurological disorders and treatments
- Gene expression and cancer classification
- Text Readability and Simplification
- Genomics and Rare Diseases
- Congenital heart defects research
- Acute Ischemic Stroke Management
- Child Development and Digital Technology
- Attachment and Relationship Dynamics
- Advanced Clustering Algorithms Research
- Chronic Disease Management Strategies
- Cardiovascular Function and Risk Factors
- Heart Rate Variability and Autonomic Control
- Acute Myocardial Infarction Research
- Child and Adolescent Psychosocial and Emotional Development
- Family and Disability Support Research
Icahn School of Medicine at Mount Sinai
2021-2024
Center for Neuroscience and Cognitive Systems
2020-2024
Italian Institute of Technology
2020-2024
University of Trento
2019-2022
Next Generation Infrastructures
2022
Mount Sinai Health System
2021
Hasso Plattner Institute
2021
Fondazione Bruno Kessler
2019-2020
Deriving disease subtypes from electronic health records (EHRs) can guide next-generation personalized medicine. However, challenges in summarizing and representing patient data prevent widespread practice of scalable EHR-based stratification analysis. Here we present an unsupervised framework based on deep learning to process heterogeneous EHRs derive representations that efficiently effectively enable at scale. We considered 1,608,741 patients a diverse hospital cohort comprising total...
This study sought to develop DL models capable of comprehensively quantifying left and right ventricular dysfunction from ECG data in a large, diverse population. Rapid evaluation function using deep learning (DL) on electrocardiograms (ECGs) can assist diagnostic workflow. However, tools estimate (RV) do not exist, whereas those (LV) are restricted quantification very low LV only. A multicenter was conducted with 5 New York City hospitals: 4 for internal testing 1 serving as external...
Musculoskeletal disorders like lower back, knee, and shoulder pain create a substantial health burden in developed countries—affecting function, mobility, quality of life1Bhattacharya A Costs occupational musculoskeletal (MSDs) the United States.Int J Ind Ergon. 2014; 44: 448-454Crossref Scopus (120) Google Scholar. These conditions are often multifactorial require clinicians to thoroughly assess their cause decide on appropriate investigative treatment approaches. However, electronic...
Abstract A major goal of precision medicine is to predict prognosis based on individualized information at the earliest possible points in development. Using early snapshots adaptive functioning and unsupervised data-driven discovery methods, we uncover highly stable autism subtypes that yield relevant later prognosis. Data from National Institute Mental Health Archive (NDA) ( n = 1,098) was used three (<72 months) generalize with 96% accuracy. Outcome data NDA 2,561; mean age, 13 years)...
Abstract Objectives Social support (SS) and social isolation (SI) are determinants of health (SDOH) associated with psychiatric outcomes. In electronic records (EHRs), individual-level SS/SI is typically documented in narrative clinical notes rather than as structured coded data. Natural language processing (NLP) algorithms can automate the otherwise labor-intensive process extraction such information. Materials Methods Psychiatric encounter from Mount Sinai Health System (MSHS, n = 300)...
Abstract Social-communication (SC) and restricted repetitive behaviors (RRB) are autism diagnostic symptom domains. SC RRB severity can markedly differ within between individuals may be underpinned by different neural circuitry genetic mechanisms. Modeling SC-RRB balance could help identify how mechanisms map onto such phenotypic heterogeneity. Here, we developed a stratification model that makes highly accurate (97–99%) out-of-sample = RRB, > subtype predictions. Applying this to resting...
Background: Evidence-based medicine (EBM) is fundamental to modern clinical practice, requiring clinicians continually update their knowledge and apply the best evidence in patient care. The practice of EBM faces challenges due rapid advancements medical research, leading information overload for clinicians. integration artificial intelligence (AI), specifically Generative Large Language Models (LLMs), offers a promising solution towards managing this complexity. Methods: This study involved...
Abstract Background Depression and anxiety are common highly comorbid, their comorbidity is associated with poorer outcomes posing clinical public health concerns. We evaluated the polygenic contribution to comorbid depression anxiety, each in isolation. Methods Diagnostic codes were extracted from electronic records for four biobanks [ N = 177 865 including 138 632 European (77.9%), 25 612 African (14.4%), 13 621 Hispanic (7.7%) ancestry participants]. The outcome was a four-level variable...
The recognition of emotional body movement (BM) is impaired in individuals with Autistic Spectrum Disorder ASD, yet it not clear whether the difficulty related to encoding motion, emotions, or both. Besides, BM has been traditionally studied using point-light displays stimuli (PLDs) and still underexplored ASD intellectual disability (ID). In present study, we investigated happy, fearful, neutral children without ID. a non-verbal task, participants were asked recognize pure-body-motion...
Abstract Aims Clinical scoring systems for pulmonary embolism (PE) screening have low specificity and contribute to computed tomography angiogram (CTPA) overuse. We assessed whether deep learning models using an existing routinely collected data modality, electrocardiogram (ECG) waveforms, can increase PE detection. Methods results create a retrospective cohort of 21 183 patients at moderate- high suspicion associate 23 793 CTPAs (10.0% PE-positive) with 320 746 ECGs encounter-level clinical...
<b>Background</b><br /> Synchrony is an essential component of interactive exchanges. In mother-infant interaction, synchrony underlies reciprocity and emotive regulation. A severe lack indeed a core issue within the communication interaction deficit that characterizes autism spectrum disorders (ASD) in accordance with DSM-5 classification. Based on emerging evidence music therapy can improve regulation ability children ASD, we aim to verify quantitatively whether: 1) ASD...
Background: Social support (SS) and social isolation (SI) are determinants of health (SDOH) associated with psychiatric outcomes. In electronic records (EHRs), individual-level SS/SI is typically documented as narrative clinical notes rather than structured coded data. Natural language processing (NLP) algorithms can automate the otherwise labor-intensive process data extraction. Data Methods: Psychiatric encounter from Mount Sinai Health System (MSHS, n=300) Weill Cornell Medicine (WCM,...
Determining the best partition for a dataset can be challenging task because of 1) lack priori information within an unsupervised learning framework; and 2) absence unique clustering validation approach to evaluate solutions. Here we present reval: Python package that leverages stability-based relative methods determine solutions as ones generalize unseen data. Statistical software, both in R Python, usually rely on internal metrics, such silhouette, select number clusters fits Meanwhile,...
Abstract Early detection and intervention are believed to be key facilitating better outcomes in children with autism, yet the impact of age at treatment start on outcome is poorly understood. While clinical traits such as language ability have been shown predict outcome, whether or not how information genomic level can unknown. Leveraging a cohort toddlers autism who all received same standardized very young provided blood sample, here we find that early engagement (i.e., <24 months)...
Abstract Background Motor difficulties are common in many, but not all, autistic individuals. These can co-occur with other problems, such as delays language, intellectual, and adaptive functioning. Biological mechanisms underpinning less well understood. Poor motor skills tend to be more individuals carrying highly penetrant rare genetic mutations. Such may have downstream consequences of altering neurophysiological excitation-inhibition balance lead enhanced behavioral noise. Methods This...
Abstract The impact of different parenting‐related variables on child psychological development is widely acknowledged. However, studies about the specific influence maternal and family dimensions early developmental outcomes in at‐risk dyads are still scarce. aim this longitudinal study was to investigate short‐ middle‐term effects prenatal postnatal features, attachment, at 3 24 months families. Forty‐two mothers with psychological, social and/or demographic risk conditions their...
Abstract Early motor difficulties are a common in many, but not all, autistic individuals. These tend to be highly present individuals carrying rare genetic mutations with high penetrance for autism. Many of these mechanisms also cause neurophysiological dysregulation excitation-inhibition balance (E:I). A predicted downstream consequence E:I imbalance circuitry would translate behaviorally into enhanced ‘motor noise’ – that is, increased variability execution actions. Here we tested the...
Convolutional Neural Networks (CNNs) are a popular deep learning architecture widely applied in different domains, particular classifying over images, for which the concept of convolution with filter comes naturally. Unfortunately, requirement distance (or, at least, neighbourhood function) input feature space has so far prevented its direct use on data types such as omics data. However, number metrizable, i.e., they can be endowed metric structure, enabling to adopt convolutional based...
Clinical note classification is a common clinical NLP task. However, annotated data-sets are scarse. Prompt-based learning has recently emerged as an effective method to adapt pre-trained models for text using only few training examples. A critical component of prompt design the definition template (i.e. text). The effect position, however, been insufficiently investigated. This seems particularly important in setting, where task-relevant information usually sparse notes. In this study we...
Despite being a unique source of information on patients' status and disease progression, clinical notes are characterized by high levels duplication redundancy. In general domain text, it has been shown that deduplication does not harm language model (LM) pretraining, thus helping reduce the training cost. Although large LMs have proven to learn medical knowledge, they still require specialized adaptation for improved downstream tasks. By leveraging real-world corpora, we first provided...