A5009033153- Machine Learning in Healthcare
- Artificial Intelligence in Healthcare
- Traumatic Brain Injury Research
- Heart Failure Treatment and Management
- COVID-19 diagnosis using AI
- Explainable Artificial Intelligence (XAI)
- Sports injuries and prevention
- Cardiovascular Function and Risk Factors
- Lung Cancer Diagnosis and Treatment
- Heart Rate Variability and Autonomic Control
- Sepsis Diagnosis and Treatment
- Radiomics and Machine Learning in Medical Imaging
- Genetics, Bioinformatics, and Biomedical Research
- Infant Nutrition and Health
- Phonocardiography and Auscultation Techniques
- Cleft Lip and Palate Research
- Cardiovascular Effects of Exercise
- Ubiquitin and proteasome pathways
- Acute Ischemic Stroke Management
- Injury Epidemiology and Prevention
- Protein Degradation and Inhibitors
- Inflammatory Bowel Disease
- Artificial Intelligence in Healthcare and Education
- Cancer-related Molecular Pathways
- Non-Invasive Vital Sign Monitoring
University of Michigan
2019-2025
Michigan United
2023
Background & AimsEndoscopic assessment of ulcerative colitis (UC) typically reports only the maximum severity observed. Computer vision methods may better quantify mucosal injury detail, which varies among patients.MethodsEndoscopic video from UNIFI clinical trial (A Study to Evaluate Safety and Efficacy Ustekinumab Induction Maintenance Therapy in Participants With Moderately Severely Active Ulcerative Colitis) comparing ustekinumab placebo for UC were processed a computer analysis that...
Heart failure (HF) is a highly heterogeneous condition, and current methods struggle to synthesize extensive clinical data for personalized care. Using from 343 HF patients, we developed mechanistic computational models of the cardiovascular system create digital twins. These twins, consisting optimized measurable unmeasurable parameters alongside simulations function, provided comprehensive representations individual disease states. Unsupervised machine learning applied twin-derived...
The extensive adoption of artificial intelligence in clinical decision support systems requires greater model interpretability. Hence, we introduce EvolveFNN, an interpretable based on the recurrent neural network that merges fuzzy logic principles with units. This is designed to train precise and understandable models using high-dimensional longitudinal electronic health records data. Through supervised learning, our method allows identification variable encoding functions significant...
Abstract Timely and accurate referral of end-stage heart failure patients for advanced therapies, including transplants mechanical circulatory support, plays an important role in improving patient outcomes saving costs. However, the decision-making process is complex, nuanced, time-consuming, requiring cardiologists with specialized expertise training transplantation. In this study, we propose two logistic tensor regression-based models to predict warranting evaluation therapies using...
Pediatric respiratory disease diagnosis and subsequent treatment require accurate interpretable analysis. A chest X-ray is the most cost-effective rapid method for identifying monitoring various thoracic diseases in children. Recent developments self-supervised transfer learning have shown their potential medical imaging, including areas. In this article, we propose a three-stage framework with knowledge from adult X-rays to aid interpretation of pediatric thorax diseases. We conducted...
Unobtrusive collection of vital signs using sensors embedded in beds, chairs, and automobile seats can longitudinally monitor patients for abnormal heart conditions outside the hospital to inform both preventative post-diagnosis care. The capacitive electrocardiogram (cECG) shows potential collecting electrical information about a patient's without requiring skin contact like regular (ECG). However, motion artifacts environmental factors easily corrupt cECG signal quality reduce diagnostic...
Missing data presents a challenge for machine learning applications specifically when utilizing electronic health records to develop clinical decision support systems. The lack of these values is due in part the complex nature which content personalized each patient. Several methods have been developed handle this issue, such as imputation or complete case analysis, but their limitations restrict solidity findings. However, recent studies explored how using some features fully available...
Acute Respiratory Distress Syndrome (ARDS) is a life-threatening lung injury for which early diagnosis and evidence-based treatment can improve patient outcomes. Chest X-rays (CXRs) play crucial role in the identification of ARDS; however, their interpretation be difficult due to non-specific radiological features, uncertainty disease staging, inter-rater variability among clinical experts, thus leading prominent label noise issues. To address these challenges, this study proposes novel...
Over the past few years, artificial intelligence (AI) has emerged as a transformative force in drug discovery and development (DDD), revolutionizing many aspects of process. This survey provides comprehensive review recent advancements AI applications within early post-market assessment. It addresses identification prioritization new therapeutic targets, prediction drug-target interaction (DTI), design novel drug-like molecules, assessment clinical efficacy medications. By integrating...
Increased risk of musculoskeletal (MSK) injury post-concussion has been reported in collegiate athletes, yet it is unknown if professional football athletes are at the same secondary injury. The objective this study was to determine MSK National Football League (NFL) increases after concussion.NFL reports from 2013 2017 were collected public websites. Concussed (n=91) equally matched a non-injured control and an athlete with incident injury.Following their return sport, concussed 2.35 times...
Monitoring blood pressure, a parameter closely related to cardiovascular activity, can help predict imminent events. In this paper, novel method is proposed customize an existing mechanistic model of the system through feature extraction from cardiopulmonary acoustic signals estimate pressure using artificial intelligence. As various factors, such as drug consumption, alter biomechanical properties system, seeks personalize information extracted vibroacoustic sensors. Simulation results for...
Background: Acute myocardial infarctions are deadly to patients and burdensome healthcare systems. Most recorded patients’ first, occur out of the hospital, often not accompanied by cardiac comorbidities. The clinical manifestations underlying pathophysiology leading an infarction fully understood little effort exists use explainable machine learning learn predictive phenotypes before hospitalization is needed. Methods: We extracted outpatient electronic health record data for 2641 case 5287...
The advanced learning paradigm, using privileged information (LUPI), leverages in training that is not present at the time of prediction. In this study, we developed logistic regression (PLR) models under LUPI paradigm to detect acute respiratory distress syndrome (ARDS), with mechanical ventilation variables or chest x-ray image features employed domain and electronic health records base domain. model training, objective was designed incorporate data from encourage knowledge transfer across...
Summary Background Heart failure (HF) is a highly heterogeneous and complex condition. Although patient care generates vast amounts of clinical data, robust methods to synthesize available data for individualized management are lacking. Methods A mechanistic computational model cardiac cardiovascular system mechanics was identified each individual in cohort 343 patients with HF. The digital twins — comprising optimized sets parameters corresponding simulations function—for HF the used inform...
Chronic heart disease is a burdensome, complex, and fatal condition. Learning the mechanisms driving development of key to early risk assessment intervention. However, many current machine learning approaches lack sufficient interpretability. Using 2,737 patients with chronic from MIMIC-III database, we trained an interpretable Tropical Geometry Fuzzy Neural Network predict one-year occurrence severe cardiac procedure or mortality (AUROC=0.663). We present 20 learned rules which explain...
Survival analysis plays a pivotal role in healthcare, particularly analyzing time-to-event data such as disease progression, treatment efficacy, and drug development. Traditional methods survival often face trade-off: they either make linear assumptions, which are interpretable but may be overly simplistic, or capture complex, non-linear relationships lack clarity ease of understanding. To overcome these challenges, we develop novel machine-learning approach. This method is an extension the...
Inflammatory Bowel Disease (IBD) is a widespread gastrointestinal disorder characterized by inflammation and ulceration, affecting millions of individuals. Mucosal injury assessment through colonoscopy plays crucial role in determining disease severity therapeutic efficacy IBD. This study draws inspiration from the Segment Anything Model (SAM) framework tailors it for ulcer segmentation using Parameter-Efficient Transfer Learning (PETL) approach incorporating binary cross-entropy loss with...
The lack of a publicly accessible abdominal X-ray (AXR) dataset has hindered necrotizing enterocolitis (NEC) research. While significant strides have been made in applying natural language processing (NLP) to radiology reports, most efforts focused on chest radiology. Development an accurate NLP model identify features NEC radiograph can support improve diagnostic accuracy for this and other rare pediatric conditions. This study aims develop privacy-preserving large models (LLMs) their...
Timely referral for advanced therapies (i.e., heart transplantation, left ventricular assist device) is critical ensuring optimal outcomes failure patients. Using electronic health records, our goal was to use data from a single hospitalization develop an interpretable clinical decision-making system predicting the need at subsequent hospitalization.Michigan Medicine patients 2013-2021 with ejection fraction ≤ 35% and least two hospitalizations within one year were used train machine...
Abstract The extensive adoption of artificial intelligence in clinical decision support systems necessitates a significant presence ML models that clinicians can easily interpret. Therefore, we developed an RNN-based interpretable method, combining the fuzzy concepts and recurrent units, to train accurate explainable on high-dimensional longitudinal electronic health records data. Through supervised learning, our method allows identification variable encoding functions rules. To demonstrate...
Evaluation of the Rate Orthopedic Injuries Concussed and Non-Concussed Players in NFL Concussions may increase risk musculoskeletal injury during 90 day period after return to play. Previous work has evaluated this effect collegiate players with consistent results. PURPOSE: To examine possible increased orthopedic among National Football League 12 weeks (90 days) play from an incident concussion compared injury. METHODS: Weekly data 2012 through 2017 was collected public websites 3-10...
Following a concussion, athletes typically display increased reaction time, compensatory gait mechanics, and altered postural control. Although commonly studied in collegiate athletics, there is lack the literature regarding post-concussive effects National Basketball Association (NBA), how they affect player performance outcomes. PURPOSE: The purpose of this study was to determine if NBA statistics were different 28 days post-concussion compared their pre-concussion metrics. METHODS:...