Explore the vast range of titles available.

How do infants extract milk during breast-feeding? We have resolved a century-long scientific controversy, whether it is sucking of the milk by subatmospheric pressure or mouthing of the nipple–areola complex to induce a peristaltic-like extraction mechanism. Breast-feeding is a dynamic process, which requires coupling between periodic motions of the infant’s jaws, undulation of the tongue, and the breast milk ejection reflex. The physical mechanisms executed by the infant have been intriguing topics. We used an objective and dynamic analysis of ultrasound (US) movie clips acquired during breast-feeding to explore the tongue dynamic characteristics. Then, we developed a new 3D biophysical model of the breast and lactiferous tubes that enables the mimicking of dynamic characteristics observed in US imaging during breast-feeding, and thereby, exploration of the biomechanical aspects of breast-feeding. We have shown, for the first time to our knowledge, that latch-on to draw the nipple–areola complex into the infant mouth, as well as milk extraction during breast-feeding, require development of time-varying subatmospheric pressures within the infant’s oral cavity. Analysis of the US movies clearly demonstrated that tongue motility during breast-feeding was fairly periodic. The anterior tongue, which is wedged between the nipple–areola complex and the lower lips, moves as a rigid body with the cycling motion of the mandible, while the posterior section of the tongue undulates in a pattern similar to a propagating peristaltic wave, which is essential for swallowing.

Smoking-related microvascular loss causes end-organ damage in the kidneys, heart, and brain. Basic research suggests a similar process in the lungs, but no large studies have assessed pulmonary microvascular blood flow (PMBF) in early chronic lung disease. To investigate whether PMBF is reduced in mild as well as more severe chronic obstructive pulmonary disease (COPD) and emphysema. PMBF was measured using gadolinium-enhanced magnetic resonance imaging (MRI) among smokers with COPD and control subjects age 50 to 79 years without clinical cardiovascular disease. COPD severity was defined by standard criteria. Emphysema on computed tomography (CT) was defined by the percentage of lung regions below -950 Hounsfield units (-950 HU) and by radiologists using a standard protocol. We adjusted for potential confounders, including smoking, oxygenation, and left ventricular cardiac output. Among 144 participants, PMBF was reduced by 30% in mild COPD, by 29% in moderate COPD, and by 52% in severe COPD (all P < 0.01 vs. control subjects). PMBF was reduced with greater percentage emphysema-950HU and radiologist-defined emphysema, particularly panlobular and centrilobular emphysema (all P ≤ 0.01). Registration of MRI and CT images revealed that PMBF was reduced in mild COPD in both nonemphysematous and emphysematous lung regions. Associations for PMBF were independent of measures of small airways disease on CT and gas trapping largely because emphysema and small airways disease occurred in different smokers. PMBF was reduced in mild COPD, including in regions of lung without frank emphysema, and may represent a distinct pathological process from small airways disease. PMBF may provide an imaging biomarker for therapeutic strategies targeting the pulmonary microvasculature.

Background Automated machine-learning systems are able to de-identify electronic medical records, including free-text clinical notes. Use of such systems would greatly boost the amount of data available to researchers, yet their deployment has been limited due to uncertainty about their performance when applied to new datasets. Objective We present practical options for clinical note de-identification, assessing performance of machine learning systems ranging from off-the-shelf to fully customized. Methods We implement a state-of-the-art machine learning de-identification system, training and testing on pairs of datasets that match the deployment scenarios. We use clinical notes from two i2b2 competition corpora, the Physionet Gold Standard corpus, and parts of the MIMIC-III dataset. Results Fully customized systems remove 97–99% of personally identifying information. Performance of off-the-shelf systems varies by dataset, with performance mostly above 90%. Providing a small labeled dataset or large unlabeled dataset allows for fine-tuning that improves performance over off-the-shelf systems. Conclusion Health organizations should be aware of the levels of customization available when selecting a de-identification deployment solution, in order to choose the one that best matches their resources and target performance level.

Background Measuring vital signs plays a key role in both patient care and wellness, but can be challenging outside of medical settings due to the lack of specialized equipment. Methods In this study, we prospectively evaluated smartphone camera-based techniques for measuring heart rate (HR) and respiratory rate (RR) for consumer wellness use. HR was measured by placing the finger over the rear-facing camera, while RR was measured via a video of the participants sitting still in front of the front-facing camera. Results In the HR study of 95 participants (with a protocol that included both measurements at rest and post exercise), the mean absolute percent error (MAPE) ± standard deviation of the measurement was 1.6% ± 4.3%, which was significantly lower than the pre-specified goal of 5%. No significant differences in the MAPE were present across colorimeter-measured skin-tone subgroups: 1.8% ± 4.5% for very light to intermediate, 1.3% ± 3.3% for tan and brown, and 1.8% ± 4.9% for dark. In the RR study of 50 participants, the mean absolute error (MAE) was 0.78 ± 0.61 breaths/min, which was significantly lower than the pre-specified goal of 3 breaths/min. The MAE was low in both healthy participants (0.70 ± 0.67 breaths/min), and participants with chronic respiratory conditions (0.80 ± 0.60 breaths/min). Conclusions These results validate the accuracy of our smartphone camera-based techniques to measure HR and RR across a range of pre-defined subgroups. Plain language summary Accurate measurement of the number of times a heart beats per minute (heart rate, HR) and the number of breaths taken per minute (respiratory rate, RR) is usually undertaken using specialized equipment or training. We evaluated whether smartphone cameras could be used to measure HR and RR. We tested the accuracy of two computational approaches that determined HR and RR from the videos obtained using a smartphone. Changes in blood flow through the finger were used to determine HR; similar results were seen for people with different skin tones. Chest movements were used to determine RR; similar results were seen between people with and without chronic lung conditions. This study demonstrates that smartphones can be used to measure HR and RR accurately. Bae et al. prospectively evaluated smartphone camera-based techniques for measuring heart rate and respiratory rate. They found measurements were accurate across a range of pre-defined subgroups.

The application of machine learning and artificial intelligence to clinical settings for prevention, diagnosis, treatment, and the improvement of clinical care have been demonstrably cost-effective. However, current clinical AI (cAI) support tools are predominantly created by non-domain experts and algorithms available in the market have been criticized for the lack of transparency behind their creation. To combat these challenges, the Massachusetts Institute of Technology Critical Data (MIT-CD) consortium, an affiliation of research labs, organizations, and individuals that contribute to research in and around data that has a critical impact on human health, has iteratively developed the “Ecosystem as a Service (EaaS)” approach, providing a transparent education and accountability platform for clinical and technical experts to collaborate and advance cAI. The EaaS approach provides a range of resources, from open-source databases and specialized human resources to networking and collaborative opportunities. While mass deployment of the ecosystem still faces several hurdles, here we discuss our initial implementation efforts. We hope this will promote further exploration and expansion of the EaaS approach, while also informing or realizing policies that will accelerate multinational, multidisciplinary, and multisectoral collaborations in cAI research and development, and provide localized clinical best practices for equitable healthcare access.

The application of machine learning and artificial intelligence to clinical settings for prevention, diagnosis, treatment, and the improvement of clinical care have been demonstrably cost-effective. However, current clinical AI (cAI) support tools are predominantly created by non-domain experts and algorithms available in the market have been criticized for the lack of transparency behind their creation. To combat these challenges, the Massachusetts Institute of Technology Critical Data (MIT-CD) consortium, an affiliation of research labs, organizations, and individuals that contribute to research in and around data that has a critical impact on human health, has iteratively developed the “Ecosystem as a Service (EaaS)” approach, providing a transparent education and accountability platform for clinical and technical experts to collaborate and advance cAI. The EaaS approach provides a range of resources, from open-source databases and specialized human resources to networking and collaborative opportunities. While mass deployment of the ecosystem still faces several hurdles, here we discuss our initial implementation efforts. We hope this will promote further exploration and expansion of the EaaS approach, while also informing or realizing policies that will accelerate multinational, multidisciplinary, and multisectoral collaborations in cAI research and development, and provide localized clinical best practices for equitable healthcare access. Author summary The development of effective and scalable algorithms that are executable at the frontlines of clinical care require robust data infrastructure, data sharing and usage capabilities that do not compromise patient confidentiality, and flexibility in creating partnerships across various departments, industries, and sectors. To achieve this in a sustainable and replicable manner, the EaaS approach has three components; 1. A global coalition of AI clinicians accessible via the MIT-CD network to support the translation, customization, innovation, and external validations of algorithms, 2. Training opportunities provided via MIT-CD affiliated institutions to educate clinically informed engineers and data literate clinicians, and introduce them to open source databases, and 3. Networking opportunities and events offered by the MIT-CD consortium to share best practices and offer chances for collaboration on innovative research in multidisciplinary teams.

In recent years, multi-resolution techniques have become increasingly popular in the image processing community. New techniques have been developed with applications ranging from edge detection, texture recognition, image registration, multi-resolution features for image classification and more. The central focus of this two-part thesis is the multi-resolution analysis of images. In the first part, we used multi-resolution approaches to help with the classification of a set of protein crystal images. In the second, similar approaches were used to help register a set of 3D image volumes that would otherwise be computationally prohibitive without leveraging multi-resolution techniques. Specifically, the first part of this work proposes a classification framework that is being developed in collaboration with NorthEast Structural Genomics Consoritum (NESG) to assist in the automated screening of protein crystal images. Several groups have previously proposed automated algorithms to expedite such analysis. However, none of the classifiers described in the literature are sufficiently accurate or fast enough to be practical in a structural genomics production pipeline. The proposed classification algorithm uses random window sampling of the regions of interest to then compute several texture and multi-resolution image descriptor features that are subsequently processed through a random forest classifier. The resulting binary classifier exceeds 90% in sensitivity and 94% in specificity. Furthermore, the classifier is able to process each image with off-the-shelf computer components at approximately 7 seconds for each image, a speed that makes this algorithm usable in high throughput settings. The second part of this work proposes a 3D image registration algorithm to register regions of emphysema as quantified by densitometry on lung CT with MR lung volumes. The ability to register quantitatively-determined regions of emphysema with perfusion MRI will allow for further exploration of the pathophysiology of Chronic Obstructive Pulmonary Disorder (COPD). The registration method involves the registration of CT volumes at different levels of inspiration (total lung capacity to functional residual capacity [FRC]) followed by another registration between FRC-CT and FRC-MR volume pairs. We propose a registration method based on a combination of cubic b-spline registrations that is relatively quick (~4.5 minutes) and accurate (~6.3%). The methods presented in this work are being used to explore the relationships between regions of emphysema and their pulmonary microvascular blood flow during longitudinal progression of COPD.

Background: Remote patient monitoring (RPM) generates vast data, yet landmark trials (Tele-HF, BEAT-HF) failed because data volume overwhelmed clinical staff. While TIM-HF2 showed 24/7 physician-led monitoring reduces mortality by 30%, this model remains prohibitively expensive and unscalable. Methods: We developed Sentinel, an autonomous AI agent using Model Context Protocol (MCP) for contextual triage of RPM vitals via 21 clinical tools and multi-step reasoning. Evaluation included: (1) self-consistency (100 readings x 5 runs); (2) comparison against rule-based thresholds; and (3) validation against 6 clinicians (3 physicians, 3 NPs) using a connected matrix design. A leave-one-out (LOO) analysis compared the agent against individual clinicians; severe overtriage cases underwent independent physician adjudication. Results: Against a human majority-vote standard (N=467), the agent achieved 95.8% emergency sensitivity and 88.5% sensitivity for all actionable alerts (85.7% specificity). Four-level exact accuracy was 69.4% (quadratic-weighted kappa=0.778); 95.9% of classifications were within one severity level. In LOO analysis, the agent outperformed every clinician in emergency sensitivity (97.5% vs. 60.0% aggregate) and actionable sensitivity (90.9% vs. 69.5%). While disagreements skewed toward overtriage (22.5%), independent adjudication of severe gaps (>=2 levels) validated agent escalation in 88-94% of cases; consensus resolution validated 100%. The agent showed near-perfect self-consistency (kappa=0.850). Median cost was $0.34/triage. Conclusions: Sentinel triages RPM vitals with sensitivity exceeding individual clinicians. By automating systematic context synthesis, Sentinel addresses the core limitation of prior RPM trials, offering a scalable path toward the intensive monitoring shown to reduce mortality while maintaining a clinically defensible overtriage profile.

Seizure detection from EEGs is a challenging and time consuming clinical problem that would benefit from the development of automated algorithms. EEGs can be viewed as structural time series, because they are multivariate time series where the placement of leads on a patient's scalp provides prior information about the structure of interactions. Commonly used deep learning models for time series don't offer a way to leverage structural information, but this would be desirable in a model for structural time series. To address this challenge, we propose the temporal graph convolutional network (TGCN), a model that leverages structural information and has relatively few parameters. TGCNs apply feature extraction operations that are localized and shared over both time and space, thereby providing a useful inductive bias in tasks where one expects similar features to be discriminative across the different sequences. In our experiments we focus on metrics that are most important to seizure detection, and demonstrate that TGCN matches the performance of related models that have been shown to be state of the art in other tasks. Additionally, we investigate interpretability advantages of TGCN by exploring approaches for helping clinicians determine when precisely seizures occur, and the parts of the brain that are most involved.

Ultrasensitive antigen recognition between T lymphocytes and cognate targets via immunological synapse (IS) formation enables live cell-based antigen-specific T cell detection. However, unpredictable antigen processing and major histocompatibility complex (MHC) turnover limit specificity. Here, intracellularly polymerized antigen-presenting cells (pAPCs) are developed for modular, persistent antigen display via kinetically driven loading. Although inanimate, pAPCs mimic cellular interactions, inducing IS hallmarks such as supramolecular activation cluster formation, cytoskeletal contraction, and trogocytosis. Incorporation of superparamagnetic nanoparticles allows label-free magnetic isolation of antigen-specific T cells, surpassing MHC-conjugated beads in sensitivity and specificity. In tumor-bearing hosts, pAPCs enrich tumor-reactive lymphocytes, enhancing adoptive T cell therapy and neoantigen-specific T cell identification. Additionally, pAPCs from engineered cells expressing monovalent human MHC enrich virus- and tumor-specific CD8 T cells from human peripheral blood mononuclear cells and human leukocyte antigen-transgenic mice, demonstrating the potential of this cell–gel hybrid platform for precise antigen-specific T cell capture. T-cell enrichment has attracted great interest but currently fails to fully replicate the complex natural immune interactions. Here, the authors report on magnetic polymerised antigen-presenting cells that mimic natural interactions to isolate rare tumour-reactive T cells, offering a platform for enhancing cancer immunotherapy and neoantigen discovery.