A5031763874- Intensive Care Unit Cognitive Disorders
- Machine Learning in Healthcare
- Plant and animal studies
- Neurobiology and Insect Physiology Research
- Insect Pheromone Research and Control
- Migraine and Headache Studies
- Cardiac Arrest and Resuscitation
- Cardiac, Anesthesia and Surgical Outcomes
- Anesthesia and Sedative Agents
- Respiratory Support and Mechanisms
- Transcranial Magnetic Stimulation Studies
- COVID-19 diagnosis using AI
- Time Series Analysis and Forecasting
- Artificial Intelligence in Healthcare
- Sepsis Diagnosis and Treatment
- Antiplatelet Therapy and Cardiovascular Diseases
- Family and Patient Care in Intensive Care Units
- Mechanical Circulatory Support Devices
- Functional Brain Connectivity Studies
- Rough Sets and Fuzzy Logic
- Acute Ischemic Stroke Management
- Age of Information Optimization
- Ion Transport and Channel Regulation
- stochastic dynamics and bifurcation
- Frailty in Older Adults
Charité - Universitätsmedizin Berlin
2021-2024
Humboldt-Universität zu Berlin
2021-2024
Freie Universität Berlin
2021-2024
Pfizer (Germany)
2024
Carnegie Mellon University
2021
Center for the Neural Basis of Cognition
2021
MIND Research Institute
2018-2021
Cornell University
2018-2021
University of Pittsburgh
2021
University of Auckland
2014-2016
Supporting the provision of intensive care medicine through telehealth potentially improves process quality. This may improve patient recovery and long-term outcomes. We investigated effectiveness a multifaceted telemedical programme on adherence to German quality indicators (QIs) in regional network units (ICUs) Germany. conducted an investigator-initiated, large-scale, open-label, stepped-wedge cluster randomised controlled trial enrolling adult ICU patients with expected stay ≥ 24 h....
Many species rely on olfaction to navigate towards food sources or mates. Olfactory navigation is a challenging task since odor environments are typically turbulent. While time-averaged concentration varies smoothly with the distance source, instaneous concentrations intermittent and obtaining stable averages takes longer than typical intervals between animals' decisions. How effectively sample from distribution determine sampling location focus in this article. To investigate which...
Background: Postoperative delirium (POD) is a critical and distressing condition, making identification of high-risk patients essential. Yet, complexity existing predictive models limits their practical use. This study evaluates fast-and-frugal decision trees (FFTrees)—simple binary prediction algorithms—and compares them to complex algorithms using routine clinical data from large German hospital.Objectives: To evaluate the utility FFTrees in predicting POD compare performance against...
Background: The prediction of the onset sepsis, a life-threatening condition resulting from dysregulated response to an infection, is one most common tasks in intensive care-related machine learning research. To assess performance such models, different evaluation strategies (fixed horizon, peak score and continuous evaluation) are commonly employed, but there no clear consensus on which approach should be used order provide clinically meaningful evaluation. Objective: approaches sepsis...
Postoperative delirium (POD) contributes to severe outcomes such as death or development of dementia. Thus, it is desirable identify vulnerable patients in advance during the perioperative phase. Previous studies mainly investigated risk factors for hospitalization and further used a linear logistic regression (LR) approach with time-invariant data. Studies have not patients’ fluctuating conditions support POD precautions. In this single-center study, we aimed predict recovery room setting...
The demands on a sensory system depend not only the statistics of its inputs but also task. In olfactory navigation, for example, task is to find plume source; allocation resources may therefore be driven by aspects that are informative about source location, rather than concentration per se. Here we explore implications this idea encoding odor concentration. To formalize notion limited, considered coding strategies partitioned range into set discriminable intervals. We developed dynamic...
Key points Agonist‐dependent oscillations in the concentration of free cytosolic calcium are a vital mechanism for control airway smooth muscle contraction and thus critical factor hyper‐responsiveness. Using mathematical model, closely tied to experimental work, we show that membrane potential accompanying have no significant effect on properties oscillations. In addition, model shows entry through store‐operated channels is oscillations, but voltage‐gated has much less effect. The predicts...
Stimulation protocols for medical devices should be rationally designed. For episodic migraine with aura we outline model-based design strategies towards preventive and acute therapies using stereotactic cortical neuromodulation. To this end, regard a localized spreading depression (SD) wave segment as central element in pathophysiology. describe nucleation propagation features of the SD segment, define new concepts hot spots labyrinths, respectively. In particular, firstly focus exclusively...
Association rules encode common patterns and structures identified in datasets. They can be derived by association rule mining (ARM) algorithms. The are human-readable allow comprehensible predictions, unlike many other types of prediction Classical ARM algorithms, like Apriori or FP-growth, cannot process interval ratio scaled data (quantitative variables) which limits their applicability. We address this restriction classical making it possible to quantitative variables on the right side a...
Mobilisation and exercise intervention in general are safe feasible critically ill patients. For patients requiring catecholamines, however, doses of norepinephrine for mobilisation the intensive care unit (ICU) not defined. This study aimed to describe practice our hospital identify that allowed a mobilisation.We conducted retrospective single-centre cohort 16 ICUs at university Germany with admitted between March 2018 November 2021. Data were collected from patient data management system....
Abstract Background Patients who experienced postoperative delirium (POD) are at higher risk of poor outcomes like dementia or death. Previous machine learning models predicting POD mostly relied on time-aggregated features. We aimed to assess the potential temporal patterns in clinical parameters during surgeries predict POD. Methods Long short-term memory (LSTM) and transformer models, directly consuming time series, were compared multi-layer perceptrons (MLPs) trained also fitted hybrid...
Artificial intelligence (AI) with its diverse domains such as expert systems and machine learning already has multiple potential applications in medicine. Based on the latest developments multifaceted field of AI, it will play a pivotal role medicine, high transformative areas, including drug development, diagnostics, patient care monitoring. In pharmaceutical industry AI is also rapidly gaining crucial role. The introduction innovative medicines requires profound background knowledge means...
(1) Background: Acute respiratory distress syndrome (ARDS) is a rare complication in multiply injured patients. Due to the rarity of ARDS development after trauma, little known about outcomes patients with trauma-associated compared non-trauma-associated ARDS. (2) Methods: This retrospective analysis included n = 1038 admitted center Charité-Universitätsmedizin Berlin between 2007 and 2018. Patients (n 62) were 976). In secondary analysis, from group 1:1 nearest neighbor matched The primary...
Abstract Machine learning (ML) holds great promise to support, improve, and automatize clinical decision-making in hospitals. Data protection regulations, however, hinder abundantly available routine data from being shared across sites for model training. Generative models can overcome this limitation by synthesize hospital a target population while ensuring privacy. Clinical time series acquired during intensive care are, difficult using established techniques, especially due uneven...
<title>Abstract</title> <bold>Motivation.</bold> Patients who experienced postoperative delirium (POD) are at higher risk of poor outcomes like dementia or death. Previous machine learning (ML) models predicting POD mostly relied on time-aggregated features. <bold>Objective.</bold> We assessed the potential temporal patterns in clinical parameters during surgeries to predict POD. <bold>Methods.</bold> Long short-term memory (LSTM) and transformer models, directly consuming time series, were...
Temporal missingness, defined as unobserved patterns in time series, and its predictive potentials represent an emerging area clinical machine learning. We trained a gated recurrent unit with decay mechanisms, called GRU-D, for binary classification between elderly - young patients. extracted series 5 vital signs from MIMIC-IV model inputs. GRU-D was evaluated means of 0.780 AUROC 0.810 AUPRC on bootstrapped data. Interpreting parameters, we found differences blood pressure missingness...
Animals use stereo sampling of odor concentration to localize sources and follow trails. We analyze the dynamics a bilateral model that depends on simultaneous comparison between concentrations detected by left right sensors. The general consists three differential equations for positions in plane heading. When landscape is an infinite trail, we reduce planar system whose has just two fixed points. Using integrable approximation (for short sensors) estimate basin attraction. In case radially...
Anticoagulation therapy with heparin is a frequent treatment in intensive care units and monitored by activated partial thromboplastin clotting time (aPTT). It has been demonstrated that reaching an established anticoagulation target within 24 hours associated favorable outcomes. However, patients respond to differently the can be challenging. Machine learning algorithms may potentially support clinicians improved dosing recommendations.This study evaluates range of machine on their...
Stimulation protocols for medical devices should be rationally designed. For episodic migraine with aura we outline model-based design strategies towards preventive and acute therapies using stereotactic neuromodulation. To this end, regard a localized spreading depression (SD) wave segment as central element in pathophysiology. describe nucleation propagation features of the SD segment, define new concepts cortical hot spots labyrinths, respectively. In particular, firstly focus exclusively...
Introduction: Left ventricular (LV) unloading with Impella has been shown to improve survival in patients cardiogenic shock treated veno-arterial extracorporeal membrane oxygenation (VA-ECMO). This study aimed investigate whether treatment VA-ECMO and (so called “ECMELLA”) during cardiopulmonary resuscitation (i.e., cardiac arrest due myocardial infarction no return of spontaneous circulation at device implantation) was associated improved outcomes. Hypothesis: ECMELLA is (primary outcome)...