Heather Couture, PhD’s Post

Exciting news for neuroimaging researchers! Check out this new blog post on ezBIDS, a user-friendly tool for standardizing and converting neuroimaging data to the BIDS format. With ezBIDS, you can easily convert your data and associated metadata without any coding or technical expertise. This tool offers automated guidance for adhering to the BIDS standard and supports both imaging and task events data. Plus, you have multiple options for managing your data, including downloading to your local system or transferring to OpenNeuro.org and brainlife.io. Don't miss out on this innovative solution for data interoperability. Read the full blog post here: https://bit.ly/40Bfrm7 [cs.DB] #neuroimaging #datastandardization

#latepost #research #engineering #biomedicalengineering #hrv #ieeexplore Glad to state that my research paper titled "Machine Learning Approach to predict Ventricular Tachycardia using Heart Rate Variability Features" is been published at IEEE Xplore. 10.1109/ICCINS58907.2023.10450074 Ventricular Tachycardia can be a major symptom of cardiac arrest!

👽How is this not the plot of a science fiction movie that your HEART 🫀generates an ELECTRIC POTENTIAL ⚡ that can be detected ON YOUR SKIN and then we can use that to tell if your heart's ARTERY IS OCCLUDED. But I think the even cooler thing is that we're starting to see new information from the good ol' 12-lead EKG that's been there the entire time; we just couldn't see it. The EKG itself hasn't changed. It's our ability to understand its data that has. I was excited to see Powerful Medical's paper with Robert Herman, MD and Stephen Smith and Pendell Meyers get published in this space, with their machine learning model doing a better job of detecting occluded coronaries than the antiquated STEMI criteria. (And if you're a biostats fan, an AUC of 0.938 is absolutely unheard of in decision support.) They're also reporting: • 58% fewer false positive cath lab activations • 30% reduction in door to balloon • Shorter hospital and ICU length of stay https://lnkd.in/gtfeWdAK

Amazing study.AUC of 0.938 is unheard of in decision support.

Advancements in biomedical image classification are breaking new ground with the integration of Geometric-Topological Analysis. A recent study reveals how combining Topological Data Analysis (TDA) and Lipschitz-Killing Curvatures (LKCs) can significantly enhance feature extraction, leading to improved accuracy in multiomics. This innovative approach is set to reshape our understanding of biological complexities, proving the power of fusing topological and geometric data. #ArtificialIntelligence #MachineLearning #DeepLearning #ComputerVision #DataScience #BigData #AIresearch #GeometricTopologicalAnalysis #BiomedicalImaging #Innovation

I am delighted to share that our paper, "Utilizing Keras Model for Dynamic Drawing Analysis in Predicting Parkinson's Disease through Spiral and Wave Patterns" has been presented at the International Conference for Innovation in Technology (INOCON) and published in IEEE Xplore! Our research focuses on identifying specific cues associated with health issues, which is crucial for developing effective clinical decision support systems. For patients with Parkinson’s disease (PD), the severity of the condition is directly reflected in impaired handwriting. Additionally, individuals with PD exhibit decreased movement and pen pressure during writing or drawing. Early detection of these symptoms can significantly improve clinical diagnosis. In our study, we propose a framework for analyzing spiral and wave patterns in both PD patients and healthy subjects. Utilizing the Keras model, we estimate drawing patterns and employ ensemble classification to predict the probability of spiral and wave sketches. Trained on patient data, our model achieves an impressive accuracy of 93.4%, 94% average recall, and 93.5% average precision, culminating in an F1 score of 93.94%. A huge thank you to Harene M and our mentors for their unwavering support throughout this process. This achievement wouldn't have been possible without your collaboration and encouragement. You can read our findings and explore our methods on IEEE Xplore here: https://lnkd.in/grHCUQCM #Research #ParkinsonsDisease #MachineLearning #Keras #INOCON #IEEE #ClinicalSupportSystems #IEEEXplore

Quantum computing is on the verge of revolutionizing healthcare as we know it, with its ability to perform complex tasks simultaneously, from Supersonic Drug Design to Precise Genomic Analysis and Predictive Health. Stay tuned for a healthier future powered by quantum computing! #QuantumComputing #HealthcareInnovation #FutureOfMedicine

Project & Outcome: Rank the severity of COVID-19 patients using radiomics and machine learning method • Collected CT images of COVID-19 patients and matched against the clinical data • Constructed radiomic feature pairs, tested the correlation between high-risk patients with recorded complications • Conducted an evaluation model based on qualitative relationship between feature pairs, which applied multiple algorithms (GA, Lasso, XGboost, ect.) for the feature selection process • Tested the accuracy of results in verification set, sorted the outcome, and carried out the survival analysis

#RecommendedPaper 📰Statistical Study of Low-Intensity Single-Molecule Recognition Events Using DeepTipTM Probes: Application to the Pru p 3-Phytosphingosine System by Rafael Daza, et al. 🔎Read the full paper here: https://lnkd.in/dCmfuQcP #atomic #force #microscopy

Selecting an appropriate metric for validating an ML model is relatively straightforward when working with a benchmark dataset. But it can become much for complicated with real world datasets. Which metric should you assess performance by? Do you need more than one? Metrics Reloaded provides a framework for categorizing most biomedical image analysis problems and a flowchart for identifying appropriate metrics. This may not be something you read end-to-end in one sitting but is very valuable for referring back to any time you start a new image analysis task. Even for those of us who have been working in the field for many years, there may be a metric we haven't heard of that is more suitable for a task -- and a good reason for selecting it over the more commonly used ones! https://lnkd.in/ernpEqyF If you enjoy these posts and want to hear more, sign up my Computer Vision Insights newsletter: https://lnkd.in/g9bSuQDP #Pathology #MedicalImaging #MachineLearning #DeepLearning #ComputerVision