Search

Label noise is omnipresent in the annotations process and has an impact on supervised learning algorithms. This work focuses on the impact of label noise on the performance of learning models by examining the effect of random and class-dependent label noise on a binary classification task: quality assessment for photoplethysmography (PPG). PPG signal is used to detect physiological changes and its quality can have a significant impact on the subsequent tasks, which makes PPG quality assessment a particularly good target for examining the impact of label noise in the field of biomedicine. Random and class-dependent label noise was introduced separately into the training set to emulate the errors associated with fatigue and bias in labeling data samples. We also tested different representations of the PPG, including features defined by domain experts, 1D raw signal and 2D image. Three different classifiers are tested on the noisy training data, including support vector machine (SVM), XGBoost, 1D Resnet and 2D Resnet, which handle three representations, respectively. The results showed that the two deep learning models were more robust than the two traditional machine learning models for both the random and class-dependent label noise. From the representation perspective, the 2D image shows better robustness compared to the 1D raw signal. The logits from three classifiers are also analyzed, the predicted probabilities intend to be more dispersed when more label noise is introduced. From this work, we investigated various factors related to label noise, including representations, label noise type, and data imbalance, which can be a good guidebook for designing more robust methods for label noise in future work.

Artificial intelligence (AI)-based solutions have revolutionized our world, using extensive datasets and computational resources to create automatic tools for complex tasks that, until now, have been performed by humans. Massive data is a fundamental aspect of the most powerful AI-based algorithms. However, for AI-based healthcare solutions, there are several socioeconomic, technical/infrastructural, and most importantly, legal restrictions, which limit the large collection and access of biomedical data, especially medical imaging. To overcome this important limitation, several alternative solutions have been suggested, including transfer learning approaches, generation of artificial data, adoption of blockchain technology, and creation of an infrastructure composed of anonymous and abstract data. However, none of these strategies is currently able to completely solve this challenge. The need to build large datasets that can be used to develop healthcare solutions deserves special attention from the scientific community, clinicians, all the healthcare players, engineers, ethicists, legislators, and society in general. This paper offers an overview of the data limitation in medical predictive models; its impact on the development of healthcare solutions; benefits and barriers of sharing data; and finally, suggests future directions to overcome data limitations in the medical field and enable AI to enhance healthcare. This perspective is dedicated to the technical requirements of the learning models, and it explains the limitation that comes from poor and small datasets in the medical domain and the technical options that try or can solve the problem related to the lack of massive healthcare data.

Photoplethysmography (PPG) is a noninvasive way to monitor various aspects of the circulatory system, and is becoming more and more widespread in biomedical processing. Recently, deep learning methods for analyzing PPG have also become prevalent, achieving state of the art results on heart rate estimation, atrial fibrillation detection, and motion artifact identification. Consequently, a need for interpretable deep learning has arisen within the field of biomedical signal processing. In this paper, we pioneer novel explanatory metrics which leverage domain-expert knowledge to validate a deep learning model. We visualize model attention over a whole testset using saliency methods and compare it to human expert annotations. Congruence, our first metric, measures the proportion of model attention within expert-annotated regions. Our second metric, Annotation Classification, measures how much of the expert annotations our deep learning model pays attention to. Finally, we apply our metrics to compare between a signal based model and an image based model for PPG signal quality classification. Both models are deep convolutional networks based on the ResNet architectures. We show that our signal-based one dimensional model acts in a more explainable manner than our image based model; on average 50.78% of the one dimensional model's attention are within expert annotations, whereas 36.03% of the two dimensional model's attention are within expert annotations. Similarly, when thresholding the one dimensional model attention, one can more accurately predict if each pixel of the PPG is annotated as artifactual by an expert. Through this testcase, we demonstrate how our metrics can provide a quantitative and dataset-wide analysis of how explainable the model is.

Early detection of Atrial Fibrillation (AFib) is crucial to prevent stroke recurrence. New tools for monitoring cardiac rhythm are important for risk stratification and stroke prevention. As many of new approaches to long-term AFib detection are now based on photoplethysmogram (PPG) recordings from wearable devices, ensuring high PPG signal-to-noise ratios is a fundamental requirement for a robust detection of AFib episodes. Traditionally, signal quality assessment is often based on the evaluation of similarity between pulses to derive signal quality indices. There are limitations to using this approach for accurate assessment of PPG quality in the presence of arrhythmia, as in the case of AFib, mainly due to substantial changes in pulse morphology. In this paper, we first tested the performance of algorithms selected from a body of studies on PPG quality assessment using a dataset of PPG recordings from patients with AFib. We then propose machine learning approaches for PPG quality assessment in 30-s segments of PPG recording from 13 stroke patients admitted to the University of California San Francisco (UCSF) neuro intensive care unit and another dataset of 3764 patients from one of the five UCSF general intensive care units. We used data acquired from two systems, fingertip PPG (fPPG) from a bedside monitor system, and radial PPG (rPPG) measured using a wearable commercial wristband. We compared various supervised machine learning techniques including k-nearest neighbors, decisions trees, and a two-class support vector machine (SVM). SVM provided the best performance. fPPG signals were used to build the model and achieved 0.9477 accuracy when tested on the data from the fPPG exclusive to the test set, and 0.9589 accuracy when tested on the rPPG data.

Atrial fibrillation (AF) is a cardiac rhythm disorder associated with increased morbidity and mortality. It is the leading risk factor for cardioembolic stroke and its early detection is crucial in both primary and secondary stroke prevention. Continuous monitoring of cardiac rhythm is today possible thanks to consumer-grade wearable devices, enabling transformative diagnostic and patient management tools. Such monitoring is possible using low-cost easy-to-implement optical sensors that today equip the majority of wearables. These sensors record blood volume variations-a technology known as photoplethysmography (PPG)-from which the heart rate and other physiological parameters can be extracted to inform about user activity, fitness, sleep, and health. Recently, new wearable devices were introduced as being capable of AF detection, evidenced by large prospective trials in some cases. Such devices would allow for early screening of AF and initiation of therapy to prevent stroke. This review is a summary of a body of work on AF detection using PPG. A thorough account of the signal processing, machine learning, and deep learning approaches used in these studies is presented, followed by a discussion of their limitations and challenges towards clinical applications.

Introduction: Current microvascular assessments may not be practical or accessible requiring experienced personnel and/or ongoing equipment costs. Piezoelectric transducers can reliably obtain finger blood pressure waves, similar to peripheral arterial tonometry devices; thus, they could be used to estimate microvascular function. We aimed to validate piezoelectric transducers as an alternative measure of microvascular function compared to EndoPAT. Methods: Twenty-five adults (aged 20–64 years) completed reactive hyperemia (5 min forearm circulatory occlusion and 3 min recovery) with piezoelectric transducers on the middle fingers and EndoPAT probes on the index fingers. Average area under the curve (AUC) of the pulse wave signal for the occluded and control arms was determined at baseline, every 30 s post-occlusion, and 10 s around the peak response. Microvascular function index (MFI) was calculated as the ratio of AUC post-occlusion to AUC baseline in the test arm, then normalized to the same ratio in the control arm. MFI at each time point was correlated with the reactive hyperemia index (RHI) from the EndoPAT. Results: The greatest significance was found between RHI and MFI at 10 s around the peak response (Spearman's r = 0.67, p = 0.0002; Pearson's r = 0.76, p = 0.00001). Conclusion: MFI is a reusable and user-friendly microvascular function assessment that could provide better access to vascular health screening. [ABSTRACT FROM AUTHOR]