Your evidence? Machine learning algorithms for medical diagnosis and prediction

Heinrichs, Bert; Eickhoff, Simon B.

doi:10.1002/hbm.24886

Cited by 72 publications

(45 citation statements)

References 43 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…As mentioned above, a particularly important application of machine learning in a healthcare context is (digital) diagnosis (see e.g. [10][11][12][13][14]. Machine learning models can detect patterns (precursor) of certain diseases within patient electronic healthcare records and inform clinicians of any anomalies.…”

mentioning

confidence: 99%

Detecting cardiac pathologies via machine learning on heart-rate variability time series and related markers

Agliari

Barra

et al. 2020

Sci Rep

View full text Add to dashboard Cite

in this paper we develop statistical algorithms to infer possible cardiac pathologies, based on data collected from 24 h Holter recording over a sample of 2829 labelled patients; labels highlight whether a patient is suffering from cardiac pathologies. In the first part of the work we analyze statistically the heartbeat series associated to each patient and we work them out to get a coarse-grained description of heart variability in terms of 49 markers well established in the reference community. These markers are then used as inputs for a multi-layer feed-forward neural network that we train in order to make it able to classify patients. However, before training the network, preliminary operations are in order to check the effective number of markers (via principal component analysis) and to achieve data augmentation (because of the broadness of the input data). With such groundwork, we finally train the network and show that it can classify with high accuracy (at most ~85% successful identifications) patients that are healthy from those displaying atrial fibrillation or congestive heart failure. In the second part of the work, we still start from raw data and we get a classification of pathologies in terms of their related networks: patients are associated to nodes and links are drawn according to a similarity measure between the related heartbeat series. We study the emergent properties of these networks looking for features (e.g., degree, clustering, clique proliferation) able to robustly discriminate between networks built over healthy patients or over patients suffering from cardiac pathologies. We find overall very good agreement among the two paved routes. Artificial intelligence (AI) is gaining a growing role in healthcare: in the last years, several devices and advanced algorithms have been successfully employed to assist medical workers (see e.g. 1-5). Among the most important goals of this partnership between humans and machines is the wide accessibility (even in low-income and remote areas) to medical assistance and the reduction of the time needed to reach a diagnosis. Of course, in order for AI-based devices to analyze large amounts of information and make (fast and correct) decisions, they first need to undergo a suitable training 6-9. Basically, during training, a machine-learning model is exposed to examples and its internal parameters are tuned accordingly; once training is over, new data are presented to the model which then uses what it has learned to explain that data. For instance, a model meant to classify images of skin lesions as benign lesions or malignant skin cancer will be trained on a dataset of skin pictures from different patients, previously labeled as benign or malignant, through which the model learns to detect in the input image specific patterns that are hallmarks of malignancies. Clearly, the more accurate the training and the better the performance. Nowadays, an accurate training is in principle possible given that each patient generates large volumes of health data such as...

show abstract

mentioning

confidence: 99%

Detecting cardiac pathologies via machine learning on heart-rate variability time series and related markers

Agliari

Barra

et al. 2020

Sci Rep

View full text Add to dashboard Cite

show abstract

“…In this sense, we interpret explanations as understanding support [59], and this latter one in the light of the critique by Kelp to both the "explanationist view" and the "manipulationist views" of understanding, that is in terms of supporting (human) understanding build a "wellconnected knowledge" [60]. According to this view, understanding how a system works does not only involve knowing a set of true propositions about the system behaviors (like in case of data about predictive accuracy and feature ranking for a given prediction), but also knowing how these propositions are interrelated, within a framework of sense-making.…”

Section: Implications For the Xai Research Fieldmentioning

confidence: 96%

As if sand were stone. New concepts and metrics to probe the ground on which to build trustable AI

Cabitza

Campagner

Sconfienza

2020

BMC Med Inform Decis Mak

View full text Add to dashboard Cite

Background We focus on the importance of interpreting the quality of the labeling used as the input of predictive models to understand the reliability of their output in support of human decision-making, especially in critical domains, such as medicine. Methods Accordingly, we propose a framework distinguishing the reference labeling (or Gold Standard) from the set of annotations from which it is usually derived (the Diamond Standard). We define a set of quality dimensions and related metrics: representativeness (are the available data representative of its reference population?); reliability (do the raters agree with each other in their ratings?); and accuracy (are the raters’ annotations a true representation?). The metrics for these dimensions are, respectively, the degree of correspondence, Ψ, the degree of weighted concordanceϱ, and the degree of fineness, Φ. We apply and evaluate these metrics in a diagnostic user study involving 13 radiologists. Results We evaluate Ψ against hypothesis-testing techniques, highlighting that our metrics can better evaluate distribution similarity in high-dimensional spaces. We discuss how Ψ could be used to assess the reliability of new predictions or for train-test selection. We report the value of ϱ for our case study and compare it with traditional reliability metrics, highlighting both their theoretical properties and the reasons that they differ. Then, we report the degree of fineness as an estimate of the accuracy of the collected annotations and discuss the relationship between this latter degree and the degree of weighted concordance, which we find to be moderately but significantly correlated. Finally, we discuss the implications of the proposed dimensions and metrics with respect to the context of Explainable Artificial Intelligence (XAI). Conclusion We propose different dimensions and related metrics to assess the quality of the datasets used to build predictive models and Medical Artificial Intelligence (MAI). We argue that the proposed metrics are feasible for application in real-world settings for the continuous development of trustable and interpretable MAI systems.

show abstract

“…Furthermore, rigid algorithm protocols and decision-making trees are subject to the consequences of the inability of AI to fully take in and interpret contextual information or delineate between relevant vs. non-relevant informational input even when employing deep machine learning (23). Contingencies are the norm in healthcare, and the human skill required to navigate and manage this offnominal, or unpredictable situations must be carefully weighed against the advantages of using AI technology (23)(24)(25)(26)(27)(28). User interface and data input methods are critical as voice recognition and interpretation is a major challenge of AI utilization (29).…”

Section: Artificial Intelligence Assisted Telemedicinementioning

confidence: 99%

Designing Futuristic Telemedicine Using Artificial Intelligence and Robotics in the COVID-19 Era

Bhaskar

Bradley

Sakhamuri

et al. 2020

Front. Public Health

102

View full text Add to dashboard Cite

Technological innovations such as artificial intelligence and robotics may be of potential use in telemedicine and in building capacity to respond to future pandemics beyond the current COVID-19 era. Our international consortium of interdisciplinary experts in clinical medicine, health policy, and telemedicine have identified gaps in uptake and implementation of telemedicine or telehealth across geographics and medical specialties. This paper discusses various artificial intelligence and robotics-assisted telemedicine or telehealth applications during COVID-19 and presents an alternative artificial intelligence assisted telemedicine framework to accelerate the rapid deployment of telemedicine and improve access to quality and cost-effective healthcare. We postulate that the artificial intelligence assisted telemedicine framework would be indispensable in creating futuristic and resilient health systems that can support communities amidst pandemics.

show abstract

Your evidence? Machine learning algorithms for medical diagnosis and prediction

Cited by 72 publications

References 43 publications

Detecting cardiac pathologies via machine learning on heart-rate variability time series and related markers

Detecting cardiac pathologies via machine learning on heart-rate variability time series and related markers

As if sand were stone. New concepts and metrics to probe the ground on which to build trustable AI

Designing Futuristic Telemedicine Using Artificial Intelligence and Robotics in the COVID-19 Era

Contact Info

Product

Resources

About