No abstract

Background Clinicians vary markedly in their ability to detect murmurs during cardiac auscultation and identify the underlying pathological features. Deep learning approaches have shown promise in medicine by transforming collected data into clinically significant information. The objective of this research is to assess the performance of a deep learning algorithm to detect murmurs and clinically significant valvular heart disease using recordings from a commercial digital stethoscope platform. Methods and Results Using >34 hours of previously acquired and annotated heart sound recordings, we trained a deep neural network to detect murmurs. To test the algorithm, we enrolled 962 patients in a clinical study and collected recordings at the 4 primary auscultation locations. Ground truth was established using patient echocardiograms and annotations by 3 expert cardiologists. Algorithm performance for detecting murmurs has sensitivity and specificity of 76.3% and 91.4%, respectively. By omitting softer murmurs, those with grade 1 intensity, sensitivity increased to 90.0%. Application of the algorithm at the appropriate anatomic auscultation location detected moderate‐to‐severe or greater aortic stenosis, with sensitivity of 93.2% and specificity of 86.0%, and moderate‐to‐severe or greater mitral regurgitation, with sensitivity of 66.2% and specificity of 94.6%. Conclusions The deep learning algorithm’s ability to detect murmurs and clinically significant aortic stenosis and mitral regurgitation is comparable to expert cardiologists based on the annotated subset of our database. The findings suggest that such algorithms would have utility as front‐line clinical support tools to aid clinicians in screening for cardiac murmurs caused by valvular heart disease. Registration URL: https://clinicaltrials.gov ; Unique Identifier: NCT03458806.

Background There is variability among clinicians in their ability to detect murmurs during cardiac auscultation and identify the underlying pathology. Deep learning approaches have shown promise in medicine by transforming collected data into clinically significant information. Objective The objective of this research is to assess the performance of a deep learning algorithm to detect murmurs and clinically significant valvular heart disease using recordings from a commercial digital stethoscope platform. Methods Using over 34 hours of previously acquired and annotated heart sound recordings, we trained a deep neural network to detect murmurs. To test the algorithm, we enrolled 373 patients in a clinical study and collected recordings at the four primary auscultation locations. Ground truth was established using patient echocardiograms and annotations by three expert cardiologists. Results Algorithm performance for detecting murmurs has sensitivity and specificity of 76.3% and 91.4%, respectively. By omitting softer murmurs, those with grade 1, sensitivity increases to 90.0%. The algorithm detects moderate-to-severe or greater aortic stenosis with sensitivity of 97.5% and specificity of 77.7% and detects moderate-to-severe or greater mitral regurgitation with sensitivity of 64.0% and specificity of 90.5%. Conclusion The deep learning algorithm's ability to detect murmurs and clinically significant aortic stenosis and mitral regurgitation is comparable to expert cardiologists. The research findings attest to the reliability and utility of such algorithms as front-line clinical support tools to aid clinicians in screening for cardiac murmurs caused by valvular heart disease.

Productive societies feature high levels of cooperation and strong connections between individuals. Public Goods Games (PGGs) are frequently used to study the development of social connections and cooperative behavior in model societies. In such games, contributions to the public good are made only by cooperators, while all players, including defectors, reap public goods benefits, which are shares of the contributions amplified by a synergy factor. Classic results of game theory show that mutual defection, as opposed to cooperation, is the Nash Equilibrium of PGGs in well-mixed populations, where each player interacts with all others. In this paper, we explore the coevolutionary dynamics of a low information public goods game on a complex network in which players adapt to their environment in order to increase individual payoffs relative to past payoffs parameterized by greediness. Players adapt by changing their strategies, either to cooperate or to defect, and by altering their social connections. We find that even if players do not know other players' strategies and connectivity, cooperation can arise and persist despite large short-term fluctuations.

This paper presents the Cogito submission to the Interspeech Computational Paralinguistics Challenge (ComParE), for the second sub-challenge. The aim of this second sub-challenge is to recognize self-assessed affect from short clips of speechcontaining audio data. We adopt a sequence classification-based approach where we use a long-short term memory (LSTM) network for modeling the evolution of low-level spectral coefficients, with added attention mechanism to emphasize salient regions of the audio clip. Additionally to deal with the underrepresentation of the negative valence class we use a combination of mitigation strategies including oversampling and loss function weighting. Our experiments demonstrate improvements in detection accuracy when including the attention mechanism and class balancing strategies in combination, with the best models outperforming the best single challenge baseline model.