The goal of the 2016 PhysioNet/CinC Challenge is the development of an algorithm to classify normal/abnormal heart sounds. A total of 124 time-frequency features were extracted from the phonocardiogram (PCG) and input to a variant of the AdaBoost classifier. A second classifier using convolutional neural network (CNN) was trained using PCGs cardiac cycles decomposed into four frequency bands. The final decision rule to classify normal/abnormal heart sounds was based on an ensemble of classifiers combining the outputs of AdaBoost and the CNN. The algorithm was trained on a training dataset (normal= 2575, abnormal= 665) and evaluated on a blind test dataset. Our classifier ensemble approach obtained the highest score of the competition with a sensitivity, specificity, and overall score of 0.9424, 0.7781, and 0.8602, respectively. IntroductionHeart auscultation is the primary tool for screening and diagnosis in primary health care [1]. Availability of digital stethoscopes and mobile devices provides clinicians an opportunity to record and analyze heart sounds (PCG) for diagnostic purposes. The goal of the 2016 PhysioNet/CinC Challenge is the development of algorithms to classify normal/abnormal heart sound recordings [2]. We proposed an ensemble of a feature-based classifier and a deep learningbased classifier to boost the classification performance of heart sounds. Method and MaterialA block diagram of the proposed approach to classify normal/abnormal PCG is shown in Fig. 1. Challenge DatabaseThe challenge database provided PCG recordings of healthy subjects and pathological patients collected at either a clinical or non-clinical environment. Details about the challenge dataset can be found in [2]. For algorithm development, in-house training and test sets were generated by randomly taking 80% and 20% of the records from each database, while keeping the same prevalence of abnormal classes. In-house training set was used for training and cross-validation of different models, and in-house test set was used for evaluation of the classification performance independently from the blind test dataset. Pre-processingEach PCG was resampled to 1000 Hz, band-pass filtered between 25 Hz and 400 Hz, and then pre-processed to remove any spikes in the PCG [3]. Furthermore, preprocessed PCGs were segmented into four heart sound states using a segmentation method proposed by Springer et al. [4]. Each PCG is comprised of more than one cardiac cycle (beat), and each beat is comprised of four heart sound states (i.e. S1, systole, S2, and diastole). Feature-based ApproachIn this approach, a variant of AdaBoost classifier [5] was trained for classification of normal/abnormal PCGs using time and frequency-domain features. Time-domain FeaturesMean and standard deviation (SD) of the following parameters were used as time-domain features (36 features): 1. PCG intervals: RR intervals, S1 intervals, S2 intervals, systolic intervals, diastolic intervals, ratio of systolic interval to RR interval of each heart beat, ratio of diastolic...

No abstract

Neuroimaging approaches have implicated multiple brain sites in musical perception, including the posterior part of the superior temporal gyrus and adjacent perisylvian areas. However, the detailed spatial and temporal relationship of neural signals that support auditory processing is largely unknown. In this study, we applied a novel inter-subject analysis approach to electrophysiological signals recorded from the surface of the brain (electrocorticography (ECoG)) in ten human subjects. This approach allowed us to reliably identify those ECoG features that were related to the processing of a complex auditory stimulus (i.e., continuous piece of music) and to investigate their spatial, temporal, and causal relationships. Our results identified stimulus-related modulations in the alpha (8-12 Hz) and high gamma (70-110 Hz) bands at neuroanatomical locations implicated in auditory processing. Specifically, we identified stimulus-related ECoG modulations in the alpha band in areas adjacent to primary auditory cortex, which are known to receive afferent auditory projections from the thalamus (80 of a total of 15107 tested sites). In contrast, we identified stimulus-related ECoG modulations in the high gamma band not only in areas close to primary auditory cortex but also in other perisylvian areas known to be involved in higher-order auditory processing, and in superior premotor cortex (412/15107 sites). Across all implicated areas, modulations in the high gamma band preceded those in the alpha band by 280 ms, and activity in the high gamma band causally predicted alpha activity, but not vice versa (Granger causality, p < 1e–8). Additionally, detailed analyses using Granger causality identified causal relationships of high gamma activity between distinct locations in early auditory pathways within superior temporal gyrus (STG) and posterior STG, between posterior STG and inferior frontal cortex, and between STG and premotor cortex. Evidence suggests that these relationships reflect direct cortico-cortical connections rather than common driving input from subcortical structures such as the thalamus. In summary, our inter-subject analyses defined the spatial and temporal relationships between music-related brain activity in the alpha and high gamma bands. They provide experimental evidence supporting current theories about the putative mechanisms of alpha and gamma activity, i.e., reflections of thalamo-cortical interactions and local cortical neural activity, respectively, and the results are also in agreement with existing functional models of auditory processing.