This work attempts to reduce the number of false alarms generated by bedside monitors in the intensive care unit (ICU), as a majority of current alarms are false. In this study, we applied methods that can be categorized into three stages: signal processing, feature extraction, and optimized machine learning. At the stage of signal processing, we ensured that the heartbeats were properly annotated. During feature extraction, besides extracting features that are relevant to the arrhythmic alarms, we also extracted a set of signal quality indices (SQIs), which we used to distinguish noise/artifact from normal physiological signals. When applying a machine learning algorithm (Random Forest), we performed feature selection in order to reduce the complexity of the models and improve the efficiency of the algorithm. The dataset used is from Reducing False Arrhythmia Alarms in the ICU: the PhysioNet/Computing in Cardiology Challenge 2015. Using the performance metric “score” from the Challenge, we achieved a score of 83.08 in the real-time category on the hidden test set, which is the highest in all published work.

Oscillatory behavior of mitochondrial inner membrane potential (ΔΨ_m) is commonly observed in cells subjected to oxidative or metabolic stress. In cardiac myocytes, the activation of inner membrane pores by reactive oxygen species (ROS) is a major factor mediating intermitochondrial coupling, and ROS-induced ROS release has been shown to underlie propagated waves of ΔΨ_m depolarization as well as synchronized limit cycle oscillations of ΔΨ_m in the network. The functional impact of ΔΨ_m instability on cardiac electrophysiology, Ca²⁺ handling, and even cell survival, is strongly affected by the extent of such intermitochondrial coupling. Here, we employ a recently developed wavelet-based analytical approach to examine how different substrates affect mitochondrial coupling in cardiac cells, and we also determine the oscillatory coupling properties of mitochondria in ventricular cells in intact perfused hearts. The results show that the frequency of ΔΨ_m oscillations varies inversely with the size of the oscillating mitochondrial cluster, and depends on the strength of local intermitochondrial coupling. Time-varying coupling constants could be quantitatively determined by applying a stochastic phase model based on extension of the well-known Kuramoto model for networks of coupled oscillators. Cluster size-frequency relationships varied with different substrates, as did mitochondrial coupling constants, which were significantly larger for glucose (7.78 × 10⁻² ± 0.98 × 10⁻² s⁻¹) and pyruvate (7.49 × 10⁻² ± 1.65 × 10⁻² s⁻¹) than lactate (4.83 × 10⁻² ± 1.25 × 10⁻² s⁻¹) or β-hydroxybutyrate (4.11 × 10⁻² ± 0.62 × 10⁻² s⁻¹). The findings indicate that mitochondrial spatiotemporal coupling and oscillatory behavior is influenced by substrate selection, perhaps through differing effects on ROS/redox balance. In particular, glucose-perfusion generates strong intermitochondrial coupling and temporal oscillatory stability. Pathological changes in specific catabolic pathways, which are known to occur during the progression of cardiovascular disease, could therefore contribute to altered sensitivity of the mitochondrial network to oxidative stress and emergent ΔΨ_m instability, ultimately scaling to produce organ level dysfunction.

Previous studies have demonstrated that microvolt T-wave alternans (MTWA) testing is a robust predictor of ventricular tachyarrhythmias and sudden cardiac death (SCD) in at-risk patients. However, recent studies have suggested that MTWA testing is not as good a predictor of "appropriate" implantable cardioverter-defibrillator (ICD) therapy as it is a predictor of SCD in patients without ICDs. To evaluate the utility of MTWA testing for SCD risk stratification in patients without ICDs. Patient-level data were obtained from 5 prospective studies of MTWA testing in patients with no history of ventricular arrhythmia or SCD. In these studies, ICDs were implanted in only a minority of patients and patients with ICDs were excluded from the analysis. We conducted a pooled analysis and examined the 2-year risk for SCD based on the MTWA test result. The pooled cohort included 2883 patients. MTWA testing was positive in 856 (30%), negative in 1627 (56%), and indeterminate in 400 (14%) patients. Among patients with a left ventricular ejection fraction (LVEF) of ≤35%, annual SCD event rates were 4.0%, 0.9%, and 4.6% among groups with MTWA positive, negative, and indeterminate test results. The SCD rate was significantly lower among patients with a negative MTWA test result than in patients with either positive or indeterminate MTWA test results (P <.001 for both comparisons). In patients with an LVEF of >35%, annual SCD event rates were 3.0%, 0.3%, and 0.3% among the groups with MTWA positive, negative, and indeterminate test results. The SCD rate associated with a positive MTWA test result was significantly higher than that associated with either negative (P <.001) or indeterminate MTWA test results (P = .003). In patients without ICDs, MTWA testing is a powerful predictor of SCD. Among patients with an LVEF of ≤35%, a negative MTWA test result is associated with a low risk for SCD. Conversely, among patients with an LVEF of >35%, a positive MTWA test result identifies patients at significantly heightened SCD risk. These findings may have important implications for refining primary prevention ICD treatment algorithms.