Search

Training deep learning models requires having the right data for the problem and understanding both your data and the models’ performance on that data. Training deep learning models is difficult when data are limited, so in this paper, we seek to answer the following question: how can we train a deep learning model to increase its performance on a targeted area with limited data? We do this by applying rotation data augmentations to a simulated synthetic aperture radar (SAR) image dataset. We use the Uniform Manifold Approximation and Projection (UMAP) dimensionality reduction technique to understand the effects of augmentations on the data in latent space. Using this latent space representation, we can understand the data and choose specific training samples aimed at boosting model performance in targeted under-performing regions without the need to increase training set sizes. Results show that using latent space to choose training data significantly improves model performance in some cases; however, there are other cases where no improvements are made. We show that linking patterns in latent space is a possible predictor of model performance, but results require some experimentation and domain knowledge to determine the best options. Published version

Background Acute kidney injury (AKI) is a known complication of COVID-19 and is associated with an increased risk of in-hospital mortality. Unbiased proteomics using biological specimens can lead to improved risk stratification and discover pathophysiological mechanisms. Methods Using measurements of ~4000 plasma proteins in two cohorts of patients hospitalized with COVID-19, we discovered and validated markers of COVID-associated AKI (stage 2 or 3) and long-term kidney dysfunction. In the discovery cohort (N= 437), we identified 413 higher plasma abundances of protein targets and 40 lower plasma abundances of protein targets associated with COVID-AKI (adjusted p Results We demonstrate that COVID-AKI is associated with increased markers of tubular injury (NGAL) and myocardial injury. Using estimated glomerular filtration (eGFR) measurements taken after discharge, we also find that 25 of the 62 AKI-associated proteins are significantly associated with decreased post-discharge eGFR (adjusted p desmocollin-2, trefoil factor 3, transmembrane emp24 domain-containing protein 10, and cystatin-Cindicating tubular dysfunction and injury. Conclusions Using clinical and proteomic data, our results suggest that while both acute and long-term COVID-associated kidney dysfunction are associated with markers of tubular dysfunction, AKI is driven by a largely multifactorial process involving hemodynamic instability and myocardial damage.

Chronic hyperglycemia in type-1 diabetes mellitus is associated with oxidative stress (OS) and sudden death. Mechanistic links remain unclear. We investigated changes in electrophysiological (EP) properties in a model of chronic hyperglycemia before and after challenge with OS by GSH oxidation and tested reversibility of EP remodeling by insulin. Guinea pigs survived for 1 mo following streptozotocin (STZ) or saline (sham) injection. A treatment group received daily insulin for 2 wk to reverse STZ-induced hyperglycemia (STZ + Ins). EP properties were measured using high-resolution optical action potential mapping before and after challenge of hearts with diamide. Despite elevation of glucose levels in STZ compared with sham-operated ( P = 0.004) and STZ + Ins ( P = 0.002) animals, average action potential duration (APD) and arrhythmia propensity were not altered at baseline. Diamide promoted early (

Reactive oxygen species (ROS)-induced ROS release (RIRR) is a fundamental mechanism by which cardiac mitochondria respond to elevated ROS levels by stimulating endogenous ROS production in a regenerative, autocatalytic process that ultimately results in global oxidative stress (OS), cellular dysfunction and death. Despite elegant studies describing the phenomenon of RIRR under artificial conditions such as photo-induced oxidation of discrete regions within cardiomyocytes, the existence, biophysical properties and functional consequences of RIRR in intact myocardium remain unclear. Here, we used a semi-quantitative approach of optical superoxide (O(2)(-)) mapping using dihydroethidium (DHE) fluorescence to explore RIRR, its arrhythmic consequences and underlying mechanisms in intact myocardium. Initially, perfusion of rat hearts with 200 μM H(2)O(2) for 40 min (n = 4) elicited two distinct O(2)(-) peaks that were readily distinguished by their timing and amplitude. The first peak (P1), which was generated rapidly (within 5-8 min of H(2)O(2) perfusion) was associated with a relatively limited (10 ± 2%) rise in normalized O(2)(-) levels relative to baseline. In contrast, the second peak (P2) occurred 19-26 min following onset of H(2)O(2) perfusion and was associated with a significantly greater amplitude compared to P1. Spatio-temporal ROS mapping during P2 revealed active O(2)(-) propagation across the myocardium at a velocity of ~20 μm s(-1). Exposure of hearts (n = 18) to a short (10 min) episode of H(2)O(2) perfusion revealed consistent generation of P2 by high (≥200 μM, 8/8) but not lower (≤100 μM, 3/8) H(2)O(2) concentrations (P < 0.03). In these hearts, onset of P2 occurred following, not during, the 10 min OS protocol, consistent with RIRR. Importantly, P2 (+) hearts exhibited a markedly greater (by 3.8-fold, P < 0.001) arrhythmia score compared to P2 (-) hearts. To explore the mechanism underlying RIRR in intact myocardium, hearts were perfused with either cyclosporin A (CsA) or 4-chlorodiazepam (4-Cl-DZP) to inhibit the mitochondrial permeability transition pore (mPTP) or the inner membrane anion channel (IMAC), respectively. Surprisingly, perfusion with CsA failed to suppress (P = 0.75, n.s.) or even delay H(2)O(2)-induced P2 or the incidence of arrhythmias compared to untreated hearts. In sharp contrast, perfusion with 4-Cl-DZP markedly blunted O(2)(-) levels during P2, and suppressed the incidence of sustained ventricular tachycardia or ventricular fibrillation (VT/VF). Finally, perfusion of hearts with the synthetic superoxide dismutase/catalase mimetic EUK-134 completely abolished the H(2)O(2)-mediated RIRR response as well as the incidence of arrhythmias. These findings extend the concept of RIRR to the level of the intact heart, establish regenerative O(2)(-) production as the mediator of RIRR-related arrhythmias and reveal their strong dependence on IMAC and not the mPTP in this acute model of OS.

Reactive oxygen species (ROS)-induced ROS release (RIRR) is a fundamental mechanism by which cardiac mitochondria respond to elevated ROS levels by stimulating endogenous ROS production in a regenerative, autocatalytic process that ultimately results in global oxidative stress (OS), cellular dysfunction and death. Despite elegant studies describing the phenomenon of RIRR under artificial conditions such as photo-induced oxidation of discrete regions within cardiomyocytes, the existence, biophysical properties and functional consequences of RIRR in intact myocardium remain unclear. Here, we used a semi-quantitative approach of optical superoxide (O(2)(-)) mapping using dihydroethidium (DHE) fluorescence to explore RIRR, its arrhythmic consequences and underlying mechanisms in intact myocardium. Initially, perfusion of rat hearts with 200 μM H(2)O(2) for 40 min (n = 4) elicited two distinct O(2)(-) peaks that were readily distinguished by their timing and amplitude. The first peak (P1), which was generated rapidly (within 5-8 min of H(2)O(2) perfusion) was associated with a relatively limited (10 ± 2%) rise in normalized O(2)(-) levels relative to baseline. In contrast, the second peak (P2) occurred 19-26 min following onset of H(2)O(2) perfusion and was associated with a significantly greater amplitude compared to P1. Spatio-temporal ROS mapping during P2 revealed active O(2)(-) propagation across the myocardium at a velocity of ~20 μm s(-1). Exposure of hearts (n = 18) to a short (10 min) episode of H(2)O(2) perfusion revealed consistent generation of P2 by high (≥200 μM, 8/8) but not lower (≤100 μM, 3/8) H(2)O(2) concentrations (P0.03). In these hearts, onset of P2 occurred following, not during, the 10 min OS protocol, consistent with RIRR. Importantly, P2 (+) hearts exhibited a markedly greater (by 3.8-fold, P0.001) arrhythmia score compared to P2 (-) hearts. To explore the mechanism underlying RIRR in intact myocardium, hearts were perfused with either cyclosporin A (CsA) or 4-chlorodiazepam (4-Cl-DZP) to inhibit the mitochondrial permeability transition pore (mPTP) or the inner membrane anion channel (IMAC), respectively. Surprisingly, perfusion with CsA failed to suppress (P = 0.75, n.s.) or even delay H(2)O(2)-induced P2 or the incidence of arrhythmias compared to untreated hearts. In sharp contrast, perfusion with 4-Cl-DZP markedly blunted O(2)(-) levels during P2, and suppressed the incidence of sustained ventricular tachycardia or ventricular fibrillation (VT/VF). Finally, perfusion of hearts with the synthetic superoxide dismutase/catalase mimetic EUK-134 completely abolished the H(2)O(2)-mediated RIRR response as well as the incidence of arrhythmias. These findings extend the concept of RIRR to the level of the intact heart, establish regenerative O(2)(-) production as the mediator of RIRR-related arrhythmias and reveal their strong dependence on IMAC and not the mPTP in this acute model of OS.