Classification of Hypoglycemic Events in Type 1 Diabetes Using Machine Learning Algorithms.

INTRODUCTION: To improve the utilization of continuous- and flash glucose monitoring (CGM/FGM) data we have tested the hypothesis that a machine learning (ML) model can be trained to identify the most likely root causes for hypoglycemic events. METHODS: CGM/FGM data were collected from 449 patients with type 1 diabetes. Of the 42,120 identified hypoglycemic events, 5041 were randomly selected for classification by two clinicians. Three causes of hypoglycemia were deemed possible to interpret and later validate by insulin and carbohydrate recordings: (1) overestimated bolus (27%), (2) overcorrection of hyperglycemia (29%) and (3) excessive basal insulin presure (44%). The dataset was split into a training (n = 4026 events, 304 patients) and an internal validation dataset (n = 1015 events, 145 patients). A number of ML model architectures were applied and evaluated. A separate dataset was generated from 22 patients (13 'known' and 9 'unknown') with insulin and carbohydrate recordings. Hypoglycemic events from this dataset were also interpreted by five clinicians independently. RESULTS: Of the evaluated ML models, a purpose-built convolutional neural network (HypoCNN) performed best. Masking the time series, adding time features and using class weights improved the performance of this model, resulting in an average area under the curve (AUC) of 0.921 in the original train/test split. In the dataset validated by insulin and carbohydrate recordings (n = 435 events), i.e. 'ground truth,' our HypoCNN model achieved an AUC of 0.917. CONCLUSIONS: The findings support the notion that ML models can be trained to interpret CGM/FGM data. Our HypoCNN model provides a robust and accurate method to identify root causes of hypoglycemic events.

Erratum: State-Resolved Mutual Neutralization of Mg

This corrects the article DOI: 10.1103/PhysRevLett.128.033401.

High-precision electron affinity of oxygen.

Negative ions are important in many areas of science and technology, e.g., in interstellar chemistry, for accelerator-based radionuclide dating, and in anti-matter research. They are unique quantum systems where electron-correlation effects govern their properties. Atomic anions are loosely bound systems, which with very few exceptions lack optically allowed transitions. This limits prospects for high-resolution spectroscopy, and related negative-ion detection methods. Here, we present a method to measure negative ion binding energies with an order of magnitude higher precision than what has been possible before. By laser-manipulation of quantum-state populations, we are able to strongly reduce the background from photodetachment of excited states using a cryogenic electrostatic ion-beam storage ring where keV ion beams can circulate for up to hours. The method is applicable to negative ions in general and here we report an electron affinity of 1.461 112 972(87) eV for 16O.

State-Resolved Mutual Neutralization of Mg

We present experimental final-state distributions for Mg atoms formed in Mg^{+}+D^{-} mutual neutralization reactions at center-of-mass collision energies of 59±12 meV by using the merged-beams method. Comparisons with available full-quantum results reveal large discrepancies and a previously underestimated total rate coefficient by up to a factor of 2 in the 0-1 eV (<10^{4} K) regime. Asymptotic model calculations are shown to describe the process much better and we recommend applying this method to more complex iron group systems; data that is of urgent need in stellar spectral modeling.

Mutual neutralisation of O+ with O-: investigation of the role of metastable ions in a combined experimental and theoretical study.

The mutual neutralisation of O+ with O- has been studied in a double ion-beam storage ring with combined merged-beams, imaging and timing techniques. Branching ratios were measured at the collision energies of 55, 75 and 170 (± 15) meV, and found to be in good agreement with previous single-pass merged-beams experimental results at 7 meV collision energy. Several previously unidentified spectral features were found to correspond to mutual neutralisation channels of the first metastable state of the cation (O+(2Do), τ ≈ 3.6 hours), while no contributions from the second metastable state (O+(2Po), τ ≈ 5 seconds) were observed. Theoretical calculations were performed using the multi-channel Landau-Zener model combined with the anion centered asymptotic method, and gave good agreement with several experimentally observed channels, but could not describe well observed contributions from the O+(2Do) metastable state as well as channels involving the O(3s 5So) state.

Ultraslow radiative cooling of Cn - (n = 3-5).

Ultraslow radiative cooling lifetimes and adiabatic detachment energies for three astrochemically relevant anions, Cn - (n = 3-5), are measured using the Double ElectroStatic Ion Ring ExpEriment (DESIREE) infrastructure at Stockholm University. DESIREE maintains a background pressure of ≈10-14 mbar and temperature of ≈13 K, allowing storage of mass-selected ions for hours and providing conditions coined a "molecular cloud in a box." Here, we construct two-dimensional (2D) photodetachment spectra for the target anions by recording photodetachment signal as a function of irradiation wavelength and ion storage time (seconds to minute time scale). Ion cooling lifetimes, which are associated with infrared radiative emission, are extracted from the 2D photodetachment spectrum for each ion by tracking the disappearance of vibrational hot-band signal with ion storage time, giving 1e cooling lifetimes of 3.1 ± 0.1 s (C3 -), 6.8 ± 0.5 s (C4 -), and 24 ± 5 s (C5 -). Fits of the photodetachment spectra for cold ions, i.e., those stored for at least 30 s, provide adiabatic detachment energies in good agreement with values from laser photoelectron spectroscopy on jet-cooled anions, confirming that radiative cooling has occurred in DESIREE. Ion cooling lifetimes are simulated using a simple harmonic cascade model, finding good agreement with experiment and providing a mode-by-mode understanding of the radiative cooling properties. The 2D photodetachment strategy and radiative cooling modeling developed in this study could be applied to investigate the ultraslow cooling dynamics of a wide range of molecular anions.