Nevin Manimala

PLoS One. 2021 Nov 12;16(11):e0259442. doi: 10.1371/journal.pone.0259442. eCollection 2021.

ABSTRACT

BACKGROUND: Children need good nutrition to develop proper immune mechanisms and psychosocial maturity, but malnutrition can affect their ability to realize this. Apart from the national demographic and health survey, which is carried out every 5 years, there have not been enough documented studies on child breastfeeding and weaning practices of caregivers in the Volta Region. We, therefore, examined child breastfeeding and weaning practices of mothers in the Volta Region of Ghana.

METHODS: A sub-national survey method was adopted and a semi-structured questionnaire was used to collect data from 396 mothers and their children. Descriptive and inferential statistics comprising frequency, percentage, chi-square, and logistic regression were employed in analysing the data. We defined exclusive breastfeeding as given only breast milk to an infant from a mother or a wet nurse for six months of life except drops or syrups consisting of vitamins, minerals, supplements, or medicines on medical advice, and prolonged breastfeeding as breastfeeding up to 24 months of age.

RESULTS: The prevalence of exclusive breastfeeding (EBF) was 43.7%. Mothers constituting 61.1% started breastfeeding within an hour of giving birth. In addition to breast milk, 5.1% gave fluids to their children on the first day of birth. About 66.4% started complementary feeding at 6 months, 22.0% breastfed for 24 months or beyond, while 40.4% fed their children on-demand. Child’s age (AOR: 0.23, 95% CI:0.12-0.43, p<0.0001), prolonged breastfeeding (AOR: 0.41, 95%CI: 0.12-0.87, p = 0.001), mother’s religion (AOR: 3.92, 95%CI: 1.23-12.61, p = 0.021), feeding practices counselled on (AOR: 1.72, 95%CI: 1.96-3.09, p = 0.023), mother ever heard about EBF (AOR: 0.43, 95%CI: 1.45-2.41, p = 0.039), child being fed from the bottle with a nipple (AOR: 1.53, 95%CI: 1.94-2.48, p = 0.003), and age at which complementary feeding was started (AOR: 17.43, 95%CI: 3.47-87.55, p = 0.008) were statistically associated with EBF.

CONCLUSION: Breastfeeding education has been ongoing for decades, yet there are still gaps in the breastfeeding practices of mothers. To accelerate progress towards attainment of the sustainable development goal 3 of ensuring healthy lives and promoting well-being for all at all ages by the year 2030, we recommend innovative policies that include extensive public education to improve upon the breastfeeding and weaning practices of mothers.

PMID:34767566 | DOI:10.1371/journal.pone.0259442

IEEE Trans Neural Netw Learn Syst. 2021 Nov 12;PP. doi: 10.1109/TNNLS.2021.3124052. Online ahead of print.

ABSTRACT

Comparing competing mathematical models of complex processes is a shared goal among many branches of science. The Bayesian probabilistic framework offers a principled way to perform model comparison and extract useful metrics for guiding decisions. However, many interesting models are intractable with standard Bayesian methods, as they lack a closed-form likelihood function or the likelihood is computationally too expensive to evaluate. In this work, we propose a novel method for performing Bayesian model comparison using specialized deep learning architectures. Our method is purely simulation-based and circumvents the step of explicitly fitting all alternative models under consideration to each observed dataset. Moreover, it requires no hand-crafted summary statistics of the data and is designed to amortize the cost of simulation over multiple models, datasets, and dataset sizes. This makes the method especially effective in scenarios where model fit needs to be assessed for a large number of datasets, so that case-based inference is practically infeasible. Finally, we propose a novel way to measure epistemic uncertainty in model comparison problems. We demonstrate the utility of our method on toy examples and simulated data from nontrivial models from cognitive science and single-cell neuroscience. We show that our method achieves excellent results in terms of accuracy, calibration, and efficiency across the examples considered in this work. We argue that our framework can enhance and enrich model-based analysis and inference in many fields dealing with computational models of natural processes. We further argue that the proposed measure of epistemic uncertainty provides a unique proxy to quantify absolute evidence even in a framework which assumes that the true data-generating model is within a finite set of candidate models.

PMID:34767511 | DOI:10.1109/TNNLS.2021.3124052

IEEE Trans Pattern Anal Mach Intell. 2021 Nov 12;PP. doi: 10.1109/TPAMI.2021.3127870. Online ahead of print.

ABSTRACT

Correspondence analysis (CA) is a multivariate statistical tool used to visualize and interpret data dependencies by finding maximally correlated embeddings of pairs of random variables. CA has found applications in fields ranging from epidemiology to social sciences; however, current methods do not scale to large, high-dimensional datasets. In this paper, we provide a novel interpretation of CA in terms of an information-theoretic quantity called the principal inertia components. We show that estimating the principal inertia components, which consists in solving a functional optimization problem over the space of finite variance functions of two random variable, is equivalent to performing CA. We then leverage this insight to design novel algorithms to perform CA at an unprecedented scale. Particularly, we demonstrate how the principal inertia components can be reliably approximated from data using deep neural networks. Finally, we show how these maximally correlated embeddings of pairs of random variables in CA further play a central role in several learning problems including visualization of classification boundary and training process, and underlying recent multi-view and multi-modal learning methods.

PMID:34767505 | DOI:10.1109/TPAMI.2021.3127870

Am J Respir Crit Care Med. 2021 Nov 12. doi: 10.1164/rccm.202109-2096LE. Online ahead of print.

NO ABSTRACT

PMID:34767493 | DOI:10.1164/rccm.202109-2096LE

Circulation. 2021 Nov 12. doi: 10.1161/CIRCULATIONAHA.121.056012. Online ahead of print.

ABSTRACT

Background: Despite the acknowledged importance of socioeconomic factors as regards cardiovascular-disease onset, and survival, the relationship between individual-level socioeconomic factors and survival after out-of-hospital cardiac arrest (OHCA) is not fully established. Our aim was to investigate whether socioeconomic variables are associated with 30-day survival after OHCA. Methods: We linked data from the Swedish Registry for Cardiopulmonary Resuscitation with individual-level data on socioeconomic factors (i.e. educational level and disposable income) from Statistics Sweden. Confounding and mediating variables included demographic factors, comorbidity and Utstein resuscitation variables. Outcome was 30-day survival. Multiple modified Poisson regression was used for the main analyses. Results: A total of 31,373 OHCAs occurring in 2010-2017 were included. Crude 30-day survival rates by income quintiles were: Q1 (low) 414/6277 (6.6%), Q2=339/6276 (5.4%), Q3=423/6275 (6.7%), Q4=652/6273 (10.4%) and Q5 (high) 928/6272 (14.8%). In adjusted analysis, the chance of survival by income level followed a gradient-like increase, with a risk ratio (RR) of 1.86 (95% CI 1.65-2.09) in the highest-income quintile vs. the lowest. This association remained after adjusting for comorbidity, resuscitation factors and initial rhythm. A higher educational level was associated with improved 30-day survival, the RR associated with post-secondary education ≥ 4 years being 1.51 (95% CI 1.30-1.74). Survival disparities by income and educational level were observed in both men and women. Conclusions: In this nationwide observational study using individual-level socioeconomic data, higher income and higher educational level were associated with better 30-day survival following OHCA, in both sexes.

PMID:34767462 | DOI:10.1161/CIRCULATIONAHA.121.056012

Phys Rev Lett. 2021 Oct 29;127(18):181102. doi: 10.1103/PhysRevLett.127.181102.

ABSTRACT

It was recently pointed out that very energetic subclasses of supernovae (SNe), like hypernovae and superluminous SNe, might host ultrastrong magnetic fields in their core. Such fields may catalyze the production of feebly interacting particles, changing the predicted emission rates. Here we consider the case of axionlike particles (ALPs) and show that the predicted large scale magnetic fields in the core contribute significantly to the ALP production, via a coherent conversion of thermal photons. Using recent state-of-the-art supernova (SN) simulations, including magnetohydrodynamics, we find that, if ALPs have masses m_{a}∼O(10) MeV, their emissivity in such rare but exciting conditions via magnetic conversions would be over 2 orders of magnitude larger than previously estimated. Moreover, the radiative decay of these massive ALPs would lead to a peculiar delay in the arrival times of the daughter photons. Therefore, high-statistics gamma-ray satellites can potentially discover MeV ALPs in an unprobed region of the parameter space and shed light on the magnetohydrodynamical nature of the SN explosion.

PMID:34767416 | DOI:10.1103/PhysRevLett.127.181102

Phys Rev Lett. 2021 Oct 29;127(18):180602. doi: 10.1103/PhysRevLett.127.180602.

ABSTRACT

The statistical character of electron beams used in current technologies, as described by a stream of particles, is random in nature. Using coincidence measurements of femtosecond pulsed electron pairs, we report the observation of sub-Poissonian electron statistics that are nonrandom due to two-electron Coulomb interactions, and that exhibit an antibunching signal of 1 part in 4. This advancement is a fundamental step toward observing a strongly quantum degenerate electron beam needed for many applications, and in particular electron correlation spectroscopy.

PMID:34767409 | DOI:10.1103/PhysRevLett.127.180602

Phys Rev Lett. 2021 Oct 29;127(18):181101. doi: 10.1103/PhysRevLett.127.181101.

ABSTRACT

We resolve the fate of the two original apparent horizons during the head-on merger of two nonspinning black holes. We show that, following the appearance of the outer common horizon and subsequent interpenetration of the original horizons, they continue to exist for a finite period of time before they are individually annihilated by unstable marginally outer trapped surfaces (MOTSs). The inner common horizon vanishes in a similar, though independent, way. This completes the understanding of the analog of the event horizon’s “pair of pants” diagram for the apparent horizon. Our result is facilitated by a new method for locating MOTSs based on a generalized shooting method. We also discuss the role played by the MOTS stability operator in discerning which among a multitude of MOTSs should be considered as black hole boundaries.

PMID:34767408 | DOI:10.1103/PhysRevLett.127.181101

Phys Rev Lett. 2021 Oct 29;127(18):180601. doi: 10.1103/PhysRevLett.127.180601.

ABSTRACT

Optical speckle fields with both non-Rayleigh statistics and nondiffracting characteristics in propagation are an important light source for many applications. However, tailoring either non-Rayleigh statistical speckles or nondiffracting speckles are only investigated independently in previous studies. Here, we report the first observation of optical speckles that remain diffraction-free over a long axial distance while keeping non-Rayleigh statistics simultaneously. We further show the enhancement of Anderson localization of light with the non-Rayleigh nondiffracting speckles. The work presented here provides a versatile framework for customizing optical fields with desired speckle patterns for applications in the fields of solid-state physics, cold atoms, and optical imaging.

PMID:34767403 | DOI:10.1103/PhysRevLett.127.180601

J Pediatr Endocrinol Metab. 2021 Nov 11. doi: 10.1515/jpem-2021-0449. Online ahead of print.

ABSTRACT

OBJECTIVES: Since the beginning of the COVID-19 pandemic, there has been a consistent decrease in the number of admissions to the emergency department, leading to a delay in the diagnosis of several pathologies. The time from onset of symptoms to the diagnosis of Type 1 diabetes is highly variable. This treatment delay can lead to the appearance of ketoacidosis.

METHODS: Retrospective study of inaugural Type 1 diabetes cases, from March 2016 to March 2021. The pandemic group was considered between March 2020 to March 2021, and the remaining period was considered as pre-pandemic. Clinical variables were analysed: duration of symptoms, weight loss and value of ketonemia and glycated haemoglobin on admission. The mean differences were considered statistically significant at p<0.05.

RESULTS: 103 inaugural episodes of Type 1 diabetes were registered. The pandemic group had a lower mean age when compared to pre-pandemic group, and 51.7% of the episodes had ketoacidosis with a higher relative risk of ketoacidosis and severe ketoacidosis, when compared the pandemic with pre-pandemic group, there was a longer symptom evolution time (34 vs. 20 days), greater weight loss occurred (9.5% vs. 6.3%), the pH and HCO₃^– values were lower (7.30 vs. 7.36 and 16.43 vs. 20.71 mmol/L respectively) and ketonemia was higher (5.9 vs. 2.3 mmol/L).

CONCLUSIONS: The COVID-19 pandemic caused a delay in the diagnosis of Type 1 diabetes, greater length of disease, greater weight loss, higher ketonemia and lower pH and HCO₃^–. There was greater ketoacidosis relative risk in pandemic group when compared to pre-pandemic group.

PMID:34766743 | DOI:10.1515/jpem-2021-0449