The mixtools package for R provides a set of functions for analyzing a variety of finite mixture models. These functions include both traditional methods, such as EM algorithms for univariate and multivariate normal mixtures, and newer methods that reflect some recent research in finite mixture models. In the latter category, mixtools provides algorithms for estimating parameters in a wide range of different mixture-of-regression contexts, in multinomial mixtures such as those arising from discretizing continuous multivariate data, in nonparametric situations where the multivariate component densities are completely unspecified, and in semiparametric situations such as a univariate location mixture of symmetric but otherwise unspecified densities. Many of the algorithms of the mixtools package are EM algorithms or are based on EM-like ideas, so this article includes an overview of EM algorithms for finite mixture models.

Abstract:We propose an algorithm for nonparametric estimation for finite mixtures of multivariate random vectors that strongly resembles a true EM algorithm. The vectors are assumed to have independent coordinates conditional upon knowing which mixture component from which they come, but otherwise their density functions are completely unspecified. Sometimes, the density functions may be partially specified by Euclidean parameters, a case we call semiparametric. Our algorithm is much more flexible and easily applicable than existing algorithms in the literature; it can be extended to any number of mixture components and any number of vector coordinates of the multivariate observations. Thus it may be applied even in situations where the model is not identifiable, so care is called for when using it in situations for which identifiability is difficult to establish conclusively. Our algorithm yields much smaller mean integrated squared errors than an alternative algorithm in a simulation study. In another example using a real dataset, it provides new insights that extend previous analyses. Finally, we present two different variations of our algorithm, one stochastic and one deterministic, and find anecdotal evidence that there is not a great deal of difference between the performance of these two variants.

No abstract

Resting-state functional connectivity (rsFC) has gained popularity mainly due to its simplicity and potential for providing insights into various brain disorders. In this vein, functional near-infrared spectroscopy (fNIRS) is an attractive choice due to its portability, flexibility, and low cost, allowing for bedside imaging of brain function. While promising, fNIRS suffers from non-neural signal contaminations (i.e., systemic physiological noise), which can increase correlation across fNIRS channels, leading to spurious rsFC networks. In the present work, we hypothesized that additional measurements with short channels, heart rate, mean arterial pressure, and end-tidal CO2 could provide a better understanding of the effects of systemic physiology on fNIRS-based resting-state networks. To test our hypothesis, we acquired 12 min of resting-state data from 10 healthy participants. Unlike previous studies, we investigated the efficacy of different pre-processing approaches in extracting resting-state networks. Our results are in agreement with previous studies and reinforce the fact that systemic physiology can overestimate rsFC. We expanded on previous work by showing that removal of systemic physiology decreases intra- and inter-subject variability, increasing the ability to detect neural changes in rsFC across groups and over longitudinal studies. Our results show that by removing systemic physiology, fNIRS can reproduce resting-state networks often reported with functional magnetic resonance imaging (fMRI). Finally, the present work details the effects of systemic physiology and outlines how to remove (or at least ameliorate) their contributions to fNIRS signals acquired at rest.

A BS TRACT: Background: Hereditary spastic paraplegia presents spasticity as the main clinical manifestation, reducing gait quality and producing incapacity. Management with botulinum toxin type A (BoNT-A) is not well elucidated. The objective of the current study was to evaluate the efficacy and safety of BoNT-A in patients with hereditary spastic paraplegias. Methods: This was a double-blind, randomized, placebo-controlled crossover trial. Each participant was randomly assigned to receive 1 injection session of either BoNT-A (100 IU/2 mL of Prosigne in each adductor magnus and each triceps surae) or saline 0.9% (2 mL). The primary outcome measure was change from baseline in maximal gait velocity, and secondary outcome measures included changes in gait at self-selected velocity, spasticity, muscle strength, Spastic Paraplegia Rating Scale, pain, fatigue, and subjective perception of improvement. We also looked at adverse events reported by the patients.Results: We enrolled 55 patients, 36 of whom were men and 41 with the pure phenotype. Mean age was 43 AE 13.4 years (range, 19-72 years), mean age of onset waws 27 AE 13.1 years (range, <1 to 55 yars), and mean disease duration was 17 AE 12.7 years (range, 1-62 years). Compared with baseline, we did not find significant differences between groups in primary and secondary outcomes, except for reduction in adductor tone (P = 0.01). The adverse events were transient and tolerable, and their incidence did not significantly differ between treatments (P = 0.17). Conclusions: BoNT-A was safe in patients with hereditary spastic paraplegias and reduced the adductor tone, but it was not able to produce functional improvement considering the doses, injection protocol, measures, and instruments used.

Background SARS-CoV-2, the virus that causes COVID-19, enters the cells through a mechanism dependent on its binding to angiotensin-converting enzyme 2 (ACE2), a protein highly expressed in the lungs. The putative viral-induced inhibition of ACE2 could result in the defective degradation of bradykinin, a potent inflammatory substance. We hypothesize that increased bradykinin in the lungs is an important mechanism driving the development of pneumonia and respiratory failure in COVID-19. Methods This is a phase II, single-center, three-armed parallel-group, open-label, active control superiority randomized clinical trial. One hundred eighty eligible patients will be randomly assigned in a 1:1:1 ratio to receive either the inhibitor of C1e/kallikrein 20 U/kg intravenously on day 1 and day 4 plus standard care; or icatibant 30 mg subcutaneously, three doses/day for 4 days plus standard care; or standard care alone, as recommended in the clinical trials published to date, which includes supplemental oxygen, non-invasive and invasive ventilation, antibiotic agents, anti-inflammatory agents, prophylactic antithrombotic therapy, vasopressor support, and renal replacement therapy. Discussion Accumulation of bradykinin in the lungs is a common side effect of ACE inhibitors leading to cough. In animal models, the inactivation of ACE2 leads to severe acute pneumonitis in response to lipopolysaccharide (LPS), and the inhibition of bradykinin almost completely restores the lung structure. We believe that inhibition of bradykinin in severe COVID-19 patients could reduce the lung inflammatory response, impacting positively on the severity of disease and mortality rates. Trial registration Brazilian Clinical Trials Registry Universal Trial Number (UTN) U1111-1250-1843. Registered on May/5/2020.

Health economic evaluations require estimates of expected survival from patients receiving different interventions, often over a lifetime. However, data on the patients of interest are typically only available for a much shorter follow-up time, from randomised trials or cohorts. Previous work showed how to use general population mortality to improve extrapolations of the short-term data, assuming a constant additive or multiplicative effect on the hazards for all-cause mortality for study patients relative to the general population. A more plausible assumption may be a constant effect on the hazard for the specific cause of death targeted by the treatments. To address this problem, we use independent parametric survival models for cause-specific mortality among the general population. Because causes of death are unobserved for the patients of interest, a polyhazard model is used to express their all-cause mortality as a sum of latent cause-specific hazards. Assuming proportional cause-specific hazards between the general and study populations then allows us to extrapolate mortality of the patients of interest to the long term. A Bayesian framework is used to jointly model all sources of data. By simulation, we show that ignoring cause-specific hazards leads to biased estimates of mean survival when the proportion of deaths due to the cause of interest changes through time. The methods are applied to an evaluation of implantable cardioverter defibrillators for the prevention of sudden cardiac death among patients with cardiac arrhythmia. After accounting for cause-specific mortality, substantial differences are seen in estimates of life years gained from implantable cardioverter defibrillators.

In this paper we describe a method to select the bandwidth used in the nonparametric EM (npEM) algorithm of Benaglia et al. (2008). This method is a generalization of the Silverman's rule of thumb used to select a bandwidth in kernel density estimation, and it results in one bandwidth for each mixture component and each block of conditionally independent and identically distributed repeated measures.