Determining day-to-day human variation in indirect calorimetry using Bayesian decision theory.

NEW FINDINGS: What is the central question of this study? We sought to understand the day-to-day variability of human indirect calorimetry during rest and exercise. Previous work has been unable to separate human day-to-day variability from measurement error and within-trial human variability. We developed models accounting for different levels of human- and machine-level variance and compared the probability density functions using total variation distance. What is the main finding and its importance? After accounting for multiple levels of variance, the average human day-to-day variability of minute ventilation, CO2 output and O2 uptake is 4.0, 1.8 and 2.0%, respectively. This is a new method to understand human variability and directly enhances our understanding of human variance during indirect calorimetry. ABSTRACT: One of the challenges of precision medicine is understanding when serial measurements taken across days are quantifiably different from each other. Previous work examining gas exchange measured by indirect calorimetry has been unable to separate differential measurement error, within-trial human variance and day-to-day human variance effectively in order to ascertain how variable humans are across testing sessions. We used previously published reliability data to construct models of indirect calorimetry variance and compare these models with methods arising from Bayesian decision theory. These models are conditional on the data upon which they are derived and assume that errors conform to a truncated normal distribution. A serial analysis of modelled probability density functions demonstrated that the average human day-to-day variance in minute ventilation ( V ̇ E ), carbon dioxide output ( V ̇ C O 2 ) and oxygen uptake ( V ̇ O 2 ) was 4.0, 1.8 and 2.0%, respectively. However, the average day-to-day variability masked a wide range of non-linear variance across flow rates, particularly for V ̇ E . This is the first report isolating day-to-day human variability in indirect calorimetry gas exchange from other sources of variability. This method can be extended to other physiological tools, and an extension of this work facilitates a statistical tool to examine within-trial V ̇ O 2 differences, available in a graphical user interface.

Spectral Transfer Learning Using Information Geometry for a User-Independent Brain-Computer Interface.

Recent advances in signal processing and machine learning techniques have enabled the application of Brain-Computer Interface (BCI) technologies to fields such as medicine, industry, and recreation; however, BCIs still suffer from the requirement of frequent calibration sessions due to the intra- and inter-individual variability of brain-signals, which makes calibration suppression through transfer learning an area of increasing interest for the development of practical BCI systems. In this paper, we present an unsupervised transfer method (spectral transfer using information geometry, STIG), which ranks and combines unlabeled predictions from an ensemble of information geometry classifiers built on data from individual training subjects. The STIG method is validated in both off-line and real-time feedback analysis during a rapid serial visual presentation task (RSVP). For detection of single-trial, event-related potentials (ERPs), the proposed method can significantly outperform existing calibration-free techniques as well as outperform traditional within-subject calibration techniques when limited data is available. This method demonstrates that unsupervised transfer learning for single-trial detection in ERP-based BCIs can be achieved without the requirement of costly training data, representing a step-forward in the overall goal of achieving a practical user-independent BCI system.