Evaluation of Respiratory Muscles Activity by Means of Cross Mutual Information Function at Different Levels of Ventilatory Effort

Abstract

Abstract-Analysis of respiratory muscles activity is an effective technique for the study of pulmonary diseases such as obstructive sleep apnea syndrome (OSAS). Respiratory diseases, especially those associated with changes in the mechanical properties of the respiratory apparatus, are often associated with disruptions of the normally highly coordinated contractions of respiratory muscles. Due to the complexity of the respiratory control, the assessment of OSAS related dysfunctions by linear methods are not sufficient. Therefore, the objective of this study was the detection of diagnostically relevant nonlinear complex respiratory mechanisms. Two aims of this work were: 1) to assess coordination of respiratory muscles contractions through evaluation of interactions between respiratory signals and myographic signals through nonlinear analysis by means of cross mutual information function (CMIF); 2) to differentiate between functioning of respiratory muscles in patients with OSAS and in normal subjects. Electromyographic (EMG) and mechanomyographic (MMG) signals were recorded from three respiratory muscles: genioglossus, sternomastoid and diaphragm. Inspiratory pressure and flow were also acquired. All signals were measured in eight patients with OSAS and eight healthy subjects during an increased respiratory effort while awake. Several variables were defined and calculated from CMIF in order to describe correlation between signals. The results indicate different nonlinear couplings of respiratory muscles in both populations. This effect is progressively more evident at higher levels of respiratory effort.

Interpretation of the auto-mutual information rate of decrease in the context of biomedical signal analysis. Application to electroencephalogram recordings

“…), the estimation of the MI is more robust against outliers and artefacts in data (Alonso et al 2007.…”

“…be a sequence acquired from a stochastic process (Alonso et al 2007). If the amplitude values of the variable ( ) t X are partitioned into I bins, a probability…”

“…It is based on the amplitude distributions of ( ) t X on different time scales, τ (David et al 2004). These amplitude distributions are estimated from histograms (Abásolo et al 2008, Alonso et al 2007, Gómez et al 2007, Jeong et al 2001.…”

“…where {} ⋅ N denotes the cardinality (the number of elements) in the set (Alonso et al 2007, Pompe et al 1998. This transformation replaces the real values of ( ) …”

Approximate entropy and auto mutual information analysis of the electroencephalogram in Alzheimer’s disease patients

“…The AMI is based on the amplitude distributions of x(t) on different time scales [10]. These distributions are estimated from histograms [4], [25]. For a fixed sequence length, if larger sampling bins are used to construct the histograms, the estimations of the average probabilities are more accurate.…”