Search

Recently, time delay stability analysis of biosignals has been successfully applied as a multivariate time series analysis method to assess the human physiological network in young adults. The degree of connectivity between different network nodes is described by the so-called link strength. Based on polysomnographic recordings (PSGs), it could be shown that the network changes with the sleep stage. Here, we apply the method to a large set of healthy controls spanning six decades of age. As it is well known, that the overall sleep architecture is dependent both on age and on gender, we particularly address the question, if these changes are also found in the network dynamics. We find moderate dependencies of the network on gender. Significantly higher link strengths up to 13% are found in women for some links in different frequency bands of central and occipital regions in REM and light sleep (N2). Higher link strengths are found in men consistently in cardio-cerebral links in N2, but not significant. Age dependency is more pronounced. In particular a significant overall weakening of the network with age is found for wakefulness and non-REM sleep stages. The largest overall decrease is observed in N2 with 0.017 per decade. For individual links decrease rates up to 0.08 per decade are found, in particular for intra-brain links in non-REM sleep. Many of them show a significant decrease with age. Non-linear regression employing an artificial neural network can predict the age with a mean absolute error (MAE) of about five years, suggesting that an age-resolution of about a decade would be appropriate in normative data for physiological networks.

Objective Physiological networks (PN) model couplings between organs in a high-dimensional parameter space. Machine learning methods, in particular artifical neural networks (ANNs), are powerful on high-dimensional classification tasks. However, lack of interpretability of the resulting models has been a drawback in research. We assess relevant PN topology changes in obstructive sleep apnea (OSA) by novel ANN interpretation techniques. Approach ANNs are trained to classify OSA based on the PNs of 48 patients and 48 age and gender matched healthy controls. The PNs consisting of 2812 links are derived from overnight biosignal recordings. The interpretation technique DeepLift is applied to the resulting ANN models, enabling the determination of the relevant features for classification decisions on individual subjects. The mean relevance scores of the features are compared to other machine learning methods (decision tree and random forests) and statistical tests on group differences. Main results The ANN interpretation results show good agreement with the compared methods and 87% of the samples could be correctly classified. OSA patients show a significantly higher coupling (p [Formula: see text] 0.001) in light sleep (N2) between breathing rate and EEG [Formula: see text] power in all electrode locations and to chin and leg muscular tone. In deep sleep (N3), OSA leads to significantly lower coupling (p [Formula: see text] 0.01) in lateral connections of EEG [Formula: see text] and [Formula: see text] power in central and frontal positions. Misclassified OSA patients had all mild/moderate AHIs and did not show PN topology changes. Both nights of these patients have been consistently misclassified as healthy. This may indicate, that the impact of respiratory events differs in subjects, thus forming different phenotypes. Significance The proposed PN analysis provides a powerful and robust method to quantify a broad range of physiological interactions. Interpretability of the ANN make them a promising tool to identify new diagnostic markers in data-driven approaches.