Characterisation of linear predictability and non-stationarity of subcutaneous glucose profiles

Abstract

Abstract-Continuous glucose monitoring is increasingly used in the management of diabetes.Subcutaneous glucose profiles are characterized by a strong non-stationarity, which limits the application of correlation-spectral analysis. We derived an index of linear predictability by calculating the autocorrelation function of time series increments and applied detrended fluctuation analysis to assess the non-stationarity of the profiles. Time series from volunteers with both type 1 and type 2 diabetes and from control subjects were analysed. The results suggest that in control subjects, blood glucose variation is relatively uncorrelated, and this variation could be modelled as a random walk with no retention of 'memory' of previous values.In diabetes, variation is both greater and smoother, with retention of inter-dependence between neighboring values. Essential components for adequate longer term prediction were identified via a decomposition of time series into a slow trend and responses to external stimuli. Implications for diabetes management are discussed.

Generalised stochastic model for characterisation of subcutaneous glucose time series

“…Please refer to Figure 2 for details. This conclusion supports the results of [9], which shows larger inter-dependence between neighbouring values in diabetes groups compared with the control group. Another way to determine the plausibility of the models is to compare the intensity of the system noise and measurement error, which are shown in Figure 3.…”

“…4) that the mean value of the system noise intensity of Type II is the smallest of the three groups, which is different from the linear model where the control group has the smallest mean value of system noise intensity. This test result confirms the suggestion from [9] that nonlinear models in response to external stimuli for some diabetic patients are better than linear models. As shown in Fig.…”

Estimation of secondary effect parameters in glycaemic dynamics using accumulating data from a virtual type 1 diabetic patient

“…44 Some of these developments also include compensation for stress 45 hyperglycaemia [42]. However, this type of treatment is still exper- 46 imental and has high cost and complexity [17,4]. Hence, it may be 47 more practical to improve upon conventional approaches such as 48 self-monitored glucose with multiple daily insulin injections [46,47].…”

Empirical mode decomposition analysis of near-infrared spectroscopy muscular signals to assess the effect of physical activity in type 2 diabetic patients

“…Due to the nonlinear nature of the NIRS signal, linear, frequency and time-frequency domain methods may not be able to fully extract the small variations from the signals, whereas nonlinear interrelationships in the NIRS data can provide more accurate information. Previous studies have demonstrated how non-stationary and nonlinear methods are needed to analyze glucose levels [20] and insulin sensitivity [21] of diabetic subjects. Also, bispectral analysis has been used for epilepsy diagnosis [22,23], sleep stages [24], cardiac abnormalities [25], EEG signals [26], and myoelectric signals [27].…”