Complexity analysis from EEG data in congestive heart failure: A study via approximate entropy

Abstract

Aim Congestive heart failure (CHF) is a very complex clinical syndrome that may lead to ischemic cerebral hypoxia condition. The aim of the present study is to analyze the effects of CHF on brain activity through electroencephalographic (EEG) complexity measures, like approximate entropy (ApEn). Methods Twenty patients with CHF and 18 healthy elderly people were recruited. ApEn values were evaluated in the total spectrum (0.2–47 Hz) and main EEG frequency bands: delta (2–4 Hz), theta (4–8 Hz), alpha 1 (8–11 Hz), alpha 2 (11–13 Hz), beta 1 (13–20 Hz), beta 2 (20–30 Hz), and gamma (30–45 Hz) to identify differences between CHF group and control. Moreover, a correlation analysis was performed between ApEn parameters and clinical data (i.e., B‐type natriuretic peptides (BNP), New York Heart Association (NYHA), and systolic blood pressure (SBP)) within the CHF group. Results Statistical topographic maps showed statistically significant differences between the two groups in the total spectrum and theta frequency band. Within the CHF group, significant negative correlations were found between total ApEn and BNP in O2 channel and between theta ApEn and NYHA scores in Fp1, Fp2, and Fz channels; instead, a significant positive correlation was found between theta ApEn and SBP in C3 channel and a nearly significant positive correlation was obtained between theta ApEn and SBP in F4 channel. Conclusion EEG abnormalities in CHF are very similar to those observed in cognitive‐impaired patients, suggesting analogies between the effects of neurodegeneration and brain chronic hypovolaemia due to heart disorder and underlying high brain sensitivity to CHF.

Brain complexity in stroke recovery after bihemispheric transcranial direct current stimulation in mice

“…In this study, we employed the measure of Approximate Entropy (ApEn), that has many advantages: it maintains good reproducibility when used with time series, it is almost unaffected by noise, and it detects changes in underlying episodic behaviour undetected by peak occurrences or amplitudes. 36 , 58 , 59 ApEn values were computed, in each mouse, for each channel and for each epoch using a homemade MATLAB software. For each LFP recording, those values were averaged among epochs to obtain a single ApEn value for each channel.…”