Granger causality analysis of sleep brain-heart interactions

We studied the networks of Granger causality (GC) between the time series of cardiac vagal autonomic activity and brain wave activities, measured respectively as the normalized high frequency (HF) component of heart rate variability and EEG power in the δ, θ, α, σ, β bands, computed in 10 healthy subjects during sleep. GC analysis was performed by vector autoregressive modeling, and significance of each link in the network was assessed using F-statistics. The whole-night analysis revealed the existence of a fully connected network of brain-heart and brain-brain interactions, with the β EEG power acting as a hub which conveys the largest number of GC links between the heart and brain nodes. These links became progressively more weak when assessed during light sleep, deep sleep, and REM sleep, thus suggesting that brain-heart GC networks are sustained mainly by sleep stage transitions.