Accounting for squared coherence marginally improves the discriminative power of frequency domain cerebral autoregulation markers in surgical aortic valve replacement patients

The dynamic component of the cerebral autoregulation (CA) is characterized via the transfer function (TF) from mean arterial pressure (MAP) to mean cerebral blood velocity (MCBv), acquired noninvasively via transcranial Doppler device. TF analysis theoretically requires the significance of squared coherence (K2). However, the practical relevance of assessing the significance of K2 is unclear. The study compares K2, TF phase (TFP), and TF gain (TFG) assessed via a moving window approach over all the frames regardless of K2 and after exclusion of those frames with K2 below the significance threshold. The approach is based on the generation of uncoupled surrogate pairs that preserve distribution and power spectral density of the original series while being fully uncoupled. MAP and MCBv were recorded in 24 patients (age: 61 ± 14 yrs, 18 males, 6 females) scheduled for surgical aortic valve replacement (SAVR). Data was acquired at supine resting and during active standing the day before the procedure (PRE) and within 7 days after (POST). In SAVR population we found that: (i) the practical relevance of accounting for the significance of K2 to discriminate time points and experimental conditions is marginal especially at time scales typical of CA; (ii) CA is preserved in both PRE and POST and this conclusion holds irrespective of K2 and its significance. Cerebrovascular control is so complex that the concomitant action of multiple regulatory mechanisms might have reduced the efficacy of the application of the criterion based on K2 significance in SAVR population.