Related Experiment Video
Updated: Dec 18, 2025
Assessing Cerebral Autoregulation via Oscillatory Lower Body Negative Pressure and Projection Pursuit Regression
Published on: December 10, 2014
Assessment of cerebral autoregulation indices - a modelling perspective
Xiuyun Liu1,2, Marek Czosnyka3,4, Joseph Donnelly3,5
1Brain Physics Laboratory, Division of Neurosurgery, Department of Clinical Neurosciences, Addenbrooke's Hospital, University of Cambridge, Cambridge, UK. liuxiuyun1@gmail.com.
Abstract:
Various methodologies to assess cerebral autoregulation (CA) have been developed, including model - based methods (e.g. autoregulation index, ARI), correlation coefficient - based methods (e.g. mean flow index, Mx), and frequency domain - based methods (e.g. transfer function analysis, TF). Our understanding of relationships among CA indices remains limited, partly due to disagreement of different studies by using real physiological signals, which introduce confounding factors. The influence of exogenous noise on CA parameters needs further investigation. Using a set of artificial cerebral blood flow velocities (CBFV) generated from a well-known CA model, this study aims to cross-validate the relationship among CA indices in a more controlled environment. Real arterial blood pressure (ABP) measurements from 34 traumatic brain injury patients were applied to create artificial CBFVs. Each ABP recording was used to create 10 CBFVs corresponding to 10 CA levels (ARI from 0 to 9). Mx, TF phase, gain and coherence in low frequency (LF) and very low frequency (VLF) were calculated. The influence of exogenous noise was investigated by adding three levels of colored noise to the artificial CBFVs. The result showed a significant negative relationship between Mx and ARI (r = -0.95, p < 0.001), and it became almost purely linear when ARI is between 3 to 6. For transfer function parameters, ARI positively related with phase (r = 0.99 at VLF and 0.93 at LF, p < 0.001) and negatively related with gain_VLF(r = -0.98, p < 0.001). Exogenous noise changed the actual values of the CA parameters and increased the standard deviation. Our results show that different methods can lead to poor correlation between some of the autoregulation parameters even under well controlled situations, undisturbed by unknown confounding factors. They also highlighted the importance of exogenous noise, showing that even the same CA value might correspond to different CA levels under different 'noise' conditions.
