A multi-objective optimisation approach for the linear modelling of cerebral autoregulation system

Area of Science:

• Neuroscience
• Biomedical Engineering
• Physiology

Background:

• Dynamic cerebral autoregulation (dCA) maintains stable cerebral blood flow (CF) despite arterial blood pressure (ABP) variations.
• Existing methods for assessing dCA face challenges in accurately discriminating between normal and impaired states.
• Modeling cerebrovascular resistance and compliance is crucial for understanding dCA.

Purpose of the Study:

• To introduce a novel multi-objective optimisation (MO) approach for configuring cerebrovascular resistance-compliance models.
• To enhance the discrimination between normal and impaired dCA using spontaneous ABP and CF fluctuations.
• To investigate the utility of MO in modeling dCA under varying physiological conditions.

Main Methods:

• Collected data from 29 subjects under normocapnic and hypercapnic (5% CO2) conditions.
• Developed cerebrovascular resistance and vessel compliance models with ABP as input and CF velocity as output.
• Employed a MO approach to fit models, optimizing for both Pearson's correlation and error.

Main Results:

• The MO approach yielded superior model configurations compared to single-objective (SO) methods.
• This improvement was particularly notable under hypercapnic conditions.
• The Pareto-optimal front provided novel insights into dCA, highlighting the myogenic mechanism's increased contribution to impairment.

Conclusions:

• The MO approach offers a more effective method for modeling dCA.
• This technique enhances the understanding of impaired dCA, especially during hypercapnia.
• The findings underscore the significant role of the myogenic mechanism in dCA regulation and impairment.