Continuous monitoring of Cerebral blood flow Autoregulation and Derived Individualised Pressure Management Targets in Traumatic Brain Injury Patients

Summary

Cerebral autoregulation is one of the brain's key protective mechanisms. It acts to maintain an adequate cerebral blood supply despite fluctuations in arterial pressure and cerebral perfusion pressure — the pressure gradient that drives blood through the brain. This protection is particularly important because neurons are highly vulnerable to even relatively short periods of inadequate oxygen and glucose delivery.

In severe traumatic brain injury (TBI), cerebral autoregulation is frequently disturbed. During the first days after injury, complex biochemical and inflammatory processes can lead to brain swelling and increased intracranial pressure, reducing cerebral perfusion pressure and placing the injured brain at further risk of ischaemia. When autoregulatory mechanisms are impaired, changes in blood pressure may be transmitted directly to the cerebral circulation, potentially contributing to secondary brain injury.

Continuous monitoring of cerebral autoregulation offers a way of assessing this vulnerability at the bedside. Several methods have already shown promise for identifying patient-specific ranges of arterial or cerebral perfusion pressure associated with better preserved autoregulation, and therefore for supporting more individualised management of patients with acute brain injury. However, current techniques remain imperfect. They rely on assumptions about the underlying physiology, can be sensitive to signal quality, time-scale and analytical choices, and may produce conflicting estimates of autoregulatory status. These limitations remain important barriers to wider clinical adoption.

Project aims

The aim of this PhD project is to develop more robust and clinically useful methods for continuous assessment of cerebral autoregulation and their derived targets for individualised blood pressure/cerebral perfusion pressure management. The student will investigate the strengths and limitations of existing techniques and develop new metrics using a spectrum of approaches including physiological signal processing, time-series analysis, probabilistic modelling and machine learning. Particular emphasis will be placed on understanding when different methods succeed or fail, how uncertainty in autoregulatory estimates can be quantified, and how complementary information from multiple physiological signals can be combined.

The project will make extensive use of the Brain Physics Laboratory's large archive of high-resolution multimodal recordings from patients with severe TBI, together with relevant datasets from international collaborators. These data provide continuous measurements of variables such as arterial blood pressure, intracranial pressure, cerebral perfusion pressure, cerebral blood-flow velocity and cerebral oxygenation, allowing candidate algorithms to be evaluated across a wide range of physiological and clinical conditions.

A key translational component of the project will be the practical implementation of successful methods within ICM+, the Brain Physics Laboratory's clinical research neuromonitoring platform. This will allow new algorithms to be tested on continuously streamed physiological data and incorporated into protocols for individualised haemodynamic management in intensive care and potentially in the operating theatre.

The broader objective is to move cerebral autoregulation monitoring from a promising research methodology towards a reliable, interpretable and clinically actionable bedside tool for personalised management of acute brain injury.

Suitable background

This project would suit graduates in Biomedical Engineering, Electrical/Electronic Engineering, Signal Processing, Computer Science, Applied Mathematics or related quantitative disciplines, as well as medically qualified graduates with a strong interest in quantitative physiology and data analysis. Experience in Python, time-series analysis, machine learning or physiological signal processing would be useful but is not essential.

Contact details

Dr Peter Smielewski - ps10011@cam.ac.uk

Opportunities

This project is open to applicants who want to do a:

• PhD