Note: This content is accessible to all versions of every browser. However, this browser does not seem to support current Web standards, preventing the display of our site's design details.

  

System identification of a cerebral blood flow model

Student(en):

Master project
Betreuer:

Tony Wood, Roy Smith, Francesca Parise
Beschreibung:

To satisfy the brainís metabolic demand, cerebral blood flow has to be tightly controlled. While limited fluctuations can be tolerated, both too low and too high blood flow over a prolonged time period can cause tissue death and lead to severe brain damage. This is usually prevented by the brainís internal feedback mechanisms which, depending on blood pressure, concentration of CO2 and cerebral metabolic rate, alter the diameter of cerebral blood vessels, such that the resulting blood flow meets the brainís demand.

demo image: Cerebral blood flow.

In the last 20 years many studies have aimed at investigating the ability of the brain's circulatory system to maintain brain perfusion constant over a wide range of blood pressure. These studies led to the understanding that this process, known as cerebral autoregulation (CA), acts as a high-pass filter which regulates blood flow at frequencies below 0.07 Hz (corresponding to time periods of 10 seconds or more) while letting high frequency changes in blood flow unaffected. Motivated by this, CA was mostly modeled using either open or closed loop linear models and analyzed by estimating the transfer function between blood pressure and blood flow velocity as a measure of blood flow. While these studies provided valuable insights, they also showed that linear models cannot adequately capture the dynamics of CA in the longer time scales where brainís blood supply is mostly regulated. Since these are the time scales of most clinical relevance, the benefit of these studies for patients with traumatic brain injury, subarachnoid hemorrhage, ischemic stroke or acute infectious diseases such as bacterial meningitis, has so far been limited.

This project aims at developing a predictive model of the regulation of cerebral blood flow over long time scales. To this end, we aim at:

  • Expanding the existing linear models and propose different models based on different hypothesized non-linearities.
  • Include the effect of CO2 concentration in the blood (a crucial factor that has for the most part been neglected in earlier modeling studies) and propose several hypotheses for how it enters the regulation process.
  • Identify the parameters of the proposed models from circulation data collected in healthy subjects using closed loop system identification techniques.
  • Time permitting, study the capability of the best identified model to predict measurements collected from patients with CA disorders; in the process we will revisit the models to generate hypotheses for the mechanisms behind the related clinical conditions.

An ideal candidate for this project should
  • possess some prior knowledge on system identification;
  • have a good experience with Matlab;
  • no prior medical knowledge is required, however the candidate should be prone to learn about the theoretical foundations of cerebral blood flow mechanism;
  • be highly motivated and propositive.

A successful theoretical work might possibly lead to a publication.



Weitere Informationen
Professor:

John Lygeros
Projektcharakteristik:

Typ:
Art der Arbeit: 50% theory, 50% simulation
Voraussetzungen:
Anzahl StudentInnen: 1
Status: done
Projektstart: as soon as possible
Semester: