| Popis: |
This thesis focuses on the development and use of analysis tools to monitor brain injury patients. For this purpose, an online amperometric analyzer of cerebral microdialysis samples for glucose and lactate has been developed and optimized within the Boutelle group. The initial aim of this thesis was to significantly improve the signal-to-noise ratio and limit of detection of the assay to allow reliable quantification of the analytical data. The first approach was to re-design the electronic instrumentation of the assay. Printed-circuit boards were fabricated and proved very low noise, stable and much smaller than the previous potentiostats. The second approach was to develop generic data processing algorithms to remove three complex types of noise that commonly contaminate analytical signals: spikes, non-stationary ripples and baseline drift. The general strategy consisted in identifying the types of noise, characterising them, and subsequently subtracting them from the otherwise unprocessed data set. Spikes were effectively removed with 96.8% success and ripples were removed with minimal distortion of the signal resulting in an increased signal-to-noise ratio by up to 250%. This allowed reliable quantification of traces from ten patients monitored with the online microdialysis assay. Ninety-six spontaneous metabolic events in response to spreading depolarizations were resolved. These were characterized by a fall in glucose by -32.0 μM and a rise in lactate by +23.1 μM (median values) for over a 20-minute time-period. With frequently repeating events, this led to a progressive depletion of brain glucose. Finally, to improve the temporal coupling between the metabolic data and the electro-cortical signals, a flow-cell was engineered to integrate a potassium selective electrode into the microdialysate flow stream. With good stability over hours of continuous use and a 90% response time of 65 seconds, this flow cell was used for preliminary in vivo experiments the Max Planck Institute in Cologne. |
