Measurement of brain activation in anaesthetised children using near-infrared spectroscopy

Submission note: A thesis submitted to the School of Engineering and Mathematical Sciences, Faculty of Science, Technology and Engineering, in total fulfilment of the requirements for the degree of Doctor of Philosophy, La Trobe University, Bundoora.

The aim of this thesis is to investigate brain activation in paediatrics using Near-Infrared Spectroscopy (NIRS) with a view of using this technique to develop a Depth of Anaesthesia measure. General anaesthesia is vital to modern healthcare. Patients can undergo complex surgical procedures without the sensation of pain or formation of memory. The patients may also be given paralytic drugs to reduce autonomic reactions to noxious stimuli. This allows lower doses of anaesthetic drugs to be administered, with advantages including reduced recovery times and minimised side effects. Problems arise when insufficient anaesthesia is given to a patient. Under this circumstance, the patient may regain consciousness during the surgical procedure, yet remain paralysed. Recently several electroencephalogram (EEG)-based commercial devices, intended to measure depth of anaesthesia have become available commercially. However, there is much debate in literature about their validity, and moreover these EEG-based devices are not suited for paediatric applications. NIRS is a non-invasive technique used to measure haemoglobin concentrations in cerebral blood. The equipment is portable and low-cost, putting it on par with existing EEG devices. With the aid of flow-metabolism coupling (haemodynamic coupling) the NIRS technique allows measurement of brain activation in a manner similar to evoked potential measurement with EEG. The NIRS approach is superior to EEG in many ways, particularly in its ability to be quantified. The investigations set out in this thesis culminated in a pilot study to determine if visually evoked haemodynamic responses in anaesthetised children can be detected using NIRS. Twenty-three children (between the ages of two and five years) were recruited into the study. These children were undergoing routine day surgery at the Royal Children’s Hospital (Melbourne, Australia). All children were induced with either propofol or sevoflurane and maintained with either isoflurane or sevoflurane. Premedication and caudal or regional blocks were common. Measurement optodes and visual stimulation goggles were placed on the patient postinduction. The protocol was to record five minutes of baseline NIRS data without visual stimulation and a further 25 minutes of recording with stimulation. All recordings were made during the maintenance phase of anaesthesia. The pilot study demonstrated that the NIRS technique was capable of detecting an evoked response in nine of 20 children (three exclusions). Additionally, the analgesic drug fentanyl diminished the response. A response could not be detected in eight of the ten children who were administered fentanyl. Further work proved to be difficult due to several limitations of the technique. There were major problems with obtaining NIRS signals from the scalp due to interference from hair pigment, even though restrictions were in place to exclude children with thick dark hair. Furthermore, several evoked response epochs needed to be averaged to remove noise and artefact for this to succeed a minimum recording of seven minutes was required, thus limiting the NIRS technique accordingly. In order for this technique to be applicable to DoA monitoring, technological advances (such as smaller probes) to reduce interference from hair, or more robust signal processing techniques, must be developed.