Thesis defense Yvonne Blokland (Donders series 200)

26 November 2015
Promotor: prof. dr. G.J. Scheffer, copromotor: dr. J. Farquhar
Moving towards awareness detection: From Brain-Computer Interfacing to anaesthesia monitoring

Detecting unintended awareness during surgery remains one of the biggest challenges in anaesthesia research and clinical practice. General anaesthesia involves the simultaneous administration of different components including a neuromuscular blocker for immobilization. If complete paralysis but not unconsciousness has been established, patients attempting to move find themselves unable to communicate their situation to the surgeon or anaesthetist. In this thesis a novel monitor of intraoperative awareness is proposed, based on detection of these movement attempts.

This paradigm follows the principles and techniques from the field of Brain-Computer Interfacing. A Brain-Computer Interface (BCI) measures a users brain signal, usually with electroencephalography, and translates this information to commands to drive a device or to enable communication. In this thesis, offline BCI performance was evaluated for healthy volunteers performing movement tasks in several conditions encountered in an intraoperative awareness situation. First, it was shown that using a movement-based paradigm a good accuracy-efficiency tradeoff could be obtained. For this, an EEG setup including 9 channels was sufficient, thus ensuring a relatively short setup time. Second, the influence of temporary paralysis, induced by a neuromuscular blocker (rocuronium), on movement-related EEG signatures was studied. In other words, while subjects attempted to move, actual neuromuscular output was chemically blocked, providing insight in the question how these attempted movements relate to executed and imagined movement. It was shown that the neural responses to attempted and executed movements are very similar, and good BCI performance can be achieved using attempted movement as a control task. Third, the effect on the EEG of another component of general anaesthesia was evaluated: volunteers were sedated by means of the hypnotic agent propofol. While the results demonstrated that the BCI may in principle work in altered states of consciousness, it became clear that sedation may present difficulties with regard to interpretation of the outcome, and thus reliability. Meanwhile, it also provided some interesting insights on the nature of sedation and states of consciousness in general.

While the main part of the thesis focusses on detecting intraoperative awareness, other applications are also addressed. One chapter is devoted to the description of a hybrid EEG-fNIRS BCI in a study with patients with tetraplegia. In these paralysed patients, as in the healthy volunteers receiving a neuromuscular blocker, attempted movements were found to be a better task for BCI control than motor imagery. The work presented in this thesis may contribute towards further development of motor-based BCIs in general and detection of intraoperative awareness in particular.