Aim 1: Determining the Generalisability of Electroencephalogram Signatures of Consciousness
Significance
At present, we are unable to accurately track consciousness under anaesthesia. Leveraging our data and similar data collected by collaborators, we will test the generalisability of our recent EEG “fingerprints” of conscious state. Reproducibility in science is a key issue, therefore we will publicly test our discoveries in these important analyses. We will then extend these analyses to identify if superior correlates of consciousness may be identified.
Hypotheses
(1) That EEG signatures of conscious states of sensory disconnection and unconsciousness (identified by Casey et al), will be readily generalisable across another dataset (provided by Prof Sheinin, Turku, Finland) and hence accurately predict conscious state using machine learning models; (2) that addition of EEG complexity and connectivity measures will enhance the classification performance of Support Vector Machine models; and (3) models in sensor space (electrode level) using reduced numbers of electrodes will identify conscious state similarly to source space.
Objective
To establish an accurate and generalisable EEG classifier to detect conscious state.
Method
We will use combinations of our established EEG methods and machine learning approaches to determine classifiers of conscious state.
Benefits
Current depth of anaesthesia monitors cannot accurately identify conscious states under anaesthesia. Identification of specific signatures of conscious state paves the way for a new generation of depth of anaesthesia monitor. Understanding the neuronal correlates of conscious state addresses one of the most important scientific questions with broad implications for anaesthesia, neurology, psychiatry, sleep science and neuroscience.
Aim 2: Determine the Electroencephalogram Signatures of Connected Consciousness after intubation
Significance: Approximately 11% of young adults experience connected consciousness after intubation. We will identify EEG correlates of this state. Combined with aim 1, this will enhance depth of anaesthesia monitoring.
Hypothesis
That high frequency-band permutation entropy will discriminate responders and non-responders on the isolated forearm test in data collected in ConsCIOUS2.
Objective
To establish an accurate and generalisable EEG marker of isolated forearm test responsiveness.
Method
Using data from 296 patients from ConsCIOUS2 we will identify if beta band permutation entropy may predict IFT responsiveness. Secondary outcomes will include additional EEG metrics.
Benefits: Improved depth of anaesthesia monitors.
Aim 3: Determining the Mechanisms of Sensory Disconnection Based on the Predictive Coding Framework
Significance
A monitor based on sound theoretical rationale and evidence that directly probes the mechanisms of sensory perception is necessary to advance anaesthesia delivery. Using the Predictive Coding framework of sensory perception we will test whether the mismatch between sensory stimuli and predictions (assessed using the evoked response potential [ERP] as a marker of prediction error) may be a marker of sensory disconnection.
Hypotheses
(1) Sensory disconnection is associated with reduced associations of sensory prediction error and the ERP across ketamine, dexmedetomidine and propofol sedation and (2) source reconstruction will identify important sources in the auditory “where pathway” that are no longer activated in states of sensory disconnection.
Objective
To establish (1) a theoretically-driven marker of sensory disconnection and (2) to identify potential brain sources that show anatomical loci for where sensory disconnection may occur. Method: We will analyse the evoked response in the EEG to identify if differences may differentiate conscious state across of disconnection induced by anaesthetic drugs with different molecular actions: dexmedetomidine, propofol and ketamine.
Benefits
We will address the critical knowledge gaps - there are no markers of sensory disconnection and limited knowledge about mechanisms about how the brain becomes disconnected from the sensory environment.
Professor Robert Sanders, University of Sydney, NSW;
Professor Jamie Sleigh, Dr Amy Gaskell
University of Auckland, Waikato Hospital, New Zealand.
The project was awarded A$69,032 funding through the ANZCA research grants program for 2024.