A virtual clinical trial of psychedelics to treat patients with disorders of consciousness

Researchers from the University of Liège and international collaborators developed a “virtual clinical trial” exploring a unique pharmacological treatment in patients who do not fully regain consciousness after a coma. The proposed treatment involves employing psychedelic drugs, such as psilocybin and lysergic acid diethylamide (LSD) that have intense, consciousness-altering effects in healthy volunteers. Those remarkable experiences are thought to be linked with increased dynamical complexity of brain activity.  In patients with disorders of consciousness (DoC), dynamic brain function is importantly reduced, potentially resulting in decreased consciousness. Therefore, the proposed treatment effects hinge upon the premise that a short-term increase in the complexity of brain activity in patients could result in improvements in the conscious state of these patients.

Performing clinical trials with psychedelic drugs is burdened with legal obstacles and necessitates careful planning and ethical considerations. However, in the study by Alnagger et al., published in Advanced Science, the authors innovatively utilised personalised computational models of patients’ brains, built from individual MRI scans (functional MRI and diffusion-weighted imaging). After constructing a computational model of each patient’s brain, they simulated the effects of LSD and psilocybin as would be performed in a clinical trial, yet virtually. The researchers demonstrated that these substances could shift brain activity in patiens with DoC towards healthier, more flexible and complex dynamics.

To study the dynamics of simulated brain activity, the authors observed how an artificial perturbation was integrated into the brain’s activity in time and magnitude. “In order to understand any dynamical system, in this case the brain, it’s often useful to perturb it” said Naji Alnagger, first author, PhD candidate of the Coma Science Group. “Imagine if you want to understand how viscous a liquid is, in other words, its dynamics, you could poke it. By observing how it reacts to the perturbation, how long it takes to return to baseline and the degree it is disturbed could reveal a lot of information as to the viscosity. The same is true here, observing how the model reacts after introducing an artificial perturbation can tell us about the nature of dynamics of the brain activity.”

The researchers first validated this method of simulated perturbations by showing that the response to perturbation was lower in states of lower consciousness such as DoC patients and healthy participants under anaesthesia compared to normal waking consciousness. They then showed that under psychedelics, perturbation induces an even greater response than in healthy individuals without the drug.

The researchers found that simulating LSD and psilocybin on patients with DoC boosted their responses to perturbations. The effect was greater in patients with some minimal signs of non-reflex behaviour (minimally conscious state) compared to patients with no detectable signs of clinical awareness (unresponsive wakefulness syndrome). Stratifying by these two diagnostic groups, the researchers also found that the simulated treatment effect was more correlated with the strength of the brain structural connectivity in the UWS patients, and with the strength of the brain functional connectivity in the MCS patients. “What was particularly noteworthy was that the MCS and UWS patients have the same level of structural connectivity strength, what differed was the efficiency of the existing connectivity, therefore it seems that, not only connectivity strength but efficient organisation was critical in sustaining the psychedelic related shifts in dynamics” said Dr. Jitka Annen, senior author and researcher at the University of Ghent.

“The results of the study suggest that if this treatment was to be carried out in real clinical settings, the type of patients who would most likely benefit would be an MCS patient with a strongly connected functional connectivity. This provides a starting point to target potential future clinical trials.” said Dr. Olivia Gosseries, director of the Coma Science Group.

Computational models offer a way to test treatments in silico, identify good potential candidates and investigate the mechanisms of a particular treatment. “This is an important proof of concept,” said Naji Alnagger. “While much more work is needed before any form of clinical application, this study shows how personalised computational modelling could one-day form a valuable part of pre-clinical work and personalised medicine.”

*The study was performed in collaboration with the University hospital of Liège (CHU de Liège), Imperial College London, Maastricht University, Sorbonne University, Universitat Pompeu Fabra, and other international partners.