Compass Pathways publishes AI technologies study results in Pyschopharmacology

Compass Pathways announced the publication of a new paper in Pyschopharmacology, which showed the results of a retrospective study that evaluates the potential of its AI technologies to support investigational COMP360 psilocybin treatment in treatment-resistant depression, or TRD. The paper was published on August 22. The peer-reviewed paper shows the AI model has the potential to predict outcomes of the investigational COMP360 psilocybin treatment in people with TRD up to 12 weeks after treatment using recordings taken during the integration session 24 hours post administration. Further research and studies are required to validate this model and to evaluate its capabilities to predict responder status pre-treatment. Using a large language model, Compass created a machine learning algorithm that assessed the emotional sentiment of 101 English speaking participants from the company’s recent randomized double-blind phase 2b study in TRD. The sentiment scores, as measured by valence and arousal, were then used in the second stage of the machine algorithm to predict patient response as measured by the Montgomery-Asberg Depression Rating Scale, or MADRS, at three weeks and three months post-drug administration. This model yielded a cross-validated accuracy of 85% and 88% at the three and twelve-week timepoints, respectively. The Area Under the Curve results were comparable with values of 88% and 85% at the three and twelve-week timepoints.