A Language Analysis Study Recorded Microdosing Effects
The practice of microdosing psychedelics has been growing in popularity over the past few years. Many microdosers report experiencing a wide range of benefits from enhanced mood, mindfulness, and creativity to reductions in symptoms of depression and anxiety.
While survey studies have consistently and persistently pointed to microdosing as a therapeutic and beneficial practice, the very few clinical studies aimed to explore the effects of microdosing have reported mixed results with most recent placebo-controlled studies casting doubt on the efficacy and viability of the practice by claiming that evidence suggests microdosing is no better than placebo.
Recently, a group of researchers questioned whether the instruments and methodologies used for assessing the effects of large-dose psychedelic experiences were sufficiently sensitive and specific to capture the subtle changes following a microdosing experience. They aimed to identify and employ more sensitive measurement methodologies capable of capturing shifts that would go unnoticed by the systems used in the past to assess microdosing effects.
The researchers crafted a testing model that aimed to "determine whether unconstrained speech contains signatures capable of identifying the acute effects of psilocybin microdoses." To explore this hypothesis, they adopted a double-blind and placebo-controlled experimental design including 34 healthy adults who were administered either a microdose of psilocybin or a placebo twice a week for a period of time. Their natural speech patterns during dosing time were then captured and analyzed by machine learning models capable of discriminating between variables known to be affected by higher doses: verbosity, semantic variability, and sentiment scores.
The results, published in Nature magazine, showed that with the exception of semantic variability, these metrics presented "significant differences between a typical active microdose and the inactive placebo condition". The classifiers utilized by machine learning were capable of distinguishing between the two conditions with high accuracy suggesting distinct language characteristics between the microdosing and placebo groups.
These results for the first time have captured the effects of microdosing on natural unconstrained speech. This experiment demonstrates the importance of designing study protocols with enough sensitivity to capture the subtle yet potentially beneficial effects of various interventions that often remain under-studies, under-appreciated, and under-utilized due to inappropriate clinical testing systems.
More sensitive testing models and methodologies are necessary to further explore the various reported benefits of microdosing.