Psilocybin microdosing does not affect emotion-related symptoms and processing: A preregistered field and lab-based study. Josephine Marschall et al. Journal of Psychopharmacology, December 17, 2021. https://doi.org/10.1177/02698811211050556
Abstract
Background: Microdoses of psychedelics (i.e. a sub-hallucinogenic dose taken every third day) can reduce symptoms of depression, anxiety and stress according to anecdotal reports and observational studies. Research with medium to high doses of psilocybin points towards potential underlying mechanisms, including the modulation of emotion and interoceptive processing.
Aims: In this preregistered study, we investigated whether psilocybin microdoses alter self-reported interoceptive awareness and whether repeated microdosing over 3 weeks modulates emotion processing and reduces symptoms of anxiety and depression.
Methods: We used a double-blind, placebo-controlled, within-subject crossover design. Participants completed the Multidimensional Assessment of Interoceptive Awareness Questionnaire 1½ h after self-administering their second dose (or placebo), and the emotional go/no-go task and the shortened Depression Anxiety Stress Scale 1½ h after self-administering their seventh dose.
Results: Our confirmatory analyses revealed that psilocybin microdosing did not affect emotion processing or symptoms of anxiety and depression compared with placebo. Our exploratory analyses revealed that psilocybin microdosing did not affect self-reported interoceptive awareness, that symptoms of depression and stress were significantly reduced in the first block compared with baseline, that participants broke blind in the second block and that there was no effect of expectations. Further research in a substance-naïve population with clinical range anxiety and depressive symptoms is needed to substantiate the potential beneficial effects of microdosing.
Keywords: Psilocybin, microdosing, psychedelics, emotion processing, interoceptive awareness, anxiety, depression, symptoms
We conducted this study to investigate the effect of repeated psilocybin microdosing on depression and anxiety symptoms, emotion processing and interoceptive awareness. We hypothesized that psilocybin microdosing would reduce symptoms of anxiety and depression, increase the processing time needed to identify negative emotions and increase interoceptive awareness. Our results suggest that the psilocybin microdose did not affect interoceptive awareness, but that there was an effect of the block-order variable. In block 1, the psilocybin-first block-order group had a lower average on two subscales scores of the MAIA compared with the Placebo-first block-order group, but this difference was no longer significant in block 2. We explored the possibility that the psilocybin-first block-order group’s interoceptive awareness increased due to repeated psilocybin microdosing after block 1 and therefore achieved higher scores in block 2 which were more similar to the placebo-first group. However, although there was an increase in average score, this change was not significant. It could well be that the block-order groups already differed in interoceptive awareness at baseline and became more similar in their scores over time regardless of condition assignment – however as we did not assess baseline scores on interoceptive awareness, this possibility cannot be verified.
The effect of repeated microdosing on emotion processing, as measured using an emotion go/no-go task, and symptoms of anxiety, depression and stress also did not differ from placebo. Our finding that psilocybin microdosing does not affect symptoms of anxiety and depression contradicts previous survey studies which reported marked reductions in negative emotionality following the repeated microdosing of psychedelic substances (Anderson et al., 2019; Johnstad, 2018; Polito and Stevenson, 2019). This discrepancy may be due to four key elements in our method, including the mental well-being of our participants at baseline, the use of psilocybin only, the duration of the microdosing period and the inclusion of a placebo condition.
First, our participants were only admitted if our pre-trial screening deemed them as physically and mentally healthy, and their symptoms of anxiety and depression at baseline were within the normal range on average. Johnstad (2018) and Anderson et al. (2019) did not include this criterion within their designs, allowing for participants with clinical range symptoms at baseline. This creates the possibility that their significant reductions in negative emotionality were in part due to higher negative emotionality at baseline, whereas our participants may have experienced a ceiling effect; they were already mentally healthy prior to microdosing and could not show further improvement during the study. However, this argument is countered by Polito and Stevenson (2019) who did explicitly focus on a non-clinical population. They reported low DASS-21 scores at baseline, yet found marked reductions in depression and stress scores. Thus, although the baseline DASS-21 scores of our participants were in the same range as those of Polito and Stevenson (2019), we failed to see additional improvements in our participants’ depression scores after microdosing. However, Polito and Stevenson (2019) did not find a reduction in anxiety scores after microdosing, which does align with our findings. Moreover, post hoc analyses of the DASS-21 subscale scores in comparison with baseline scores did reveal significant reductions in the stress and depression subscales in block 1, but regardless of condition.
Second, previous results demonstrated by Johnstad (2018), Anderson et al. (2019) and Polito and Stevenson (2019) were based on psychedelic microdoses in general, while we chose to focus on specifically psilocybin. It is thus possible that previous results were driven by the effects of psychedelic substances other than psilocybin. However, Bershad and colleagues (2019) investigated the effect of three different microdoses of LSD (6.5, 13 and 25 μg) and also did not find a significant effect of these doses on emotion processing nor on negative emotionality. Relatedly, in our study, we had little control over the specific amount of psilocybin that participants consumed, due to natural variability in different batches of psilocybin-containing truffles. Next to that, it is possible that we also manipulated other active compounds found in the psilocybin-containing truffles and that these influenced our results.
Third, our participants consumed the microdoses for a shorter duration (3 weeks) compared with those in the research by Polito and Stevenson (2019; 6 weeks) and likely in the research by Johnstad (2018) and Anderson et al. (2019), although here the specific duration of microdosing was not reported. While the appropriate time necessary for the benefits of microdosing to take effect is unknown, it is known that serotonergic antidepressants can take up to 2 months before measurable effects arise (Harmer et al., 2009), potentially because their effects are due to certain downstream changes in brain structure and function (Erb et al., 2016; Hanson et al., 2011). The argument that effects of microdoses may also require a longer period of repeated dosing rests on two key findings: that depression and stress-related disorders are associated with neural atrophy in the prefrontal cortex (PFC; Christoffel et al., 2011) and that serotonergic psychedelics can increase structural and functional plasticity in the PFC (Ly et al., 2018; Olson, 2018), thereby potentially counteracting the neurobiological markers of these disorders. It is possible that a period of consistent microdosing which succeeds 3 weeks is required for such changes to develop and we can expect an effect on emotion processing and mood-related symptoms only after these changes have occurred.
Nevertheless, Cameron et al. (2019) administered microdoses of the serotonergic psychedelic DMT to rats every third day for 7 weeks and revealed no markers of increased neural plasticity. In fact, the researchers found a decrease in dendritic spine density in PFC of female rats. Important to note is that the original association between serotonergic psychedelics and neuronal plasticity is based on the effect of a single large serotonergic psychedelic dose. Single large doses of DMT and LSD were found to promote spinogenesis, synaptogenesis and neural plasticity in cortical neuron cultures of rats 24 h after administration (Ly et al., 2018). Taken together, this evidence, although limited in its generalizability to humans, may indicate that regardless of the duration of the dosing period, psilocybin microdoses are simply not potent enough to trigger structural changes in the cortex.
Another factor that could account for the apparent conflict with previous findings is our implementation of a placebo-control group. It is plausible that previous results were driven by participants’ expectations rather than the chemical components of the doses. This would mean that the sole act of taking doses improved participants’ mental health scores, regardless of whether the doses were placebos or psilocybin microdoses. Indeed, Kaertner et al. (2021) found that positive expectancy scores at baseline predicted changes in well-being after 4 weeks of microdosing. In our participants, we observed an overall decrease in the depression and stress scores from baseline to block 1, irrespective of the condition that the participants were assigned to. This effect, termed the ‘placebo effect’, is especially relevant in clinical and pharmacological research and refers to the situation when blinded participants in the placebo condition experience a reduction in symptoms either due to their positive expectations towards the treatment condition or due to previous conditioning of the treatment condition (Meissner et al., 2011). Especially in antidepressant research, placebo doses evoke reductions in symptoms comparable with the antidepressant (Kirsch, 2014). Such potent placebo responses may also be contributing to previous findings regarding the effects of psilocybin microdosing and make it difficult to assess whether psilocybin microdosing is effective beyond expectations and conditioning.
Of relevance, through an exploratory analysis of our participants’ condition guesses, we found that participants broke blind regarding their condition in the second block. This confound had the potential to further contribute to response expectancy effects. However, we found no difference between psilocybin and placebo conditions in our outcome measures in either block, which indicates that explicit expectations likely did not influence our results. Moreover, in contrast to Kaertner et al. (2021), we found no effect of expectation in further post hoc exploratory analyses. We propose two possible reasons for this lack of an effect of psilocybin microdosing on outcome measures: either the placebos and psilocybin microdoses were equally ineffective at influencing a change in response to the scales that we used (potentially the measures were not sensitive enough), or the placebo effect was equally as effective as the microdosing effect but was guided by processes other than explicit expectations, such as previous conditioning (e.g. as demonstrated by Amanzio and Benedetti, 1999). Previous experience with psychedelic substances could evoke a placebo effect based on conditioning. Most participants in this study had taken psychedelics before and may therefore have been subject to such conditioning. Thus, including a placebo condition may have dampened our effect of interest.
Finally, we need to consider the possibility that microdosing does not affect depression and anxiety at all, as our findings consistently indicate. Previously reported beneficial effects may be related to other confounding factors, as mentioned above, and experimental research thus far fails to show these hypothesized effects of microdosing on clinically relevant outcome measures (Bershad et al., 2019; Family et al., 2020).
We note five key limitations of our study. First, our sample suffers from selection bias, since participants were self-selected from a microdosing workshop. As a result, most of our participants had tried psychedelics previously, which means that they may have broken blind easier or may have been desensitized to the microdosing effects. Second, the psilocybin doses were made by the participants using dried psilocybin truffles, meaning that we cannot be sure of the exact amounts of psilocybin in the individual doses that the participants consumed. It is possible that the degree of psilocybin content varied across participants and thereby obscured our results. Third, we encountered a large drop-out rate during this project and several participants did not sufficiently comply with the behavioural guidelines to be included in the analyses. This resulted in small sample size relative to existent observational studies and in a further selection bias (i.e. only motivated participants likely stayed in). Moreover, due to such sample size, our study may have been underpowered to detect true effects, particularly the interaction effect hypothesized for the emotional go/no-go task. Our post hoc power analysis suggested that our design, given our observed data, was insufficiently powered to detect this effect. Simulated data in the hypothesized direction, however, yielded sufficient power with a large effect size. Of course, as noted earlier, this analysis based on simulated data remains speculative and we encourage future studies to plan their sample size according to expected RT patterns. Fourth, we measured the effects in our study only after self-administration of a dose, and not between doses or after each block. Thus, our results may be confounded by the acute effect of the psilocybin dose, which may differ from its persistent effect after the acute chemical-induced symptoms have subsided. However, Szigeti et al. (2021) did assess both acute and post-acute effects and found no significant microdose vs placebo differences in psychological outcomes when accounting for participants breaking blind. Last, our study is a combined field and lab-based study, meaning that the results may not be readily generalizable or replicable, for example, in a more clinical setting.