More realistic forecasting of future life events after psilocybin for treatment-resistant depression

Major depressive disorder (MDD) is a major contributor to global disease burden and roughly 30% of patients meet criteria for treatment-resistant depression (TRD). Cognitive accounts of depression emphasise biased negative beliefs about the self, the world and the future, whereas the depressive realism hypothesis argues that depressed individuals may actually perceive reality more accurately. Recent empirical work tends to support the cognitive-bias account, showing that depressed patients exhibit pessimistic biases when forecasting future events. Meanwhile, classical psychedelics such as psilocybin have been reported to produce enduring increases in optimism, psychological wellbeing and trait openness after single doses in healthy volunteers, and early clinical research has shown rapid antidepressant effects in patients with cancer-related and treatment-resistant depression. Neurobiological targets implicated in both depression and the psychedelic state include the serotonin 2A receptor (5-HT2A R) and the default mode network (DMN). The present study aimed to test whether patients with TRD display an unjustified pessimism bias when predicting personal future life events, and whether a composite intervention involving psilocybin plus psychological support could remediate that bias. Using a behavioural forecasting task—the Prediction Of Future Life Events (POFLE)—the investigators compared TRD patients before and after two oral psilocybin dosing sessions (10 mg then 25 mg one week apart) and contrasted any change with that seen in matched, untreated healthy control subjects assessed over the same interval. The primary hypothesis was that TRD patients would show a baseline pessimism bias that would be reduced following psilocybin therapy.

Serafini and colleagues conducted an open-label pilot study with a mixed-model design. The sample comprised 15 clinically assessed TRD patients and 15 healthy control subjects matched on age, gender and education. The patients received two supervised oral doses of psilocybin: an initial safety dose of 10 mg followed one week later by a treatment dose of 25 mg. Controls were recruited via word of mouth and received no intervention. Neither the patients nor the research team were masked to treatment assignment. Ethical and regulatory approvals were obtained and written informed consent was provided by all patients. Depressive symptoms and cognitive forecasting were the main outcomes. Depressive symptoms were measured using the Beck Depression Inventory (BDI), a 21-item self-report scale. Cognitive bias and forecasting accuracy were assessed using the POFLE task, which asks participants to estimate the probability that 40 life events (20 desirable, 20 undesirable) will occur within the next 30 days; the task was split into two equivalent versions (A and B), counterbalanced across participants. Participants provided baseline predictions and then were contacted 30 days later to record which events had actually occurred; predictions were repeated after the treatment interval and again checked 30 days later. Statistical analyses used SPSS v23. Within-group comparisons used paired t-tests or Wilcoxon signed-rank tests as appropriate, and between-group comparisons used independent t-tests or Mann–Whitney U tests. Depressive symptoms over multiple time points were analysed with the nonparametric Friedman test and Dunn's post hoc correction. Correlations used Pearson or Spearman coefficients depending on distribution. Effect sizes were calculated using Hedges' g and 95% confidence intervals were reported. Bias scores were computed on a scale from -1 (maximal pessimism) to 1 (maximal optimism), with 0 indicating no bias; accuracy was quantified as the discrepancy between predicted probabilities and actual event occurrence.

Participant characteristics indicated a predominantly Caucasian sample (93.33%), mostly male (73.33%) and with post-secondary education (80%). At baseline, TRD patients had markedly higher depressive symptom scores on the BDI than controls (patients: M = 34.33, SD = 7.44; controls: M = 3.67, SD = 3.83; p < 0.001, g = 5.0). Patients showed a large and statistically significant reduction in BDI score one week after the second psilocybin session (post-treatment M = 12.13, SD = 9.80) compared with baseline [t(14) = 7.900, p < 0.001; 95% CI (16.17, 28.23), g = 1.9]. No significant BDI change occurred in controls over the same period. Temporal assessment across dosing days showed that BDI scores did not decline naturally prior to treatment (no difference between baseline and dosing day 1), but were significantly reduced by dosing day 2 compared with baseline and dosing day 1. Scores fell further one week after both sessions, indicating progressive improvement following each dose. On the forecasting task at baseline, patients estimated similar probabilities for desirable (M = 0.29, SD = 0.15) and undesirable (M = 0.23, SD = 0.15) events, yet reported experiencing significantly more desirable (M = 4.60, SD = 1.76) than undesirable events (M = 1.40, SD = 1.35) in the ensuing 30 days [t(14) = 5.967, p < 0.001, g = 1.5]. By contrast, controls predicted markedly higher likelihoods for desirable than undesirable events at baseline (desirable M = 0.57, SD = 0.09) and their experienced event counts matched their expectations. Between-group comparisons showed that patients predicted significantly fewer desirable events than controls at baseline [t(28) = 5.820, p < 0.001, g = 2.2], despite similar actual event rates across groups during that period. After psilocybin treatment, patients' predictions shifted: they gave higher probabilities for desirable (M = 0.44, SD = 0.20) than undesirable events and reported more desirable events actually occurring in the subsequent 30 days. Forecasting accuracy increased significantly in patients post-treatment (pre M = 0.71, SD = 0.12; post M = 0.76, SD = 0.07; t(14) = 1.857, p = 0.042, g = 0.6), whereas controls showed no significant change in accuracy between baseline and repeat assessment. There were no significant between-group differences in accuracy at baseline or follow-up. Bias scores confirmed a baseline pessimism in patients (mean bias M = -0.10, SD = 0.12), significantly below zero [t(14) = -3.260, p = 0.006, g = 1.1]. Following treatment the mean bias shifted to near zero (M = 0.02, SD = 0.08), a significant reduction in pessimism [t(14) = -2.714, p = 0.017, g = 0.7], and no longer differed from zero. Controls showed negligible bias at both time points. Correlational analyses indicated that greater baseline pessimism was associated with more severe depressive symptoms among patients (Spearman rs = -0.55, p = 0.017). Change scores showed that reductions in BDI were significantly related to reductions in pessimism for patients (r = -0.57, p = 0.014).

Serafini and colleagues interpret their findings as evidence that the combination of psilocybin and psychological support alleviated an unjustified pessimism bias in patients with TRD, and that this cognitive change occurred in parallel with marked reductions in depressive symptoms. The investigators note that patients were excessively pessimistic at baseline relative to their actual life circumstances, and that after treatment they made more accurate forecasts that resembled the behaviour of healthy controls. The authors situate these results within prior psychedelic research showing rapid antidepressant effects and enduring increases in optimism and wellbeing following single doses of classic psychedelics. Potential neurobiological mechanisms discussed include engagement of the 5-HT2A receptor, temporary disruption and subsequent reintegration of the default mode network, and a proposed "reset" or plasticity-enhancing state that may permit revision of entrenched cognitive biases in an enriched therapeutic context. The discussion also acknowledges that optimism and pessimism can each have adaptive functions, and argues for a balanced, reality-focused perspective as optimal for mental health. Key strengths highlighted by the study team are the use of an objective, behavioural measure (the POFLE) rather than only subjective self-report, and the focus on a cognitive facet of the depressive triad that can be prospectively validated against actual events. At the same time, the authors emphasise several limitations: the open-label design, small sample size, and the fact that control subjects did not undergo the same treatment procedures (so non-pharmacological components of the intervention, such as psychological support, cannot be ruled out). They also acknowledge the cross-sectional nature of the follow-up interval reported here, the inability to fully exclude natural symptom decline, and the unresolved question of causality—whether cognitive-bias change is a direct effect of psilocybin or an epiphenomenon of mood improvement. Finally, the investigators propose that a common mediating factor may underlie both symptom change and bias change and recommend further controlled, prospective studies—ideally incorporating neuroimaging paradigms that probe prospection—to clarify mechanisms, specificity and durability of the effects reported here.

The study concludes that psilocybin administered with psychological support was associated with alleviation of pessimism bias and large reductions in depressive symptoms in a small sample of TRD patients, as measured behaviourally by the POFLE task. The authors state that this is the first behavioural demonstration of psychedelic-induced change in a cognitive bias integral to depression and call for further controlled research to establish causality, reliability, specificity and the durability of these findings.