Decreased mental time travel to the past correlates with default-mode network disintegration under lysergic acid diethylamide

Lysergic acid diethylamide (LSD) is a classic psychedelic whose psychological effects are largely attributed to serotonin 2A receptor agonism, but the links between its neurobiology and subjective phenomenology remain incompletely understood. Previous work has associated spontaneous mental time travel—the tendency during mind-wandering to recollect past autobiographical episodes or imagine future experiences—with the brain's default-mode network (DMN), and has shown that DMN resting-state functional connectivity (RSFC) correlates with frequency of such self-referential cognition and is elevated in depression. Conversely, acute psychedelic states have been reported to reduce DMN RSFC and to occasion experiences described as ego-dissolution, suggesting a possible mechanistic connection between DMN disruption and altered temporal focus of spontaneous thought. Speth and colleagues therefore set out to combine linguistic analysis of spontaneous mentation reports with resting-state fMRI to test whether LSD reduces linguistic markers of mental time travel to the past, and whether any such change relates to LSD-induced decreases in intra-DMN connectivity. The study used a placebo-controlled, within-subject crossover design in healthy volunteers, analysing mentation reports obtained about 2.5 hours after intravenous administration of 75 µg LSD or saline placebo and relating these behavioural measures to contemporaneous DMN RSFC derived from task-free fMRI scans.

The study used a placebo-controlled, within-subjects crossover design. Twenty healthy volunteers were recruited and completed screening; 19 contributed mentation reports used in the main linguistic analyses (one male was excluded due to a missing LSD-condition report). Participants had prior experience with classic psychedelics and met clinical exclusion criteria (e.g. no personal history of diagnosed psychiatric illness). Dosing sessions were separated by at least two weeks and order of LSD versus placebo was balanced across participants; volunteers were blind to order, but the research team were not. Each participant received 75 µg LSD intravenously (in 10 mL saline infused over two minutes) or 10 mL saline placebo. After infusion, participants underwent a habituation period and then two eyes-closed, task-free BOLD fMRI scans (each ~7 minutes 20 seconds) on a 3T system approximately one hour post-dosing; the first structured interview began soon after scanning, roughly 2.5 hours post-dosing, and mentation reports used for analysis were taken at the start of these interviews. Subjective effect intensity during scanning was measured using a visual analogue scale. Five participants’ fMRI data were excluded (one did not complete scanning, four due to motion), leaving n=15 for the imaging analyses. Quantitative linguistic analysis was applied to verbatim transcriptions of the mentation reports. Six independent, blinded raters identified instances of cognitive agency and ensuing "mental spaces"—linguistic constructs reflecting past, present or future scenarios—drawing on theta-role theory and the cognitive-semantic mental spaces framework. Raters classified agency perspective (first/second/third person, singular/plural) and whether the resulting mental space pertained to past, present or future. While the Methods section at one point refers to 77 reports to be analysed, the extracted text shows that 38 reports were actually collected (19 per condition); this discrepancy is noted in the extracted material but not resolved. fMRI preprocessing used band-pass filtering (0.01–0.08 Hz) and multivariate ICA to extract components; a canonical DMN template derived from Human Connectome Project data was used to calculate intra-DMN RSFC for each condition. Paired t-tests compared within-subject intra-DMN connectivity between LSD and placebo and reductions in DMN RSFC (expressed as z-score changes) were treated as an index of DMN disintegration for correlation with linguistic measures. Statistical tests on behavioural data included repeated-measures MANOVA to test omnibus effects across past/present/future references, post-hoc paired t-tests and planned comparisons focused on past-referenced mental spaces, and Pearson correlations for association analyses. Inter-rater reliability was assessed (Cronbach's alpha reported).

Behavioural sample size for mentation reports was 19 participants per condition, yielding 38 reports in total. Raters identified on average 2.03 cognitive agencies per report (SD=1.67) and inter-rater agreement was very high (Cronbach's alpha=0.92). Participants' verbal reports were lengthier after LSD (mean words per report 252.11, SD=146.88) than after placebo (mean 155.58, SD=114.77); the extracted text states this difference as statistically significant but does not clearly report the complete t-statistic and p-value for that comparison. An omnibus repeated-measures MANOVA testing the effect of LSD on counts of mental spaces for past, present and future did not reach significance (F(3,16)=1.91, p=0.168). A planned comparison, however, indicated a selective reduction in references to the past under LSD: mean past-references after LSD were 0.079 (SD=0.2) versus 0.71 (SD=1.04) after placebo. The extracted t-statistic for this planned comparison is given as t(18)=2.5 but the p-value is incomplete in the extraction; the authors report this as a significant effect. There were no significant differences between conditions for references to the present (placebo mean=1.00, SD=0.90; LSD mean=1.04, SD=0.79; t(18)≈-0.114, p=0.911) or the future (placebo mean=0.64, SD=0.63; LSD mean=0.51, SD=1.08; t(18)=0.5, p=0.624). Correlation analyses examined relationships between subjective intensity of LSD effects during scanning, DMN disintegration (between-condition differences in intra-DMN RSFC), and reductions in past-focused mental spaces. The association between subjective intensity and reduced past-references was in the predicted direction (r=0.34, r²=0.17) and approached but did not reach conventional significance (p=0.076). The relationship between greater DMN disintegration and fewer past-references was also in the predicted direction (r=0.51, r²=0.26) and again narrowly missed conventional significance in the extracted text (p=0.054). Lifelong use measures for LSD, psilocybin and cannabis, time since last use, weekly alcohol use and similar covariates did not correlate with the main outcomes (all p>0.05). For imaging, five participants were excluded from fMRI analyses (leaving n=15), and the authors note that this sample size may be small for correlational neuroimaging analyses.

Speth and colleagues interpret their findings as showing that acute LSD selectively reduces spontaneous mental time travel to the past while leaving present- and future-focused spontaneous mentation relatively unaffected, and that this behavioural effect aligns with LSD-induced reductions in DMN integrity. The selective nature of the effect is highlighted: despite longer and more elaborate verbal reports under LSD, participants produced fewer linguistic instances of past-focused mental spaces. The near-significant correlations—stronger subjective intensity and greater DMN disintegration relating to fewer past-references—are presented as supporting evidence that the effect is drug-related and mechanistically linked to DMN changes. The authors situate these results within existing literature linking the DMN, hippocampal-parahippocampal circuitry and autobiographical memory; they propose that disruption of hippocampal–DMN connectivity may reduce the input of autobiographical content that sustains a narrative sense of self, a process related to the phenomenology termed ego-dissolution. The paper contrasts spontaneous, often mundane past-reflections captured in the present linguistic coding with the clinically observed phenomenon of vivid reliving of salient autobiographical memories during psychedelic-assisted psychotherapy, noting that cued recollection in therapeutic settings is a different task and may not be directly comparable to spontaneous unprompted reports. The authors also raise the possibility that enhanced suggestibility under psychedelics could bias cued-memory findings, arguing that measures of spontaneous cognition, as used here, are relatively insulated from such expectancy effects. Key limitations acknowledged include that participants were still acutely intoxicated when providing retrospective reports, so it is unclear whether reduced past-references reflect altered in-the-moment mentation or altered retrospective description; small fMRI sample size for correlational analyses (n=15) which may limit reliability; and the alternative interpretation that the finding could reflect a non-specific cognitive perturbation rather than a psychedelic-specific effect. The authors propose future tests to address these issues, such as comparing other intoxicants or exhilarating experiences, studying patient populations in therapeutic contexts, probing post-acute effects, and further examining hippocampal–DMN circuitry. They also suggest that the observed reduction in past-focused thinking might mechanistically align with therapeutic approaches that reduce ruminative, past-oriented thinking in depression (for example mindfulness), and advocate further research to test whether psychedelics produce enduring shifts in temporal focus that could be therapeutically relevant.

The study concludes that acute intravenous LSD at 75 µg selectively decreased spontaneous linguistic markers of mental time travel to the past, and that this effect correlated with the reported intensity of subjective effects and with decreased intra-DMN connectivity. The authors link these findings to the construct of ego-dissolution and to the notion of a decomposed narrative self, and they suggest potential relevance for psychedelic-assisted approaches to conditions characterised by excessive past-focused rumination, such as depression. The paper emphasises the need for further research to establish the robustness, specificity and clinical implications of these findings.