REM (rapid eye movement) sleep has been theorised as a particularly favourable state for psi reception. During REM, external sensory input is blocked by active thalamic gating, internal narrative production is high, and the critical, evaluative processing of waking life is suspended. Researchers proposed that if telepathic or clairvoyant information can be received by the mind, dreams might be the medium through which it most readily surfaces into awareness — much as unconscious material is considered more accessible during dreaming than waking.
The most rigorous experimental investigation of dream telepathy was conducted at the Maimonides Medical Center Dream Laboratory in Brooklyn, New York, by psychiatrist Montague Ullman, psychologist Stanley Krippner, and later Charles Honorton. Over approximately 14 years and 450+ formal sessions, the laboratory produced the primary empirical dataset for dream-state psi research.
The standard protocol involved a sleeping subject monitored by EEG for REM onset. A sender, isolated in a separate room, was given a randomly selected art print (the "target") and instructed to concentrate on it mentally during the subject's REM period. At the end of each REM period, the subject was awakened and asked to describe dream content. In the morning, independent judges received the dream transcripts and a pool of art prints (including the target) and ranked them for correspondence — without knowing which was the actual target. Statistical analysis compared achieved rankings against chance expectation.
Results across the formal database showed a 63% hit rate against a 50% chance expectation, with overall statistical significance below p = 0.001. Effect sizes were modest but consistent across experimenters and subjects.
Malcolm Bessent participated in precognitive dream experiments at Maimonides — a design in which the target image was selected after the subject's dream was recorded. Bessent's pre-cognitive dream transcripts were judged by independent evaluators to correspond significantly above chance to images selected the following day. These trials, among the strongest in the Maimonides database, were published in the Journal of the American Society for Psychical Research.
The full programme was summarised by Ullman and Krippner in their book Dream Telepathy (1973), which remains the standard reference for the Maimonides research.
Subsequent replication attempts have been mixed. A 1999 meta-analysis by Milton and Wiseman examined 30 studies conducted outside Maimonides and found no significant overall effect, concluding that the Maimonides findings had not been independently replicated. Proponents noted that most replication attempts used reduced protocols without the full sleep-lab monitoring apparatus, potentially eliminating the experimental conditions that made the original work effective. The debate remains unresolved.
Beyond the laboratory, spontaneous pre-cognitive dream reports constitute a large portion of the SPR's historical archive of claimed psi. Researcher W.E. Cox compiled statistics on railway accident occupancy rates, finding suggestive evidence that trains involved in accidents carried fewer passengers than comparable trains on non-accident days — a finding attributed by some to pre-cognitive avoidance, though methodological objections have been raised. Anecdotally documented cases include dreams reported by associates of President Lincoln in the days before his assassination, and the collection of Titanic premonition dreams assembled by journalist and researcher J.C. Barker in 1966.
Shared dreaming refers to reports by two or more individuals of dreams that share unusual specific content on the same night without prior communication. Anecdotal cases are common in close relationships, particularly couples, twins, and parent–child dyads. Experimental attempts to induce shared dreaming have been conducted by multiple researchers but have not produced replicable controlled results.
Cross-correspondence in dreams — analogous to the SPR's mediumship cross-correspondences — has been reported in case collections but not investigated under controlled conditions. The theoretical interest lies in whether shared dream content between geographically separated individuals could represent a form of non-local information exchange.
The laboratory verification of lucid dreaming (Hearne 1975; LaBerge 1980) opened new experimental avenues for dream-state psi research. Subjects able to signal lucidity via pre-agreed eye movements could in principle perform intentional psi tasks during verified REM sleep, allowing a cleaner separation between the dreaming state and post-awakening narration. Some pilot studies have explored psi performance in lucid-dream conditions, with preliminary positive results, but the research programme remains nascent.
This page presents documented research and reported phenomena. No claims about validity or mechanism are made by this archive.