Using a Sleep Laboratory for a Question It Was Not Built For

The Brooklyn work took advantage of an instrument that had only recently become available. Electroencephalography could identify rapid eye movement sleep as it happened, which meant a subject could be woken during a dream rather than asked about one at breakfast. The protocol used that directly. A subject slept wired to the monitor in one room. A sender in another, awake, concentrated for the night on a single art print chosen at random from a sealed pool, usually a reproduction with strong and unusual imagery. Whenever the monitor showed the subject entering dreaming sleep, an experimenter woke them over an intercom and recorded a description of the dream.

In the morning the subject added associations. The transcripts and the pool of candidate prints then went to judges who had not been present, who ranked how well each transcript matched each print. Above-chance ranking of the correct print was the reported result across several series, and Montague Ullman, Stanley Krippner and Charles Honorton presented it to a general readership in 1973. The sender stayed awake through the night with the print and a set of related objects, on the theory that sustained engagement with the image mattered more than passive viewing. Whether that made any difference was never tested, since no version of the protocol omitted it.