Investigation of unconscious precognition in the visual attention system

Abstract

Precognition can be defined as an anomalous correlation between current cognitive activity and a future event. Using behavioural and physiological measures, a number of previous studies have reported evidence for unconscious precognition during a variety of task conditions. The current thesis presents five experiments that were designed to test for unconscious precognition in the visual attention system while participants were engaged in a short term visual memory task. Each trial consisted of a study and test phase. In the study phase, participants were required to memorise an array of four stimuli while their eye movements were recorded. After a brief retention interval, a probe stimulus was presented for a yes/no recognition test. Two conditions were employed and were randomly determined. In the old condition, the probe was a stimulus viewed during study, termed the target. In the new condition, the probe was a novel stimulus. Experiments tested for the presence of precognition by examining whether there was a difference in the degree to which visual attention was allocated to items during the study phase of old and new trials. Two further studies were also carried out involving simulations that aimed to establish the extent to which a previously described artefact, termed the expectation bias, may impact on the results. Experiment 1 suggested that participants spent more time attending to target stimuli in old compared to new trials, a result that appeared to provide evidence for precognition. However, the data was considered unreliable due to inadequate randomisation. An exact replication of Experiment 1 was carried out in Experiment 2 with adequate randomisation, but failed to find evidence for precognition. Experiment 3A was a further attempt to replicate the preliminary results of Experiment 1 using more extensive randomisation procedures while Experiment 3B explored the potential role of the probe stimulus in generating a precognitive effect. However, no support for the precognitive hypothesis was found in either experiment. A fully balanced design was employed in Experiment 4 in order to control for potential confounds such as position and saliency effects. The results supported the precognitive hypothesis and suggested that less attention was allocated to targets in the old condition. An exploratory analysis also examined the relationship between several standardised stimulus variables and the apparent precognitive effect observed in Experiment 4. The results revealed a suggestive relationship between the size of the effect and item ratings of familiarity and visual complexity. Simulations of an expectation bias in Experiments 5A and 5B together with post-hoc examination of the data from the current series of experiments suggest that this artefact is not a plausible explanation for the observed effects. The thesis ends with a discussion of several methodological issues that may impact on both the interpretation of positive results and the conclusions that may be reached from this body of data as a whole. Finally, suggestions for further work are made.