Insomniacs commonly complain that they are unable to get to sleep at night due to unwanted thoughts, worries and concerns. The present study investigated whether brief training in identifying and elaborating an interesting and engaging imagery task for use during the pre-sleep period can reduce unwanted pre-sleep cognitive activity and sleep onset latency. Forty one people with insomnia were given one of three instructional sets to follow on the experimental night; instructions to distract using imagery, general instructions to distract, or no instructions. Based on previous findings reported by Salkovskis & Campbell (1994) 'Behaviour Research and Therapy 32 (1994) 1' and ironic control theory (Wegner, 1994) 'Psychological Review 101 (1994) 34', it was predicted that (1) "imagery distraction" would be associated with shorter sleep onset latency and less frequent and distressing pre-sleep cognitive activity compared to the "no instruction" group and that (2) "general distraction" would be associated with longer sleep onset latency and more frequent and distressing pre-sleep cognitive activity compared to the "no instruction" group. Support was found for the first but not the second prediction. The success of the "imagery distraction" task is attributed to it occupying sufficient "cognitive space" to keep the individual from re-engaging with thoughts, worries, and concerns during the pre-sleep period. In addition, "imagery distraction" involved a very specific alternative cognitive task hence the operating process was given a feature positive search, conditions where mental control is likely to be achieved.