Video was back-projected onto a screen fastened to the front of the head coil, made visible to the subject by a mirror. The experiment comprised 10 trials, each of which consisted of a 60s task period flanked by two 30s fixation periods. During the fixation periods the display was dark except for a fixation cross which remained present throughout each trial. During the task periods, coloured squares subtending 1.8° were displayed against a dark background in the left and right upper quadrants, centred 5° lateral and 3° superior to the fixation cross. In each of these two locations, stimuli appeared for 56ms and were separated from each other by a further 56ms, giving a presentation rate of 9 stimuli per second. The offset of a stimulus on the left coincided with the onset of a stimulus on the right, and vice versa. Target squares were red; non-target squares were green. In each location, targets occurred with probability 0.0006 for each 112ms period during the first 18s of each trial (1 target over the 10 trials) and with probability 0.007 during the remaining 42s of each trial (27 targets over the 10 trials). The average interval between targets was 16s.
Subjects were instructed to maintain fixation at the centre of the display throughout each trial, and to begin each trial by attending covertly to one of the two stimulus locations and ignoring the other. This starting location was counter-balanced across trials. On detecting a target stimulus in the attended location, subjects had to respond by shifting attention covertly to the opposite location, and by moving the index finger of the dominant hand in the direction of this shift. The rest of the trial then proceeded in the same manner, attending at the new (previously ignored) location and ignoring the old (previously attended). Attention thus alternated between the two spatially separated streams of rapid serial visual stimuli, shifting from one stream to the other on detection of each attended target. Essentially, this task consisted of two oddball paradigms, presented side by side and alternately attended.
As our interest lay not in psychophysical measures of ability to direct attention but in the neurophysiological effects associated with successful direction of attention, the task was designed so that most subjects would perform at or near ceiling. Before entering the scanner, subjects practised the task till they felt comfortable with it. Once in the scanner, subjects were permitted to rest for up to three minutes between trials in order to avoid visual fatigue and to relieve anxiety.