Each subject was asked to complete the experimental procedure once on each of five occasions, where the occasions were separated by as little as 24 hours and by as much as three weeks. Each session required approximately 45 minutes to complete.
In each trial, subjects were presented with one repeating rhythmic auditory pattern and were asked to respond by striking a key on a synthesizer keyboard synchronous with the perceived starting point of the pattern. Subjects were asked to continue to strike the key at the beginning of each repetition until confident that they had perceived the starting point. Once subjects were confident of their response, they were asked to press a mouse button ending the trial.
The PsyLog software  running on a NeXT workstation was modified to present the stimuli and gather the responses. Subjects were presented with a short set of practice trials and then a set of 115 experimental trials. Subjects were informed that they could rest at any time that they became fatigued, and that the software would wait for them.
A single rhythmic stimulus was composed of a fixed number of beats: equal intervals of time which could either be empty or be filled with a percussive sound at the beginning of the interval. Thus a measure could be represented by a binary number where each binary digit represents a beat: a zero representing an empty interval and a one representing an interval with a percussive sound at its beginning. The set of stimuli for Experiment 1 consisted of all of the unique rhythmic patterns of length 8 or less, 115 patterns in all.
In each trial, the stimulus initially began with a short beat length of 10 ms which quickly slowed to a steady beat length of 250 ms. The beat on which the stimulus was initiated was chosen at random for each presentation of the stimulus. The combination of these two methods minimized the subjects' ability to associate the beginning of the presentation of the first beat of the stimulus with the perceived beginning of a measure within the repeating pattern.
The percussive sound used in this experiment was a synthesized musical cowbell produced by a Roland MT-32 MIDI wavetable synthesis module and delivered to the subjects binaurally via Sennheiser HD--414--SL headphones. The user responded by striking a key on a Kawai K5 Digital Synthesizer keyboard. Several variables were measured for each keypress: the time of response in milliseconds relative to the beginning of the presentation of the stimulus; the time of response in milliseconds relative to the beginning of the pattern as represented internally by the computer software; and the velocity of the response as an integer between 1 and 127.