Temporal context influences time perception

Seven experiments demonstrate that rhythmic context effects on time perception are attributable to broadly-tuned filters sensitive to absolute durations. All experiments use the same time perception task; on each trial, participants compare two target durations ranging from 420 to 580 milliseconds ( ms), showing response biases when certain types of irrelevant distracting durations are between the target durations. Distractors consisted of a series of short tones; the inter-onset intervals of these tones were the independant variables of all the experiments. Unlike other reported cases of context effects on time perception, the distractors occur outside (rather than during) the target durations. These biases varied systemmatically as a function of the absolute duration of the distractors---not of any higher-order rhythmic organization. Biases were largest when distractors were 200 ms shorter or longer than the target durations, falling off gradually as the distractors became more similar to the targets. For both short (420--580ms) and long (920--1080 ms) targets, biases disappeared when distractors were 500ms shorter or longer than the targets. These data support an interval timer model of timing over an oscillator model since absolute duration rather than rhythmic organization or rhythmic relation to the targets creates this effect. The absolute duration of the distractors rather than rhythmic organization generated the effect, suggesting that the timer encodes time absolutely, like an interval timer. The fact that distractors differing greatly from the targets did not induce biases suggests that there is no global rate parameter, as suggested by most oscillator models. The author presents a type of interval timer model, dubbed the Spreading Interval Activation Model (SIAM) of a central timer to account for this effect and proposes three future experiments to test this model against oscillator models.