Sensitivity to time change: The role of context and skill

Three experiments used a 2-alternative, forced-choice procedure to examine the performance of skilled (musician) and unskilled (nonmusician) participants in detecting small time changes embedded within the test cycles of rhythmical patterns. Relational versus statistical properties of rhythms as well as rate and magnitude of to-be-detected time changes were varied in Experiments 1 and 2. Experiment 3 continued to vary structural properties of sequences (relational vs. statistical) and of experimental sessions. In general, listeners were better with relationally simple than with complex patterns in spite of the fact that the former were, on average, more variable; in addition, musicians were superior to nonmusicians only with simple sequences. In Experiment 3, more marked effects of relational structure within both patterns and sessions were evident. Results were interpreted in terms of an oscillator model of dynamic attending.