Improving the continuous tracking paradigm to investigate implicit motor learning

Abstract

In three experiments we investigated factors that undermine conclusions about implicit motor learning in the continuous tracking paradigm. In Experiment 1, we constructed a practice phase in which all three segments of the waveform pattern were random, in order to examine whether tracking accuracy decreased as a consequence of time spent on task. Tracking error was lower in the first segment than in the middle segment and lower in the middle segment than in the final segment, indicating that tracking accuracy decreased as a function of increasing time-on-task. In Experiment 2, the waveform pattern was controlled so that the middle segment, which was repeated in each practice trial, was as difficult to track as random segments, which were not repeated. Substitution of the repeated pattern with a random pattern (in a transfer test) resulted in significantly decreased tracking performance, suggesting that characteristics of the repeated pattern had been learnt. However, the time-on-task effect between the first segment (random) and the second segment (repeated, but effectively random due to so few repetitions) was absent in the first practice block, raising the possibility that despite our attempts to match segment complexity, Segment 2 was easier to perform. In Experiment 3, the time-on-task effect between the first segment (random) and the second segment (repeated) in the first practice block returned when participants first completed a 1 block of warm-up trials in which all three segments were random, suggesting that complexity of the repeated waveform pattern we used was matched with random ones. These findings confirm the importance of controlling for waveform complexity when employing the continuous tracking task paradigm and show that time spent tracking can influence performance. Several suggestions for refining the continuous tracking paradigm for implicit learning research are outlined.