Fitts' law when errors are not allowed: Quantification of reciprocating trajectories and estimating information processing

There are circumstances in which humans are required to move in environments wherein accuracy should be maximized, such as golf putting or surgeons' hand movements during surgical procedures. Fitts' law expresses human movement by movement time and task difficulty, which is determined by one's distance from the target and the targets' size. Additionally, this law is considered as the most universal expression of human movement. However, it is calculated based on the possibility of failure, mainly because a tap error ratio of ~4% occurred in the continuous tap task used by Fitts. Thus, this study aimed to examine how movement changes can occur with an error rate of 0%, and whether Fitts' law can be applied to human movement in such a setting. The continuous tap task was performed under two conditions: a conventional one, and a new condition where the devised error rate becomes 0%. To measure movement change during tapping, the variation of the entire trajectory was quantified by principal component analysis, and changes in the quantified values of the trajectory toward the target were examined. For the new condition, no tapping errors occurred; the quantified value of the trajectory toward the target decreased compared with the conventional condition. It was considered that this related to the process of feedback control, and that this process related to an increase in movement time per tap. We suggest that increased information processing may account for these changes. Furthermore, the Fitts' law model, shown from the regression lines for movement time and difficulty, displayed a high fit for both conditions. However, it was difficult to evaluate movement time with the highest index of difficulty for the new condition using the model for the conventional condition. Therefore, we conclude that the conventional model may need to be modified for conditions where the error rate is 0%.