A Dynamic Model of Behavior in a Discrete Open-Loop Self-Paced Motor Skill

Some results of a study of the sequential response accuracy exhibited by individual subjects engaged in a particular skilled positioning task requiring discrete open-loop self-paced behavior are presented. The task resembled a discrete self-paced tracking task. A mathematical deterministic discrete model capable of predicting the sign and magnitude of the subject´s sequential response error was formulated. This model, based solely on the similarity of previous sequential responses to the required response, contained three coefficients which were fitted to each subject´s data by linear regression techniques. There was often a fairly high degree of correlation between a subject´s actual and model predicted error. Zero mean error and the presence or absence of the range effect which are observable gross phenomena are direct consequences of the behavior described by the model. Because of the possibility of statistical artifacts, the model was also checked against computerized pseudosubjects whose deterministic and random behavioral components were programmed. The paper is presented as an effort to further basic knowledge of discrete open-loop self-paced behavior and of fine-grain analysis and modeling of the accuracy of skilled sequential responses.