Pupil responses to task requirement in goal-directed movements

Objectively measuring the operators’ task workload in goal-directed motor tasks such as surgical operations, is important for performance and safety. This thesis presents an approach for objectively measuring task workload in goal-directed movements using an important eye response: the pupil diameter. We demonstrate how to capture movement-related pupil size changes during motor tasks, investigate how the pupil responds to task requirement, and show that the pupil diameter can be employed as an objective and quantitative indicator of task workload in motor tasks. In particular, we studied tasks where a tool is used to manipulate a target. The challenges include how to quantitatively define the task requirement, accurately detect the movement-related pupil responses from other event-evoked pupil changes, and separate the pupil responses from consecutive movements. We quantify the task requirements using Fitts’ index of difficulty. The movement-related pupil events are captured from a time window aligned at a specific moment of tool movement which is automatically detected from recorded task videos. To separate overlapping movement responses, each tool movement is divided into two phases: the Transport and Landing phases. Three experiments were conducted to verify correlations between pupil responses and task difficulty of goal-directed movements, including two discrete target-pointing tasks and a continuous target-pointing task. We also investigated the pupil responses to task difficulty in a realistic situation: that is, during performance of a simulated surgical task, where we found that the pupil responses to the subtasks were related to the task difficulty. Overall, we found the pupil diameter can be employed to objectively measure task workload in goal-directed movements, by conducting three experiments of discrete and continuous target-point tasks and a study of real-life motor task. The findings constitute the foundation for developing methods to objectively and quantitatively evaluate task workload of motor tasks using pupil diameter, and have a variety of implications in enhancing psychophysiological interactions in human-centered HCI and evaluating mental workload in high skill-demanding domains such as driving, aviation, and surgery.