Abstract
We have been working on the development of the method for quantitative assessment of mental workload (MWL). The quantitative assessment of MWL is helpful for improvement of the usability of various computer systems, the working environment, and the schedule management. Methodological requirements include (1) being able to compare MWL under various working conditions, (2) caputuring the time course of MWL, (3) assessing it in relation to individual differences, and (4) executing assessment without giving subjects extra load. Some autonomic responses meet the first two requirements. They may be the promising measures for MWL if a proper methodology is applied in order to satisfy the 3rd and 4th requirements. In the present paper, we show the results of our basic experiments. The first one concerns the differences in autonomic response patterns among subjects caused by a mental arithmetic, and the second concerns the similarities and differences in them as a function of MWL caused by various tasks. The findings of two experiments suggest the importance of the multidimensional use of the indices which are differently influenced by the two autonomic nervous systems, the exsistence of individual response specificity under metal tasks, the difficulty in prediction for individual response patterns to MWL using those caused by physial tasks. The necessity of careful consideration for respiration and posture at the interpretation of autonomic indices was reaffirmed.
