Leveraging Changes in Mental Workload during Task Execution to Mitigate Effects of Interruption

Task interruption often has a significant negative impact on a user’s productivity and affective state. Cognitive theorists have argued that interrupting the execution of primary tasks at moments of lower mental workload would mitigate effects of interruption, yet knowing just where these moments occur remains elusive. In this article, we present empirical results from three user experiments aimed at further explicating the relationship among mental workload, task execution, and effects of interruption. Principal results include (i) a user’s mental workload exhibits momentary decreases at subtask boundaries during task execution; (ii) workload decreases more at boundaries higher in a task model than at boundaries lower in a model; (iii) the decrease in workload is not uniform within the same level of a model; and (iv) interrupting task execution at subtask boundaries with lower workload meaningfully mitigates effects of interruption. These results demonstrate that mental workload is an effective predictor of how opportune different moments in a task are for interruption and contribute further answers to where moments of lower workload exist within the structure of a task. Practical implications of these results are discussed for computational systems that reason about when to interrupt users engaged in tasks.