In this paper, we showed that fine-grained linguistic complexity has measurable effects on cognitive load. We designed the tasks in our experiment to require continuous attention. The language task clearly affected performance on the primary steering task: we saw the effect of the secondary task in all of our measures. Furthermore, we found effects of linguistic ambiguity and complexity in our measures of cognitive load: during the disambiguating region, we observed significantly higher steering deviation, which indicated that people are allocating more mental resources to the linguistic task, hence impeding steering performance. We also found evidence for a measurable effect of linguistic complexity in our pupillometric measure ICA: the ICA was significantly higher during the disambiguating region and the following second for the ORC condition compared to SRC. This experiment has provided early support for both the ConTRe steering task and the ICA as useful measures for the assessment of language-induced cognitive load. We have shown several aspects of the suitability of the ConTRe steering task for measuring the fine-grained effects of cognitive load in a smooth, continuous driving environment. A key characteristic of the task is that is uses a continuous measure instead of turns or abrupt stimuli, making it suitable for use with tasks that require careful synchronization, such as measuring the workload that accrues to particular grammatical turns. It extends the assortment of solutions available for measuring driver distraction in simulator environments and was created to compensate certain drawbacks of other driving tasks. In summary, we successfully demonstrated that a more sensitive task is able to reveal more subtle effects on driving performance. A more fine-grained evaluation of driving performance enhances the possibilities to investigate cognitive workload and the effect of secondary tasks on cognitive workload.
Measuring linguistically-induced cognitive load during driving using the ConTRe task
2013
8 Seiten, Bilder, Tabellen, 24 Quellen
Conference paper
English