Neural Indicators of the Depth of Cognitive Processing for User-Adaptive Neurotechnological Applications
The ability to infer implicit user variables in real-time and an in unobtrusive way would open a broad variety of applications such as adapting the user interface in human-computer interaction or safety relevant assistance systems in industrial workplaces. Such information may be extracted from behavior, peripheral physiology and brain activity. Each of these sensors has its advantages and disadvantages suggesting that finally all available features should be fused. While in Brain-Computer Interface (BCI) research powerful methods for the real-time extraction of information from brain signals have been developed, comparatively little effort was spent on the extraction of hidden user states. In this work, we propose a novel experimental paradigm to study the feasibility of quantifying how deeply presented information is processed in the brain. An investigation of event-related potentials (ERPs) demonstrates the effectiveness of our task in inducing different levels of cognitive processing and shows which features of brain activity provide discriminative information.