New Approaches for Detecting Workload and Stress

Markus Guhe
CogWorks Laboratories
Rensselaer Polytechnic Institute

Workload is a difficult concept for which no commonly accepted
definition exists. Researchers use the term very differently, 
which reflects that most current measures are problematic. I 
report on two projects in which we explore how well new, non-
intrusive technologies are suited to detect different levels 
of workload and the correlated stress. In both projects we 
develop models in the cognitive architecture ACT-R, which makes 
it possible to determine the cognitive processes a user is 
using when performing a given task. In one of the projects we 
use cameras to monitor facial features and head position of a 
user as well as an "emotional mouse", which measures
physiological signals like Galvanic Skin Response. In the 
second project we use functional Near Infra Red imaging (fNIR) 
to determine the oxygenation level of the frontal brain 
regions. Combining these approaches we get a thorough 
understanding of the underlying cognitive mechanisms and can 
identify reasons why the participants of our experiments are 
likely to experience stress in certain experimental conditions.