Doktorsvörn við tækni- og verkfræðideild- Eydís Huld Magnúsdóttir

Cognitive workload classification with psychophysiological signals for monitoring in safety critical situations

  • 18.1.2019, 13:00 - 15:00

Eydís Huld Magnúsdóttir will defend here doctoral thesis "Cognitive workload classification with psychophysiological signals for monitoring in safety critical situations" on the 18th of January at 13:00 in room V102.

 Doctoral committee:

- Main supervisor:  Dr. Jón Guðnason, Associate Professor at the School of Science and Engineering, Reykjavik University

- Co-supervisor:     Dr. Kamilla Rún Jóhannsdóttir, Associate Professor at the School of Buisness, Reykjavik University

- Co-supervisor:     Dr. Arnab Majumdar, Reader in Transport Risk Management at Imperial College


Dr. Paco Saez, Reader in Air Traffic Management at Cranfield University


Monitoring cognitive workload has the potential to improve both performance and fidelity of individuals facing safety-critical situations in their working environments. Psychophysiological signals, in particular from speech and the cardiovascular system, are an opportune choice for monitoring individuals providing minimum intrusion and disruption. For reasons perhaps mostly rooted in individual differences, current methods are limited in that moving beyond binary classification has proved challenging. The aim of the present research was to investigate the potential of speech- and cardiovascular signals both separately and in conjunction, for cognitive workload detection, using short-term variability for screen and heart rate based classification schemes. The aim was further to explore alternative methods in order to take into consideration individual differences in the cardiovascular signals that describe the extent of the reactions of the whole cardiovascular system to cognitive workload. For this purpose, new method where a single distance measure describing the hemodynamic reactions of individuals during tasks compared to their own baseline is introduced. A total of 100 university students participated in a study providing extensive information on reactions to cognitive workload that can be used for individual comparisons.

The results showed that trinary classification is well achievable with the methods introduced and that the two signals do compliment each other in the cognitive workload classification task. The proposed distance measure showed obvious reactions to cognitive workload during tasks and that these reactions are highly various between individuals.