J. Scheirer, Raul Fernandez, J. Klein, Rosalind W. Picard

Abstract

Using social science methods to induce a state of frustration in users
we collected physiologi cal
video and behavioral data
and developed a strategy for coupling these data with realworld events The eectiveness of the proposed strat egy was tested in a study with thirtysix sub jects
where the system was shown to reliably synchronize and gather data for aect analysis Hidden Markov Models were applied to each sub jects physiological signals of skin conductivity and blood volume pressure in an eort to see if regimes of likely frustration could be automat ically discriminated from regimes when all was proceeding smoothly This pattern recognition approach correctly classied these two regimes  of the time Mouseclicking behavior was also synchronized to frustrationeliciting events
and analyzed
revealing NN distinct patterns of clicking responses Keywords A