Automated stress detection using keystroke and linguistic features: An exploratory study

  • Authors:
  • Lisa M. Vizer;Lina Zhou;Andrew Sears

  • Affiliations:
  • Department of Information Systems, UMBC, Baltimore, MD 21250, USA;Department of Information Systems, UMBC, Baltimore, MD 21250, USA;Department of Information Systems, UMBC, Baltimore, MD 21250, USA

  • Venue:
  • International Journal of Human-Computer Studies
  • Year:
  • 2009

Quantified Score

Hi-index 0.00

Visualization

Abstract

Monitoring of cognitive and physical function is central to the care of people with or at risk for various health conditions, but existing solutions rely on intrusive methods that are inadequate for continuous tracking. Less intrusive techniques that facilitate more accurate and frequent monitoring of the status of cognitive or physical function become increasingly desirable as the population ages and lifespan increases. Since the number of seniors using computers continues to grow dramatically, a method that exploits normal daily computer interactions is attractive. This research explores the possibility of detecting cognitive and physical stress by monitoring keyboard interactions with the eventual goal of detecting acute or gradual changes in cognitive and physical function. Researchers have already attributed a certain amount of variability and ''drift'' in an individual's typing pattern to situational factors as well as stress, but this phenomenon has not been explored adequately. In an attempt to detect changes in typing associated with stress, this research analyzes keystroke and linguistic features of spontaneously generated text. Results show that it is possible to classify cognitive and physical stress conditions relative to non-stress conditions based on keystroke and linguistic features with accuracy rates comparable to those currently obtained using affective computing methods. The proposed approach is attractive because it requires no additional hardware, is unobtrusive, is adaptable to individual users, and is of very low cost. This research demonstrates the potential of exploiting continuous monitoring of keyboard interactions to support the early detection of changes in cognitive and physical function.