Computer users can be distinguished from one another based on differences in their typing rhythms. Our research extends this idea to ask whether a user's level of anxiety or stress can also be determined, but based on *changes* in typing rhythms. Thus, our primary research objective is to answer a single scientific question in a laboratory study: Do typing rhythms change when a typist is under stress, such that the change can be measured and detected with a standard keyboard? Why would anyone want to do this? There are at least two reasons: (1) people under stress tend to make more errors at the keyboard/console, which can affect job accuracy and reliability; and (2) insiders who abuse their legitimate access to sensitive computer systems are likely to show signs of stress in the execution of a nefarious act, such as cyber espionage. In each of these examples, remote detection of typist stress, through the use of the common keyboard, would alert supervisory personnel to a potential problem that could be averted. In addition to having a major effect on error prevention in national critical infrastructure controls, and on insider detection in cyber space, the impact of this research in health care could also be substantial. For example, subtle changes in typing rhythms could indicate early onset dementia, episodes of cognitive decline, or even a tendency for musculoskeletal disorders such as carpal tunnel syndrome or digital flexor tendinitis. As this research develops, so will the range of beneficial uses.
TWC: Small: Safer Computing through Biometric Stress Detection