Evaluation of the ergonomic design quality of work systems or the usability of technical system components (e.g., software products; see ISO, 1998) requires the evaluation of mental workload. >From an occupational safety and health perspective, an evaluation of mental workload is mandatory for compliance with legal regulations within the European Union (e.g., Occupational Safety and Health Directive 89/591 EEC, 1989; Video Display Unit Directive 90/270 EEC, 1990) for certain positions--for example, for jobs at workstations with visual display units in the office as well as in process control operations. The intention is to protect workers against the risk of impairing effects and to provide for optimal working conditions to promote their well-being and the reliability and productivity of the work system (Nachreiner, 1995, 1998). This calls for suitable and practical methods for the assessment and evaluation of mental workload at the workplace (Nachreiner, 1999).
Besides performance measures and subjective ratings, psychophysiological measures--especially heart rate (HR) and heart rate variability (HRV)--have a long tradition of being used for that purpose, perhaps because of their assumed relative simplicity and nonreactivity. Heart rate and its derivates are thus among the most practical and most commonly used methods for the assessment of mental workload imposed on an operator (Meshkati, Hancock, Rahimi, & Dawes, 1999). Although HR has been used since the 19th century for indicating mental processes (e.g., Lehmann, 1899; Mosso, 1892; Zoneff & Meumann, 1903), it is only one of several parameters of the complex homeostatic cardiovascular system reflecting energetic, thermoregulatory, respiratory, and/or emotional processes, in addition to cognitive processes. Therefore, measures of HRV in the time and frequency domain have recently been preferred in attempts to decompose different processes affecting cardiovascular regulation (e.g., Boucsein & Backs, 2000; Van Dellen, Aasman, Mulder, & Mulder, 1985).
Among the available HRV measures, the suppression of spectral power of the 0.1-Hz component of HRV has not only been considered attractive and promising (O'Donnell & Eggemeier, 1986) but even advocated as a sensitive and diagnostic indicator of mental workload (Mulder & Mulder, 1987; Wierwille & Eggemeier, 1995), until recently (Boucsein & Backs, 2000; Luczak, 1998). This is based on the assumption that HRV allows one to differentiate among different levels and kinds of mental workload in information processing--for example, automated versus controlled information processing (Mulder, Mulder, Meijman, Veldman, & Van Roon, 2000), with the latter requiring more effort and thus leading to a suppression of HRV.
In recent years this measure has been used with increasing frequency for analyzing mental workload in real or simulated work situations (Izso, 2001; Mulders, Meijman, O'Hanlon, & Mulder, 1982; Richter, Wagner, Heger, & Weise, 1998; Tattersall & Hockey, 1995; Wiethoff, 1997) and appears to have become an established, standard procedure. However, HRV sometimes seems to be used without any consideration of its psychometric properties (e.g., Izso & Wiethoff, 1997), despite the lack of systematic and comprehensive studies investigating these properties. Based on a closer inspection of the relevant literature and some recent research experience, serious doubts concerning the validity of this measure have been raised (Apparies, Riniolo, & Porges, 1998; Caldwell et al., 1994; Eilers, 1999; Manzey, 1998; Nachreiner, Grzech-Sukalo, & M6hlmann, 1999; Nickel, Nachreiner, Zdobych, & Yanagibori, 1998; Paas, Van Merrienboer, & Adam, 1994; Veltman & Gaillard, 1996; Wilson, 1992) that urgently call for systematic and, preferably, experimentally controlled analyses of the psychometric properties of this proposed indicator of mental workload. …