Eye tracking experiments are an important contribution to human computer interaction (HCI) research. Eye movements indicate attention, information processing, and cognitive state. Oculomotor activity is usually captured with high temporal resolution eye tracking systems, which are expensive and not affordable for everyone. Moreover, these systems require specific hard- and software. However, affordable and practical systems are needed especially for applied research concerning mobile HCI in everyday life. This study examined the reliability/validity of low temporal resolution devices by comparing data of a table-mounted system with an electrooculogram. Gaze patterns of twenty participants were recorded while performing a visual reaction and a surveillance task. Statistical analyses showed high consistency between both measurement systems for recorded gaze parameters. These results indicate that data from low temporal resolution eye trackers are sufficient to derive performance related oculomotor parameters and that such solutions present a viable alternative for applied HCI research.
TopIntroduction
Multiple studies have suggested the use of eye movement based measures as a tool to objectively monitor cognitive state and performance (e.g. de Rivecourt, Kuperus, Post, & Mulder, 2008; Stern, Boyer, Schroeder, Touchstone, & Stoliarov, 1994, 1996) due to the obvious and tight connection between eye movements and cognition as well as the predominantly visual nature of the tasks. For example, dwell time has been used as a global measure to assess workload or vigilance by evaluating points of interest and distribution of visual attention (Ahlstrom & Friedman-Berg, 2005; Alfredson, Nählinder, & Castor, 2003; Lavine, Sibert, Gokturk, & Dickens, 2002; Marshall, 2007; Stern et al., 1994, 1996). Furthermore, results from psychological research indicate that information from oculomotor parameters, i.e. fixations and saccades, can be used to assess cognitive performance (de Rivecourt et al., 2008) and fixations provide information about cognitive demands in reading (Rayner, 1998, 2009), process modelling (Pinggera et al., 2013), and visual search (Liversedge & Findlay, 2000). They are also sensitive to task specific involvement of memory (Geyer, von Mühlenen, & Müller, 2007) and may be used as indicators of hazard perception (Underwood, Chapman, Brocklehurst, Underwood, & Crundall, 2003; Velichkovsky, Rothert, Miniotas, & Dornhöfer, 2003). Saccades provide information about central activation and fatigue (Galley, 1989; Schleicher, Galley, Briest, & Galley, 2008) as well as shifting of the focus of attention (Chapman & Underwood, 1998; Underwood et al., 2003). Furthermore, gazing patterns are related to personality (e.g. Nitzschner, Nagler, Rauthmann, Steger, & Furtner, 2015; Rauthmann, Seubert, Sachse, & Furtner, 2012).