Practical Neurophysiological Analysis of Readability as a Usability Dimension

Abstract

Abstract. This paper discusses opportunities and feasibility of integrating neurophysiologic analysis methods, based on electroencephalography (EEG), in the current landscape of usability evaluation methods. The rapid evolution and growing availability of low-cost, easier to use devices and the accumulated knowledge in feature extraction and processing algorithms allow us to foresee the practicality of this integration. The work presented in this paper is focused on reading and readability, identified as a key element of usability heuristics, and observable in the neurophysiologic signals' space. The experiments are primarily designed to address the discrimination of the reading activity (silent, attentive and continuous) and the verification of decreasing readability, associated with the user's mental workload analysis. The results obtained in the series of experiments demonstrate the validity of the approach for each individual user, and raise the problem of intersubject variability and the need for designing appropriate calibration procedures for different users. Keywords. Usability analysis, neurophysiologic signals, EEG IntroductionThe current usability evaluation methods range from interpretative to predictive, the former based on the observation and study of the actual use of an artifact during its development cycle, and the latter exploring external analyses performed by experts equipped with standards, heuristics and modeling techniques [1]. The methods capture either the behavior and perceptions of users or the interactive attributes of the artifact, and converge to evaluation conclusions based on the qualitative or quantitative analysis of empirically collected data. Alternative, or complementary, methods based in the measurement of physical and physiological signals of the human user (e.g. eye-movements, heart rate (HR) and heart rate variation (HRV), skin conductance (SC), or electroencephalography (EEG)) have been used more frequently in contexts with critical requirements for human performance [2], and tested in dedicated labs, as opposed to the use of the former empirical methods, typically adopted by usability labs participating in the design and development of interactive artifacts for the general user or consumer.The evolution of the technological landscape leads us reassess the opportunities for expanding and improving the set of tools for usability evaluation. First, capture devices are cheaper, more reliable, less intrusive, and usable with increasing autonomy. Examples of this evolution are HR or SC measurement devices, which have become portable and wireless and even EEG systems have left the controlled conditions of clinical settings (see for example, the design of dry electrode devices [3]. Second, the capacity to process these signals evolved dramatically, both in computing power and in the understanding of the algorithms that extract meaning from the human data.As a result of this evolution, we can envision a feasible integration of human physical and physiological information in comm...