The statistical challenge of scan-path analysis

By knowing what people look at while performing tasks, we would be able to better design devices that facilitate users´ interaction. Eye tracking is a useful technique to study visual attention but the data are often spatially and temporally complex to analyze. To this aim more sophisticated analytical tools are needed to give meaning to data, especially when both the spatial and temporal components are to be taken in synchrony. In this paper we review statistical methods for spatial and temporal analysis of eye-movement data pinning down advantages and disadvantages while proposing new directions of research.