Fig. 2

Simplified details of the ScanMatch method used to calculate similarity in scanning patterns. The top panel shows the fixations of two participants overlaid on an example stimulus. A grid-based region-of-interest (ROI) mask is also overlaid with a combination of lower- and upper-case letters used to define each ROI. The middle panel demonstrates the generation of letter sequences used to describe each participants’ search behaviour and the substitution matrix that provides information on the relationship between two ROIs. The bottom panel shows how two participants scanning behaviour are compared. The colour-coded substitution matrix is used to calculate an alignment score which indexes how similar two participants’ scanning behaviour is. Red squares indicate a high score because the two participants fixated in the same ROI whereas blue squares indicate a low score because the two participants fixated in different ROIs. Alignment scores are then normalised to generate a score between 0 (low similarity) to 1 (high similarity). See Cristino et al. (2010) for a full description of the ScanMatch method