What factors constitute this priority map? Studies have shown that many factors influence the guidance of eye movements. These factors can be generally grouped into task-independent, or
bottom-up features, and
task-relevant features. Examples of bottom-up features are luminance contrast (Reinagel & Zador,
1999; but see Einhäuser & König,
2003), color contrast (Itti, Koch, & Niebur,
1998), energy (Ganguli, Freeman, Rajashekar, & Simoncelli,
2010), and saliency (Itti & Koch,
2001; Itti et al.,
1998); examples of task-specific features are relevance (resemblance to search target; Einhäuser, Rutishauser, & Koch,
2008); learned context about target location (Ehinger, Hidalgo-Sotelo, & Torralba,
2009); factors that are ecologically relevant but not specific to the task, such as intrinsic value (Gottlieb,
2012); and exploratory strategies for minimizing uncertainty (Gottlieb, Oudeyer, Lopes, & Baranes,
2013; Renninger, Verghese, & Coughlan,
2007). In addition, change in direction between successive saccades (Wilming, Harst, Schmidt, & König,
2013) and natural statistics of saccade magnitude and direction (Tatler & Vincent,
2009) also enable saccade prediction. Predictive models of eye movements are derived from priority maps comprising one or more of the above factors.