Accounting for priming in visual search: Episodic retrieval does not explain priming of pop-out

Pop-out visual search performance is not only influenced by bottom-up saliency but also by previous task history. If the target in such a task remains the same from one trial to the next performance is faster and more accurate than if the target identity changes (Maljkovic and Nakayama, 1994; Sigurdardottir et al., 2008). Maljkovic and Nakayama argued that this reflects feature facilitation, that attention is drawn to the feature defining the target on the preceding trials. This conception is challenged in episodic retrieval accounts (Huang et al., 2004; see also Hillstrom, 2000), where it is assumed that priming reflects a higher-level episodic memory representation of the search on the last trial. Here we show that the interaction between repetitions of different features critical for the episodic retrieval account applies only to a limited subset of tasks – in particular very difficult single-feature searches. We argue that feature facilitation is the most parsimonious accounts of priming of pop-out but episodic retrieval account have an important, albeit limited, in visual search. We conclude that a dual- or multi-stage account is needed to explain the heterogeneous results in the priming literature.