B.3.6.4.2. Perceptual locus

A second conclusion of this experiment is that this integration probably occurred within a cognitive perceptual sub-system dedicated to processing the feature involved in both signals (i.e. colour). First, the expected two-way interaction was observed between salience (neutral, target salient), relevance (uncued, cued) and salience-relevance perceptual similarity (similar vs. different). This interaction directly supported the hypothesis: the integration between salience and relevance effects was stronger when both processes were perceptually similar than different. The fact that perceptual similarity influenced the integration between endogenous and exogenous processes suggested that they interacted within the visual system, and thus, logically, that both of them occurred at least partly therein. Second, only when both signals concerned colour, a genuine integration was evidenced by the Race Model inequality. Moreover, this integration was significantly stronger in the similar than in the different condition. This not only confirmed that salience and relevance effects emerged within the perceptual system, but also that they genuinely integrated therein. More precisely, this integration probably occurred in the cognitive some sub-system specialized in processing colour.

Even if this point was never addressed directly by neurophysiology, to our knowledge, it is highly consistent with the available data concerning both salience and feature-based endogenous orienting. On the one hand, several studies suggested that salience emerged from competition in specific feature maps within the visual system (Beck & Kastner, 2005 ; Knierim & Van Essen, 1992; Nothdurft et al., 1999), for instance in a colour processing map, perhaps in V4 for colour-based salience (Burrows & Moore, 2009). This salience enhancement might be based on a decrease of the suppressive influence of neighbours (Knierim & Van Essen, 1992, but see Rauschenberger, 2010). On the other hand, cueing a relevant feature also seemed to favour this feature's processing within its specific visual feature map (e.g. Anlle-Vento et al., 1998; Bichot et al., 2005; Luck et al., 1997; Schoenfeld, 2007), probably through feed-back excitatory connections (Corbetta & Shulman, 2002; Miller & Cohen, 2001).

A biased competition view of visual selective attention seems to be the most evident account of the present results. Indeed, this perspective predicted that endogenous and exogenous processes would modulate the perceptual processes directly, salience as a contextual modulation, relevance as a descending influence from templates in WM on perceptual processes. The fact that salience and relevance emerged within the perceptual system is congruent with these hypotheses, as well as with many neurophysiological data. However, the hypothesis that endogenous and exogenous processes integrate at this perceptual locus, suggested by the present behavioural data, would probably deserve direct neurophysiological evaluation.