B.3.6.5. Conclusion

In summary, both salience and relevance appeared to modulate performances in a visual search task by biasing the competition between items in the perceptual system, as suggested by the biased competition hypothesis (Desimone and Duncan, 1995; Duncan, 2006). Consistently, both signals are not indefinitely independent, since they could genuinely integrate at some locus when they concerned the same perceptual dimension. The influence of similarity on this integration suggested that these signals combined within the perceptual system. These results encouraged to think visual selective attention as an emergent property of the competition between objects taking place in the perceptual system, and endogenous and exogenous modes of orienting as two ways of biasing the competition. Indeed, this perspective could naturally account for relevance and salience effects, as well as for their genuine integration when they concerned the same dimension. Finally, the present results also invite questioning the salience map hypothesis, and considering the issue of "attentional selection" from the grounded cognition perspective.