B.2.2. Working memory and attentional tasks

B.2.2.1. Resistance to interference

One other condition in which resistance to interference could be disrupted is when working memory (WM) capacity is limited in one way or another, for instance in the realm of inter-individual variations. Indeed, high WM span individuals were shown to perform better than low WM span individuals in, for instance, the Stroop test (Kane & Engle, 2003) or the anti-saccade task (Unsworth, Schrock, & Engle, 2004), and to be less sensitive to the "cocktail party effect" (Conway, Cowan, & Bunting, 2001). Higher WM capacities seem to allow resisting interference more efficiently, perhaps through inhibition. Similar conclusions about resistance to interference were drawn from manipulations of WM by dual-tasking (e.g. de Fockert, Rees, Frith, & Lavie, 2001; Roberts, Hager, & Heron, 1994). Mitchell et al. (2002) confirmed the results of Roberts et al. (1994) and found that an executive WM load (an n-back task) impaired performances when a reflexive (pro-)saccade had to be "resisted", no matter a saccade had to be made (anti-saccade condition) or not (no-saccade condition).

More specifically focused on visual selective attention, Lavie and de Fockert (2005) suggested that resisting AC by a salient onset distractor also depended on WM. They proposed their participants to achieve concurrently an "irrelevant singleton" visual search task (Theeuwes, 1992), and a second task whose demands on WM resources varied. In the no-distractor condition, performances did not vary with WM load, which might suggest that attentional orienting was not dependant on WM resources (but see below). On the contrary, distraction increased with loading WM, suggesting that resistance to interference was specifically impaired by the WM load. The authors interpreted these results in the terms of the load hypothesis (Lavie, 2005). In its entirety, this hypothesis predicted opposite effects from perceptual versus cognitive loads. First, increasing the perceptual load (e.g. through the number of items or their visual complexity) should decrease the influence of salient distractors, due to an exhaustion of available "perceptual resources". Recently, Torralbo and Beck proposed that the perceptual load was actually determined by the competition within the visual cortex. Second, increasing the cognitive load should increase AC. Indeed, Lavie (2005) claimed that WM general resources were necessary to control the visual search, especially in competitive situations. This latter claim was supported by their experiments. Sala and Courtney (2007; Courtney, Roth, & Sala, 2007) recently proposed an account of WM mechanisms that could shed light on this dependency. Unfortunately, as we argued below, the incentive to resist interference was probably quite low, due to the experimental paradigm used. Therefore, the data of Lavie and de Fockert (2005) would have to be confirmed through a different paradigm. A second point deserved investigation. These experiments showed no influence of WM on a non-competitive search (i.e. in absence of the salient distractor). Notwithstanding, the processes involved in visual search were also shown to be dependent on WM.