B.2.7.2. Cognitive load

The second major hypothesis of the present experiments concerned the AC variations with the executive WM load. The expected effect was evidenced in the first experiment, when no sound was heard in the control condition. In the second experiment, however, when the auditory list was heard in both conditions, the effect was observed only in the size-3 condition, that is, when the perceptual load was lower. This difference between Experiments 1 and 2 might perhaps be due to the presence of noise in the ST. Noise could be distracting by itself (Dalton & Behm, 2007). Thus, participants might have had to resist the distraction induced by the noise in the single task of Experiment 2. Since WM and resistance to interference were thought to interact, it is possible that the presence of noise in the single participated in minimizing the WM effects in the second experiment. However, this interpretation is only speculative. The lack of significance of the interaction between WM and condition might also be due to a mere statistical type-2 error. Indeed, the four-way interaction of the general ANOVA was not significant, while the expected effects of cognitive and perceptual load on AC were significant at this general level.

The variation of AC was taken as a cue of the abillity to resist interference. In the present experiments, the incentive toward the irrelevant salient distractor was strictly controlled. Indeed, this latter was made completely irrelevant. Moreover, indirect incentive toward salience was also avoided by cueing it endogenously and by keeping it unsalient. Even in these conditions, which maximized the incentive to resist salience-induced interference, the attentional capture increased with WM load. These results confirmed those of Lavie and de Fockert (2005).

On the other hand, at odd with Lavie and de Fockert (2005), loading WM also impaired visual search in absence of salient distractors. This difference was expected for at least two reasons. First, the WM task was more executive in the present experiments. Yet, Han and Kim (2004) showed that the executive nature of the concurrent task was crucial to impair a visual search task. Second, the target was not salient and could not be selected on the basis of exogenous signals, as was the case in Lavie and de Fockert (2005). This was consistent with the hypothesis that only endogenous processes were disrupted by WM load.

Several effects of WM on selective attention, observed in the present experiments, might advantageously be described in the terms of the biased-competition hypothesis (Sala & Courtney, 2007; Duncan et al., 1996; 2008). First, loading WM induced a general RT increase in the visual search. This effect was significant whether or not the salient distractor was present. Second, the specific influence of this salient distractor, the AC, was also increased by WM load. Third, these effects of WM (on simple visual search and on AC) were correlated. This correlation will be described in the next section, before construing the results in the biased-competition framework.