Controlling spatial attention without central attentional resources: Evidence from event-related potentials

Abstract

The present study examined whether the control of spatial attention requires central attentional resources using a modified Psychological Refractory Period paradigm. We varied across experiments whether Task 1 was a two- or four-choice speeded task and whether it was auditory or visual. Task 2 (unspeeded response) was to identify a visual letter in a specific target colour, while ignoring letters in other colours. We measured the N2pc effect (reflecting lateralized attentional allocation) elicited by Task 2 as a function of the stimulus–onset asynchrony (SOA) between Task 1 and Task 2. The question was whether spatial attention could shift to the Task 2 stimulus at short SOAs, while central attention was still allocated to Task 1. For the two-choice Task 1, Task 2 elicited a strong N2pc effect (indicating capture) even at short SOAs, regardless of whether Task 1 was auditory (Experiment 1) or visual (Experiment 2). But for the four-choice Task 1, the N2pc effect elicited by Task 2 was attenuated strongly at short SOAs, both for the visual Task 1 (Experiment 3) and the auditory Task 1 (Experiment 4). N2pc attenuation was also observed in Experiment 5, which mapped four auditory stimuli onto two responses for Task 1. This finding suggests that the attenuation is due to the difficulty of stimulus classification on Task 1, not the number of responses. Experiment 6 showed that the attenuation of N2pc effect on Task 2 was due to central operations on Task 1, not the mere presence of the Task 1 stimulus. We propose that controlling spatial attention without central resources is possible, but the quality of attentional settings degrades if concurrent tasks impose a sufficiently great load on working memory (especially the load related to stimulus classifications).

Keywords:

• Central attention
• N2pc effect
• Psychological refractory period
• Spatial attention

Acknowledgements

We thank Andrew Morgan at Oregon State University Enterprise Computing Service for providing technical support and building the customized response box for this study. We also thank Benoit Brisson, Charles Folk, and Robert McCann for their helpful comments on earlier drafts of this paper.

Notes

¹In our study, we measured the N2pc effect with respect to target stimuli that participants intended to attend. We therefore deliberately use the term “shift of spatial attention” rather than “attention capture” in this paper to avoid the controversial implication that the shifts were involuntary. Note, however, that some researchers have argued that involuntary attention capture (both capture by salient but irrelevant objects and contingent capture) shares some underlying processing mechanism with voluntary deployment of attention (e.g., Hickey, McDonald, & Theeuwes, 2006; Leblanc, Prime, & Jolicœur, 2008).

²It is logically possible that the exclusion of trials where Task 2 was incorrect served to eliminate the very trials where spatial attention was blocked during Task 1 processing. This seems unlikely, a priori, because there generally was no substantial increase in errors at short SOAs. Nevertheless, to investigate this possibility we conducted a follow-up N2pc analysis including trials where Task 2 was incorrect. As in the original data analyses, the N2pc effect was similar across SOAs, F<1.0, and pairwise comparisons revealed no significant differences between any two SOAs, Fs<1.0.

³Following Miller et al.'s (1998) jackknife method, we calculated n grand-average waveforms for each subset of n–1 participants (i.e., with a different participant being removed each time). The onset latency was determined as the time at which the N2pc effect reached the threshold of 0.3 µV. The values from the 100 ms SOA and 900 ms SOAs were submitted to a one-tailed t-test. Using this method, we found no statistically significant difference in the N2pc onset latency between these two SOAs in any of the present experiments (Experiments 1–6).