Recently a number of investigators (e.g., Arend et al. 2006; Olivers & Nieuwenhuis, 2005, 2006) have reported the beneficial effects of an "unfocused" attentional mode in the temporal allocation of attention while selecting two targets in the attentional blink (AB) paradigm. An 'overinvestment' hypothesis to account for this benefit has been considered, suggesting that the AB is in fact produced by an 'over-allocation' of attention to the T1 task. To investigate the mechanism underlying the beneficial effect of this unfocused attentional mode the present investigators recorded ERPs while participants took part in a study designed after Arend et al. to attenuate the AB. Separate participants (N = 12) took part in two conditions. In the 'motion' condition, a flowfield of moving dots appeared behind the fixation point two seconds prior to the onset of the RSVP stream and remained for the duration of the trial. This condition was designed to remove attention from the locus of the RSVP stream. In the control 'static' condition an identical number of non-moving dots were placed on the screen in the same locations as in the motion condition. Whereas the 'static' condition showed a normal AB, the 'motion' condition revealed a significantly attenuated AB as was found in Arend et al. (2006). No difference was found for T1 accuracy across conditions. ERPs from a 64-channel (BrainProducts UK) system revealed a slow negative potential shift, similar to the contingent negative variation (CNV), which developed between fixation and RSVP onset. This CNV exhibited a topographical maximum at parietal sites, was larger over the right as compared to the left hemisphere, and importantly was significantly larger before no AB as compared to AB trials in the "motion" group only. These results are discussed in terms of the over-investment hypothesis.