Localization of spatial attention processes with the aid of a probe technique.

A sudden visual onset is thought to 'attract attention to its location' within less than 100 ms. We attempted to measure the effect of this attentional process on the event-related potential (ERP) to a probe presented about 140 ms after the onset, and to delineate the spatiotemporal characteristics of such an effect, if any. ERPs were recorded from 30 channels from 6 subjects while they performed a target detection task. Both targets and probes could be located in each of the 4 quadrants (eccentricities 6.1 degrees and 7 degrees, respectively). For a given single target, the subsequent probe was either presented near the location of the target ('valid target') or at the diagonal opposite ('invalid target'). Appropriate 'neutral' conditions (probes preceded by no target, or by simultaneous targets in all quadrants) were applied, and ERPs to probes were corrected for the contribution of the ERPs to targets. The earliest effect of (in)validity was found at about 120 ms after probe onset for lower field probes. This effect consisted of enhanced posterior positivity for valid relative to neutral relative to invalid conditions. This positivity was superposed on a contralateral, extrastriate negative ongoing wave peaking at about 150 ms ('N150'). Source localization suggested that the (in)validity effects originate from deep medial parietal areas. The source corresponding to the N150 activity was not influenced by (in)validity. An earlier deflection to the probe at 80 ms ('NP80') depended on location, but not on (in)validity, and seemed to be of striate origin. Results are discussed in terms of a model postulating an attention-independent 'input module' from which activation is fed to a 'location module' embodying the actual attention mechanism.