Visuospatial attention:: how to measure effects of infrequent, unattended events in a blocked stimulus design

This fMRI study investigates the differences between a blocked and event-related analysis in a cued target detection task, the so-called Posner paradigm, using a hybrid design. Validly and invalidly cued trials were presented intermingled in different blocks containing 50 %, 75 %, or 100 % valid trials. Four analyses were conducted: (i) an event-related analysis comparing invalid and valid trials, (ii) a blocked analysis comparing blocks with 50 % valid and invalid trials to blocks with 100 % valid trials, (iii) a blocked analysis detecting differences between block models when modeled as epochs or chains of events, and (iv) a blocked analysis that modeled blocks as chains of events to scale regressors equally to the event-related analysis. Irrespective of the type of analysis (blocked or event related), significant activation of the right intraparietal sulcus was observed. A larger cluster size was evident in the blocked analysis, which can be attributed to higher efficiency. In addition to this common right parietal activation, the event-related analysis revealed activations in right superior parietal cortex and left intraparietal sulcus. In contrast, the blocked analysis yielded additional activity in the right occipitoparietal junction. No influences of the block model (epoch versus chain of events) were found in regions activated in the blocked or event-related analysis, respectively. In summary, using a hybrid design and both event-related and blocked analysis techniques, we show both sustained and transient neural processes underlying reorienting of visuospatial attention. (C) 2004 Elsevier Inc. All rights reserved.