Abstract
Engaging attention at fixation with a task is widely used in brain imaging (fMRI) to control for the powerful effects of spatial attention. While the difficulty of a fixation task can modulate the processing of peripheral stimuli (Lavie 1995; Rees et al., 1997), little is known about cortical responses caused by attentive fixation tasks. We found that presenting a simple letter counting task at fixation strongly activated ventral and dorsal mid-level cortical areas, including spatial locations assigned to the periphery with retinotopic mapping.
Our intention was to control spatial attention with a baseline fixation task, while measuring responses to moving stimuli in the periphery. While peripheral stimuli were presented (16s ON-16s OFF), subjects counted the occurrence of a target (upright ‘T’s, 0.8 deg in size) amongst non-targets (inverted ‘T’s), and reported by button press whether target count was ‘odd’ or ‘even’. Surprisingly, the presentation of moving stimuli in the periphery generated only weak BOLD responses. To test whether the fixation task may have caused this, we contrasted the attentive fixation task with passive fixation of a small dot in the centre of a blank screen. As expected, we found foveal activation in early visual areas. However, we also found extra-foveal responses in ventral lateral occipital area (LOC) and dorsal area hMT/V5. This pattern was consistent across subjects, and significantly correlated to the task period as identified by both modelled (GLM) and model-free (MELODIC) analysis.
These results lend support to a role for ventral area LOC in shape perception, (Kourtzi et al., 2001) and hMT/V5 in letter reading (Demb et al., 1997). We conclude that the choice to employ an active fixation task that locks attention on certain visual features may have important consequences. Among these are the induction of a trade-off in cortical activity between regions-of-interest that serve central and peripheral vision.
Supported by the James S McDonnell Foundation and the Wellcome Trust.