Endogenous control of visual spatial attention switching: A neurophysiological approach.

摘要

Temporal dynamics and cortical spatial location of the control mechanisms for the switching of visual attention across space (without eye movements) were examined with behavioural measures and the brain-imaging techniques of Functional Magnetic Resonance Imaging (FMRI) and Evoked-Response Potentials (ERPs). Subjects monitored multiple-frame displays (presented at 1.6 Hz) on which digits were repeated in numeric sequence, and responded whenever a digit occurring out of sequence was detected (e.g., monitor ascending digits and respond to errors such as the first '7' in the example: 1,2,3,7,5,6,7,8,1,2,...). There were 4 digits on each frame (one in each quadrant), 5 experimental, and 2 control conditions. Subjects fixated eyes on a central cross while attention either remained fixed on (1) the lower left digit, or (2) the lower right digit, or attention switched on each successive frame between (3) the left-most digits, (4) the right-most digits, or (5) the bottom digits. FMRI was performed on conventional MRI scanners using a spiral k-space pulse sequence. 6 oblique slices were acquired, using a 5 inch surface coil centered dorsally over the parietal occipital boundary. Activation images were generated by calculating pixel-by-pixel split-half t-test and percent difference maps. A region lateralized to the right superior parietal cortex was active when attention switched between locations, independent of which locations in the visual field. This area was less active when attention was fixed in one location. Areas in the left superior parietal and right inferior parietal cortices were also implicated. ERPs from a 64 electrode scalp montage were processed to reveal temporal dynamics. Attentional effects (switching versus fixed) occurred for 3 components. The N1 component may index the orienting of attention in the switching condition, the P2 may be a preparatory response before the attentional shift, and the attention switching effect within the 250 to 450 ms range may index the disengaging of attention in the switching condition. Support was provided for the hypothesis of a critical period during which the timing of the disengagement of attention is sensitive to the offset of the attended stimulus.