The left lateral frontopolar (LFP) cortex showed dimension change-related activation in previous event-related functional magnetic resonance imaging studies of visual singleton feature search with non-brain-lesioned participants. Here, we tested the hypothesis that LFP actively supports changes of attention from the old to the new target-defining dimension in singleton feature search. Singleton detection was selectively slowed in this task when the target-defining dimension changed in patients with left LFP lesions, compared with patients with frontomedian lesions as well as with matched controls without brain lesions. We discuss a potential role of LFP in change detection when the optimal allocation of dimension-based attention is not clearly defined by the task.

Four-dot masking is a new form of visual masking that does not involve local contour interactions or spatial superimposition of the target stimulus and the mask (as, e.g., in pattern or metacontrast masking). Rather, the effective masking mechanism is based on object substitution. Object substitution masking occurs when low-level visual information representations are altered before target identification through iterative interaction with high-level visual processing stages has been completed. Interestingly, object substitution interacts with attention processes: Strong masking effects are observed when attentional orientation toward the target location is delayed. In contrast, no masking occurs when attention can be rapidly shifted to and engaged onto the target location. We investigated the neural basis of object substitution masking by studying the interaction of spatial attention and masking processes using functional magnetic resonance imaging. Behavioral data indicated a two-way interaction between the factors Spatial Attention (valid vs. invalid cueing) and Masking (four-dot vs. pattern masking). As expected, spatial attention improved performance more strongly during object substitution masking. Functional correlates of this interaction were found in the primary visual cortex, higher visual areas, and left intraparietal sulcus. A region-of-interest analysis in these areas revealed that the largest blood oxygenation level-dependent signal changes occurred during effective four-dot masking. In contrast, the weakest signal changes in these areas were observed when target visibility was highest. The data suggest that these areas represent an object substitution network dedicated to the generation and testing of a perceptual hypotheses as described by the object substitution theory of masking of Di-Lollo et al. [Competition for consciousness among visual events: The psychophysics of reentrant visual processes. Journal of Experimental Psychology: General, 129 , 481–507, 2000].

Objects characterized by a unique visual feature may pop out of their environment. When participants have to search for such “odd-one-out” targets, detection is facilitated when targets are consistently defined within the same feature dimension (e.g., color) compared with when the target dimension is uncertain (e.g., color or motion). Further, with dimensional uncertainty, there is a cost when a given target is defined in a different dimension to the preceding target, relative to when the critical dimension remains the same. Behavioral evidence suggests that a target dimension change involves a shift of attention to the new dimension. The present fMRI study revealed increased activation in the left frontopolar cortex, as well as in posterior visual areas of the dorsal and ventral streams, specific to changes in the target dimension. In contrast, activation in the striate cortex was decreased. This pattern suggests control of cross-dimensional attention shifts by the frontopolar cortex, modulating visual cortical processing by increased activation in higher-tier visual areas and suppression of activation in lower-tier areas.