| Although brain networks supporting the memorization of color alone or location alone have been well studied, network activity underlying the integration of these features is still poorly understood. We used functional magnetic resonance imaging (fMRI) and magnetoencephalography (MEG) in combination with synthetic aperture magnetometry (SAM) to identify brain regions revealing changes in blood-oxygenation levels and oscillatory neural activity during encoding tasks for single features or the conjunction of features. In the single feature tasks, subjects were required to memorize a set of target colors while ignoring their locations (color task), or memorize a set of target locations while ignoring their colors (location task). In the color and location conjunction task, subjects were required to memorize pairings of specific colors and locations. A set of targets were presented repeatedly (encoding) with recognition tests interspersed among the encoding blocks. Behavioral data indicate that participants were able to successfully learn colors, locations, and color and location pairs. Memorizing conjunctions of colors and locations was the easiest task, while memorizing colors alone was the most difficult task. In the fMRI experiment we found that the left inferior frontal gyrus (BA 45/46) was preferentially activated for encoding of color alone, and the bilateral inferior parietal lobule (BA 40) and superior parietal lobule and precuneus (BA 7) and posterior middle temporal and occipital gyrus (BA 37/19) were preferentially activated for encoding of location alone. These regions showed reduced fMRI signals for encoding of color and location pairs relative to the single preferential feature condition. In the MEG experiment we found that theta band (4--8 Hz) power in the right medial precuneus (BA 7) and right posterior middle temporal and occipital gyrus (BA 37) was reduced for conjoint encoding of color and location relative to encoding of location alone. Our results suggest that the encoding and retention of single features while ignoring task-irrelevant features was associated with an increase in physiological response, or alternatively, the encoding and retention of conjoined features resulted in a decrease. The reduction in responses to conjoined features relative to preferred single features was interpreted as reflecting increased efficiency. |
