The Role of Frontopolar Cortex in Subgoal Processing during Working Memory

• Published in: NeuroImage (Orlando, Fla.) Vol. 15; no. 3; pp. 523 - 536
• Main Authors: Braver, Todd S., Bongiolatti, Susan R.
• Format: Journal Article
• Language: English
• Published: United States Elsevier Inc 01.03.2002
• Subjects: Adolescent; Adult; Arousal - physiology; Attention - physiology; Brain Mapping; Dominance, Cerebral - physiology; Female; Goals; Humans; Magnetic Resonance Imaging; Male; Mental Recall - physiology; Paired-Associate Learning - physiology; Pattern Recognition, Visual - physiology; Prefrontal Cortex - physiology; Problem Solving - physiology; Psychomotor Performance - physiology; Reaction Time - physiology; Retention (Psychology) - physiology; Semantics
• ISSN: 1053-8119; 1095-9572
• DOI: 10.1006/nimg.2001.1019

Neuroimaging studies have implicated the anterior-most or frontopolar regions of prefrontal cortex (FP-PFC, e.g., Brodmann's Area 10) as playing a central role in higher cognitive functions such as planning, problem solving, reasoning, and episodic memory retrieval. The current functional magnetic resonance imaging (fMRI) study tested the hypothesis that FP-PFC subserves processes related to the monitoring and management of subgoals, while maintaining information in working memory (WM). Subjects were scanned while performing two variants of a simple delayed response WM task. In the control WM condition, subjects monitored for the presence of a specific concrete probe word (LIME) occurring following a specific abstract cue word (FATE). In the subgoal WM condition, subjects monitored for the presence of any concrete probe word immediately following any abstract cue word. Thus, the task required semantic classification of the probe word (the subgoal task), while the cue was simultaneously maintained in WM, so that both pieces of information could be integrated into a target determination. In a second control condition, subjects performed abstract/concrete semantic classification without WM demands. A region within right FP-PFC was identified which showed significant activation during the subgoal WM condition, but no activity in either of the two control conditions. However, this FP-PFC region was not modulated by direct manipulation of active maintenance demands. In contrast, left dorsolateral PFC was affected by active maintenance demands, but the effect did not interact with the presence of a subgoal task. Finally, left ventral PFC regions showed activation in response to semantic classification, but were not affected by WM demands. These results suggest a triple dissociation of function within PFC regions, and further indicate that FP-PFC is selectively engaged by the requirement to monitor and integrate subgoals during WM tasks.