Working memory is one of the most fundamental building blocks of human cognition. It is an integral part of most mechanisms involved in higher-level cognitive function such as fluid intelligence, reading comprehension, and decision-making (Conway et al., 2005; Curtis & Lee, 2010). Current theory implicates the prefrontal cortex (PFC), specifically the dorsolateral prefrontal cortex (dlPFC), as a critical site for maintaining spatial locations in working memory. This is largely due to two critical findings from monkey studies of working memory: (1) neural activity in dlPFC persists during working memory delay periods and (2) dlPFC lesions cause impairments in remembering information in the contralesional visual field (Funahashi et al., 1989, 1993). Based on this animal model of human working memory, we tested the hypothesis that lesions to human dlPFC will impair spatial working memory performance in the contralesional visual field. We recruited patients with PFC lesions (n=13) and a neurologically healthy comparison group (n=10) that was matched in age and education. We used a memory-guided saccade task, similar to that used in the monkey studies, to assess spatial working memory performance, and used a visually-guided saccade task to control for any visual or oculomotor deficits. We evaluated performance by measuring saccade gain, spatial error, and reaction time. Surprisingly, only a few PFC patients had impaired working memory performance. Those that had lesions to the precentral sulcus (PCS) had lower gains and larger errors on the memory-guided saccade task. Additionally, performance in the contralesional hemifield was worse than in the ipsilesional hemifield. Importantly, these performance deficits were not present in the visual control task ruling out visuomotor deficits. These deficits were also independent of lesion size, patient age, and time since surgery. Those patients whose lesions spared the PCS, including damage to the dorsal convexity, were completely unimpaired. Therefore, the dlPFC is not necessary for spatial working memory. Although incompatible with data from animal studies, these findings are consistent with human neuroimaging studies showing persistent delay period activity in the PCS, but not in more anterior portions of the dlPFC (Srimal & Curtis, 2008). Past human lesion studies examining the role of PFC in verbal or spatial working memory tasks have been inconsistent (Ploner et al., 1999; D’Esposito & Postle, 1999). Taken together, these results suggest that theories of the role of the dlPFC in human working memory needs revision.


Sunday, Nov 16, 2014,WCC Hall A-C, 11:00 – 12:00 PM