Abstract
Persistent activity in human frontal precentral and intraparietal sulci (PCS and IPS) sustains working memory (WM) representations over retention intervals (Jerde et al., 2012). Moreover, the fidelity of WM also depends on the efficient allocation of memory resources among multiple items competing for representation (Klyszejko et al., 2014). Here, we tested whether PCS and IPS may contribute differentially to the allocation of WM resources. We measured the effect on WM accuracy that repetitive transcranial magnetic stimulation (TMS) had when applied to retinotopic frontal and parietal cortex during the delay period of a visual spatial WM task requiring an uneven distribution of resources. On each trial, participants generated a memory-guided saccade (MGS) to the location of one of two items maintained over a memory delay. Critically, they were cued before the trial which of the two items was more likely to be tested, and therefore, should allocate more WM resources to prioritize the storage of that item. We applied rTMS during the delay period to superior PCS (sPCS) and IPS (IPS2) defined using a novel fMRI-based population receptive field (pRF) mapping technique (Mackey et al., 2017). In the absence of rTMS, MGS were more accurate to high-compared to-low prioritized targets, confirming past work demonstrating the flexible allocation of WM resources (Klyszejko et al., 2014; Emrich et al., 2017; Bays et al., 2009). Perturbation to IPS2 caused a non-selective worsening of MGS accuracy to both high and low priority targets. In contrast, sPCS perturbation removed the benefit of priority, as if the strategic allocation of resources was impaired. These results support a framework whereby sPCS enables an efficient allocation of WM resources, while IPS supports WM representations themselves. Together, these results demonstrate the nuanced role each region plays with regard to representing versus prioritizing information.
Meeting abstract presented at VSS 2018