Central as well as peripheral attentional bottlenecks in dual-task performance activate lateral prefrontal cortices

Abstract

Human information processing suffers from severe limitations in parallel processing. In particular, when required to respond to two stimuli in rapid succession, processing bottlenecks may appear at central and peripheral stages of task processing. Importantly, it has been suggested that executive functions are needed to resolve the interference arising at such bottlenecks. The aims of the present study were to test whether central attentional limitations (i.e., bottleneck at the decisional response selection stage) as well as peripheral limitations (i.e., bottleneck at response initiation) both demand executive functions located in the lateral prefrontal cortex. For this, we re-analysed two previous studies, in which a total of 33 participants performed a dual-task according to the paradigm of the psychological refractory period (PRP) during fMRI. In one study (N=17), the PRP task consisted of two two-choice response tasks known to suffer from a central bottleneck (CB group). In the other study (N=16), the PRP task consisted of two simple-response tasks known to suffer from a peripheral bottleneck (PB group). Both groups showed considerable dual-task costs in form of slowing of the second response in the dual-task (PRP effect). Imaging results are based on the subtraction of both single-tasks from the dual-task within each group. In the CB group, the bilateral middle frontal gyri and inferior frontal gyri were activated. Higher activation in these areas was associated with lower dual-task costs. In the PB group, the right middle frontal and inferior frontal gyrus were activated. Here, higher activation was associated with higher dual-task costs. In conclusion we suggest that central and peripheral bottlenecks both demand executive functions located in lateral prefrontal cortices. Differences between the CB and PB groups with respect to the exact prefrontal areas activated and the correlational patterns suggest that the executive functions resolving interference at least partially differ between the groups.