55, P = 0.032), Time (F1,15 = 5.26, P = 0.037) and Region (F1,15 = 6.45, P = 0.023), and a Deforolimus order Trial × Time × Region (F1,15 = 8.23, P = 0.012) interaction. Region-specific tests confirmed that a trend towards a Trial × Time interaction was only evident over

the parietal-occipital scalp region (F1,15 = 3.97, P = 0.06). The within-modality anova revealed a main effect of Trial (F1,15 = 5.55, P = 0.032) and a Trial × Time × Region (F1,15 = 8.23, P = 0.012) interaction. Region-specific tests confirmed that a trend towards a Trial × Time interaction was only evident over the parietal-occipital scalp region (F1,15 = 3.98, P = 0.06). The behavioral data did not exhibit any overt indication of a classical local switch cost. However, in light of the current findings regarding alpha oscillatory processes and as suggested by a reviewer, we sought to probe deeper into the behavioral data in order to explore the relationship of the relative

behavioral success of a given task-set reconfiguration to the current findings in the oscillatory domain. Certainly prior work has shown links between the effectiveness of alpha-band deployment mechanisms and subsequent task success (Thut et al., 2006; Kelly et al., 2009). To do this, we undertook a post hoc analysis in which we sorted individual trials based on RT. On an individual participant basis, we split KU-60019 experimental trials based upon the median RT within a given condition (i.e. repeat-auditory, switch-auditory, repeat-visual and switch-visual). Dividing each of these original four conditions by the median of the RT distribution yielded what we will refer to as ‘fast’ and ‘slow’ conditions for each participant and for each of the original conditions. The reasoning behind this approach is that a fast-switch trial reflects a more successful task-set reconfiguration than a slow-switch trial. This comes with the necessary caveat that a raw RT value on any given trial is by no means a direct index of successful task-set reconfiguration. That is, a relatively fast response on a switch

trial is not a pure index Cell press of a successful switch but necessarily indexes the multiple underlying neural events that give rise to the stochastic nature of RT. Thus, in an attempt to bolster the relevance of fast and slow trials to the successful instantiation of a new task set, we performed the following additional analysis. First, both hit trials (a correct response on a go trial) and false alarm (FA) trials (a mistaken response on a no-go trial) were included in the RT distributions of each of the experimental conditions. Next, after performing the median splits of these distributions, the proportion of hits relative to false alarms was calculated [i.e. hits/(hits + FAs)] yielding what we will refer to as the success rate. Behavioral success rates were then submitted to a 2 × 2 × 2 repeated-measures anova with factors of Modality (visual vs. auditory), Trial (switch vs. repeat) and Speed (fast RTs vs. slow RTs).