Switch-Related and General Preparation Processes in Task-Switching: Evidence from Multivariate Pattern Classification of EEG Data

Abstract

The cued-trials task-switching paradigm is used to investigate the processes involved in preparation to change task. Task switch trials typically show poorer performance than task repeat trials, suggesting that additional or more time-consuming preparation processes are required to switch tasks. However, behavioral and neuroimaging studies have so far been unable to decipher whether preparing for a switch in task involves distinct cognitive processes to those required more generally on both switch and repeat trials. The current study addresses this question using a novel multivariate pattern misclassification analysis of frequency band-specific local topographical patterns in human EEG activity that was elicited by cues varying in information value. Within the alpha frequency band, misclassification analysis produced evidence for an early switch-related preparation process over right frontal cortex, as well as a later task readiness preparation process over right parietal cortex. This represents compelling evidence for dissociable switch-related and task readiness preparation processes that show distinct time course and spatial activation patterns.

Task Switching Processes

“…Importantly, noninformative cues did not elicit either switch positivity. The early switch positivity temporally overlaps alpha band activity over the right VLPFC (IFG) (Mansfield, Karayanidis, & Cohen, 2012), a region strongly linked to inhibitory control (Aron & Poldrack, 2006; see also consistent right IFG activity evident in fMRI ALE meta-analysis, Figure 3), and is consistent with disengagement from the now irrelevant task set (Meiran, Hsieh, & Chang, 2011). In contrast, the late switch positivity is consistent with fMRI evidence that proactive control involves activation of task-specific cortical regions in anticipation of the upcoming target (Wylie, Javitt, & Foxe, 2006).…”

Theta frontoparietal connectivity associated with proactive and reactive cognitive control processes

“…The amplitude of this early switch-positivity is associated with both response time and response threshold -a measure of cautiousness in decision-making (Karayanidis et al, 2009). Using multivariate pattern classification analyses, Mansfield et al (2012) showed activity in a right frontal source for switch relative to non-switch trials within the same latency range (see also Lavric et al, 2008). These studies also showed later activation in parietal sources for trials that permit task-specific preparation, irrespective of whether the task switches or repeats (Lavric et al, 2008;Mansfield et al, 2012).…”

“…Given neuroimaging evidence that preparation to switch and implementation of a switch response involve distinct networks (for review see Ruge et al, 2013), theta oscillatory synchronization for proactive and reactive control may be expected to involve distinct networks. Additionally, as previous work with this paradigm has shown increased alpha power when preparing to switch (Mansfield et al, 2012; see also Sauseng et al, 2006), we examine whether alpha oscillatory networks are also evident during switch preparation and/or implementation.…”

“…In a cued-Wisconsin Card Sorting task, cues associated with a rule switch led to sustained theta band activity across frontal and parietal sites, as well as a decrease in centro-parietal alpha band power (Cunillera et al, 2012). However, using a cued-trial paradigm, Mansfield et al, (2012) found differences in low-mid frequency oscillations between switch and repeat cues, relative to a non-informative cue baseline, over right frontal and bilateral parietal sites in time windows corresponding the early switchpositivity and pre-target negativity, respectively. In sum, these studies suggest a role for low frequency oscillations in task-switching, with differentiable effects at timelines consistent with proactive and reactive control.…”

“…Despite the utility of the task-switching paradigm, there are few studies examining the role of oscillatory activity during task-switching (Gladwin et al, 2006;Mansfield et al, 2012;Sauseng et al, 2006). For example, using a task-switching paradigm that did not differentiate between proactive and reactive control processes, Sauseng et al (2006) reported that switching tasks was associated with a reduction in upper alpha (10-13 Hz) power as well as a large increase in theta, but not alpha, coherence across fronto-posterior networks.…”