Hemispheric asymmetries and cognitive flexibility: An ERP and sLORETA study

Abstract

Although functional cerebral asymmetries (FCAs) affect all cognitive domains, their modulation of the efficacy of specific executive functions is largely unexplored. In the present study, we used a lateralized version of the task switching paradigm to investigate the relevance of hemispheric asymmetries for cognitive control processes. Words were tachistoscopically presented in the left (LVF) and right visual half field (RVF). Participants had to categorise the words either based on their initial letters, or according to their word type. On half of the trials the task changed (switch trials) whereas on the other half it stayed the same (repeat trials). ERPs were recorded and the neural sources of the ERPs were reconstructed using standardised low resolution brain electromagnetic tomography (sLORETA). In the word type task, participants were faster on repeat trials when stimuli were presented in the RVF. In contrast, in the initial letter task participants were faster on repeat trials and in general more accurate after stimulus presentation in the LVF. In both tasks, no hemispheric asymmetries in reaction times were observed on switch trials. On the electrophysiological level, we observed a left lateralization of the N1 that was mediated by activation in the left extrastriate cortex as well as a greater positivity of the P3b after stimulus presentation in the RVF compared to the LVF that was mediated by activation in the superior parietal cortex. These results show that FCAs affect the neurophysiological correlates of executive functions related to task switching. The relation of neurophysiological and behavioural asymmetries is mediated by task complexity, with more complex tasks leading to more interhemispheric interaction and smaller left–right differences in behavioural measures. These findings reveal that FCAs are an important modulator of executive functions related to cognitive flexibility.

Introduction

Cognitive control processes mediated via basal ganglia-prefrontal loops play a major role in the organisation of human behaviour and include mechanisms related to inhibition (Sehlmeyer et al., 2010), selection (Beste, Saft, Andrich, Gold, & Falkenstein, 2008) and correction of erroneous actions (Beste, Willemssen, Saft, & Falkenstein, 2009). A key aspect of cognitive control is cognitive flexibility, the ability to swiftly switch between different tasks. One of the major experimental paradigms to investigate this aspect of executive functions is the task switching paradigm (Allport et al., 1994, Karayanidis et al., 2003, Kiesel et al., 2010, Monsell, 2003). In cued task-switching experiments, a cue at the beginning of each trial indicates which task out of a set of two or more the participant has to perform in this trial. On so-called switch trials the task that has to be performed changes compared to the trial before, whereas on repeat trials it does not. Typically, participants react slower on switch than on repeat trials, a phenomenon called switch costs (Jamadar et al., 2010, Rogers and Monsell, 1995, Wylie and Allport, 2000).

Performance in paradigms that assess executive functions is influenced by information processing in the bottom-up channel (e.g. Knudsen, 2007). However, depending on the type of stimuli used, one hemisphere is more efficient in processing than the other. For example, the left hemisphere is more efficient in processing verbal stimuli than the right hemisphere (Corballis, 2009, Hugdahl, 2005). It has been shown that these so-called functional cerebral asymmetries (FCAs) modulate the efficacy of executive functions related to response inhibition processes, with initial stimulus presentation in the non-dominant hemisphere leading to a less efficient inhibition process (Measso and Zaidel, 1990, Ocklenburg et al., 2011). Although likely, it is not clear whether FCA’s also modulate executive functions involved in task switching. Even more important is the question of which neurophysiological processes are involved and in what brain areas this modulation takes place. This information is necessary to achieve a more comprehensive mechanistic explanation of the neural events subserving task switching. To this end, we use a cued task switch paradigm with tachistoscopic presentation of verbal stimuli in the left (LVF) or right visual field (RVF).

Participants had to categorise the words either based on their initial letters, or according to their word type. The simpler initial letter task can be solved by relying on a perceptual analysis of spatial features of the two initial letters, a cognitive process that is mediated by the right hemisphere (Vogel, Bowers, & Vogel, 2003). The more complex name identity task, however, can only be solved by using a verbal information processing strategy that is mediated by the left hemisphere (Corballis, 2009, Hugdahl, 2005). Therefore, we would expect participants to be more accurate and faster on RVF compared to LVF trials when performing the word type task, but more accurate and faster on LVF compared to RVF trials when performing the initial letter task. These hemispheric asymmetries should be reduced on switch compared to repeat trials, since switch trials are more complex than repeat trials as they include additional cognitive processes including prospective reconfiguration processes, active control processes, as well as passive task interference processes (Rushworth, Passingham, & Nobre, 2005). This increasing task complexity has been linked to reduced hemispheric asymmetries (Hausmann, Kirk, & Corballis, 2004) as well as greater interhemispheric interaction in order to solve a task (Banich and Belger, 1990, Bayer et al., 2008, Weissman and Banich, 2000, Welcome and Chiarello, 2008). Moreover, we expect participants to be faster and more accurate on repeat- compared to switch-trials.

On the neurophysiological level, FCAs should be evident in two different cognitive processing stages when examined using event-related potentials (ERPs): On the one hand, FCAs for processing of verbal stimuli have an effect on stimulus-driven attentional processing as reflected by a left-lateralization of the N1 ERP-components. The N1 is a negative ERP component that is supposed to reflect mechanisms that orient attention towards visual stimuli (Beste et al., 2010, Beste et al., 2008, Herrmann and Knight, 2001, Hillyard and Anllo-Vento, 1998, Wascher and Beste, 2010) or mechanisms involved in the categorisation of these stimuli (Grossi, Savill, Thomas, & Thierry, 2010). The N1 is the earliest ERP component that reflects recognition of verbal stimuli (Spironelli & Angrilli, 2009) and the greater efficacy of the left hemisphere in processing of verbal stimuli is reflected by a left-lateralized N1 (Grossi et al., 2010, Proverbio et al., 2002, Spironelli and Angrilli, 2007). Consequently, brain areas that have been linked to the N1 like the extrastriate cortex, dorsal occipito-parietal and ventral occipito-temporal areas (Gomez-Gonzalez et al., 1994, Herrmann and Knight, 2001, Yamazaki et al., 2000) should show FCA-dependent activation differences.

In addition to the early stimulus-driven attentional processes reflected by the N1, FCAs have also been shown to modulate later ERP components (Ocklenburg et al., 2011). In this regard the P3b is important for task switching performance (Gajewski et al., 2010, Hsieh, 2006, Kok, 2001). The P3b has been suggested to reflect a memory guided stimulus evaluation process (Kok, 2001). One of the main theoretical accounts of P3b function is the context-updating or schema revision approach (Donchin, 1981, Polich, 2007). This processing capacity approach suggests that a stimulus entering the processing system elicits a memory comparison process which checks whether the current stimulus is identical to the previous stimulus or not. Should the incoming stimulus be different compared to the trial before, the subject has to allocate additional attentional resources to this stimulus and the neural representation of the stimulus environment is updated. This process leads to a more pronounced P3b potential (Polich, 2007). In accordance with this approach of P3b function, a greater positivity of the P3b following a switch cue compared to a cue that indicates a repeat was observed in cued task-switching experiments (Barcelo et al., 2006, Jost et al., 2008, Nicholson et al., 2006, Nicholson et al., 2005, but see Rushworth et al., 2005). According to the suggestion that the P3b is reduced when a task is more demanding (Polich, 2007), we expect a smaller P3b amplitude after stimulus presentation in the LVF, since the right hemisphere is non-dominant for the processing of verbal information.