The neurophysiological basis of reward effects on backward inhibition processes

Highlights

• Neural mechanisms of reward-related changes in backward inhibition are examined.

• Rewards reduce the magnitude of backward inhibition.

• Rewards affect response and/or conflict monitoring processes in the ACC.

• Perceptual and attentional selection and context updating are unaffected by rewards.

Abstract

The ability to flexibly switch between tasks is an important faculty in daily life. One process that has been suggested to be an important aspect of flexible task switching is the inhibition of a recently performed task. This is called backward inhibition. Several studies suggest that task switching performance can be enhanced by rewards. However, it is less clear in how far backward inhibition mechanisms are also affected by rewards, especially when it comes to the neuronal mechanisms underlying reward-related modulations of backward inhibition. We therefore investigated this using a system neurophysiological approach combining EEG recordings with source localization techniques. We demonstrate that rewards reduce the strength of backward inhibition processes. The neurophysiological data shows that these reward-related effects emerge from response and/or conflict monitoring processes within medial frontal cortical structures. Upstream processes of perceptual gating and attentional selection, as well as downstream processes of context updating and stimulus-response mapping are not modulated by reward, even though they also play a role in backward inhibition effects.

Introduction

In many daily situations we are required to switch between different tasks. Yet, this switching goes along with performance costs, i.e. slower responses and higher error rates. For successful task switching, both an efficient activation of a new task and an inhibition of the no longer relevant previous task are important (Allport et al., 1994, Jamadar et al., 2010, Mayr and Keele, 2000). The process that inhibits the most recently performed task upon switching to a new one is referred to as backward inhibition (BI), and serves the suppression of interferences arising from previous tasks (Allport et al., 1994, Allport and Wylie, 1999, Costa and Friedrich, 2012, Mayr and Keele, 2000). A stronger BI is thought to be related to a better task-switching performance, as it facilitates the activation of a new task set (Mayr and Keele, 2000). However, a strong BI can also be disadvantageous, since the inhibition of a currently irrelevant task can persist over time making it difficult to perform a previously inhibited task when it becomes relevant again (Allport et al., 1994, Allport and Wylie, 1999). To examine these processes, experimental paradigms assessing the time costs of overcoming the inhibition of a recently abandoned task set which becomes relevant again have been developed to measure task set inhibition (Mayr and Keele, 2000). Performance costs related to BI are observed in task sequences in which a task A is repeated from n-2 trials (e.g. ABA task triplet/BI condition), compared to when that task A has no n-2 trial sequence history (e.g. CBA task triplet/non-BI condition).

Executive control functions are generally known to be modulated by reward manipulations and it has been shown that rewards improve performance in many kinds of cognitive control tasks (Braem et al., 2012, Libera and Chelazzi, 2006, Veling and Aarts, 2010). In this context, Notebaert and Braem (2015) proposed that different reward components are associated with different kinds of cognitive control behavior: While the hedonic aspect of reward promotes explorative behavior and flexibility, the learning component of reward induces exploitative behavior increasing stability, and the motivational component of reward promotes anticipatory behavior (Notebaert and Braem, 2015). With regard to task switching, some studies showed that reward can reduce the switch costs (Kleinsorge and Rinkenauer, 2012, Savine et al., 2010, Shen and Chun, 2010). As BI is central to the “magnitude” of emerging switch costs, it seems that BI effects should be modulated by rewards as well. Yet, evidence for this assumption is sparse: Jiang and Xu (2014) reported that reward modulates inhibitory processes underlying task switching but they do not provide insights into the underlying neurophysiological mechanisms and functional neuroanatomical structures. In the current study, we combine EEG (event-related potentials, ERPs) with source localization techniques (i.e., sLORETA) to answer the question which neurophysiological processes within the processing cascade from early attentional processes to response selection mechanisms are changed in timing or intensity by reward modulation of the BI effect and what functional neuroanatomical networks are involved. In this context, it needs however to be mentioned that we used a between-subject manipulation with two different group conditions (i.e. performance-based reward for every trial vs. no reward for any trial) whereas the reward schedule of Jiang and Xu, randomly rewarded one third of their trials indicating reward trials with an additional monetary symbol (Jiang and Xu, 2014), With respect to the theory by Notebaert and Braem (2015), we investigated the effects of motivation/reward anticipation while Jiang and Xu investigated the effects of reinforcement history. It is known that reward modulates conflict monitoring and response selection processes (Braem et al., 2012, Krebs et al., 2013). Compared to the non-BI condition (CBA), the reactivation of the recently abandoned task intensifies response selection and conflict monitoring processes and causes a conflict indicating the need for additional allocation of control in the BI condition (ABA). However, the exertion of this control is assumed to carry an inherent subjective cost. As a consequence, the exertion of cognitive control depends on the expected value of control. In other words, the allocation of control is driven by a cost-benefit analysis which is likely to be modulated by rewards (Shenhav et al., 2013). Based thereon, we expect that rewards affect the BI and non-BI condition differently. Inasmuch as the N2 component has been demonstrated to reflect cognitive control and conflict monitoring processes (Botvinick et al., 2004, Deng et al., 2015, Donkers and van Boxtel, 2004, Folstein and Van Petten, 2008, Huster et al., 2013, Larson et al., 2014), we expect that the N2 component shows relevant differences related to the reward modulation of the BI effect. The anterior cingulate cortex (ACC) is highly associated with conflict-related N2 (Folstein and Van Petten, 2008, Yeung and Cohen, 2006). It plays an important role in conflict monitoring and in assessing the need for cognitive control (Botvinick et al., 2004, Cavanagh and Frank, 2014, Folstein and Van Petten, 2008, Holroyd and McClure, 2015, Kerns et al., 2004, Shenhav et al., 2013). Hence, reward-related modulations of the BI effect should be associated with the ACC. Yet, reward is supposed to have effects on attentional processes as well: It has been shown that the N1 and P2 components are modulated by reward, suggesting changes in attentional processes and the distinctive allocation of attentional resources (Chmielewski et al., 2015, Doñamayor et al., 2012, Flores et al., 2015, Stock et al., 2015, Sugimoto and Katayama, 2013, Yu and Zhou, 2006). Based thereon, it is possible that attentional selection processes during backward inhibition are also modulated by rewards. Matching this, the N1, which is known to reflect attentional selection processes (Beste et al., 2010, Gajewski et al., 2013, Herrmann and Knight, 2001, Luck et al., 1990, Wascher and Beste, 2010), has been demonstrated to be larger in the BI condition than in the control condition (Sinai et al., 2007). This suggests that attentional selection processes in the BI condition are intensified to re-activate the recently abandoned task, which makes it possible that these mechanisms are also modulated by rewards. Given that dopaminergic innervation, which also carries reward signals (e.g. Schultz, 1998), is less strong in occipital and parietal areas than in medial frontal (ACC) regions (Nieoullon, 2002), it is however possible that attentional processes show less reward modulation effects than response selection and conflict monitoring processes.

Another process which could be involved in backward inhibition is reflected by the P3 ERP. During task switching, the P3 has been linked to processes of context-updating and stimulus-response re-mapping (e.g. Finke et al., 2012, Gajewski and Falkenstein, 2011, Polich, 2007). Given that the BI effect arises in the situation, where a recently abandoned task becomes relevant again, it is possible that the BI effect relates to difficulties in context-updating and the decision of stimulus-response mapping, which can be reflected in the P3. Some studies found that the P3 components are also modulated by the reward, which has been ascribed to the increased attention to reward predictive cues (Krebs et al., 2013) and the updating of the internal environment (Broyd et al., 2012, Flores et al., 2015, Gruber and Otten, 2010). On these grounds, we expected that reward modulation might be shown in the P3 components.