Varying task difficulty in the Go/Nogo task: The effects of inhibitory control, arousal, and perceived effort on ERP components

Abstract

Similar to other executive functions, inhibitory control is thought to be a dynamic process that can be influenced by variations in task difficulty. However, little is known about how different task parameters alter inhibitory performance and processing as a task becomes more difficult. The aim of this study was to investigate the influence of varying task difficulty, via manipulation of reaction time deadline (RTD), on measures of inhibitory control, perceived effort, and task-related arousal (indexed by skin conductance level). Sixty adults completed a visual Go/Nogo task (70% Go) after being randomly assigned to one of three task difficulty conditions: High, Medium and Low, with RTDs of 300, 500 or 1000 ms, respectively. Results revealed incremental increases in Go/Nogo errors and greater perceived effort with increasing difficulty. No condition differences were found for arousal, but the amplitude of the Nogo N2 increased and peaked earlier with increasing task difficulty. In contrast, the Nogo P3 effect was reduced in the High condition compared to the Low and Medium conditions. Finally, the amplitude of N1 and P2 showed differential effects, with Nogo N1 increasing with task difficulty, while the Nogo P2 decreased. This study provides valuable baseline behavioural and ERP data for appropriately manipulating difficulty (via RTD) in Go/Nogo tasks — highlighting the potential key role of not only the N2 and P3, but also the N1 and P2 components for task performance.

Introduction

Inhibitory control refers to the ability to successfully suppress thoughts, behaviour and irrelevant stimuli (Aron et al., 2004). Crucial for the proper functioning of many other cognitive capacities (Clark, 1996), inhibitory control is an important, but often unnoticed, feature of everyday life: Its effective execution potentially means the difference between safely crossing a busy road or endangering oneself to oncoming traffic.

Among the most commonly employed paradigms used to investigate inhibitory processing is the Go/Nogo task, which requires participants to respond to a frequently presented Go stimulus, while withholding a response to a rare Nogo stimulus. Event-related potentials (ERPs) to Go/Nogo tasks typically contain two inhibition-related components: an augmented N2 for Nogo relative to Go stimuli, primarily at frontal sites (e.g. Falkenstein et al., 1999, Fallgatter and Strik, 1991, Oddy et al., 2005), and a more anterior focus for the Nogo P3, where P3 is larger for Nogo than Go stimuli at frontal and central leads. The Nogo N2 has been suggested to reflect the pre-motor ‘need’ for inhibition (Kok, 1986), but more recent research has instead linked the N2 to response conflict (Donkers and van Boxtel, 2004, Nieuwenhuis et al., 2003). By contrast, the Nogo P3 has primarily been related to motor inhibition in recent years (Smith et al., 2006, Smith et al., 2007, Smith et al., 2008, Smith et al., 2010). But further work has also suggested that it may not be linked to inhibition itself, but more to the evaluation of the inhibitory process (Band and van Boxtel, 1999, Bruin et al., 2001). Notably, both components appear to be modulated by different neurobiological pathways (Beste et al., 2008, Beste et al., 2010) supporting the idea that they reflect different inhibition-related sub-processes.

Like other executive functions, inhibitory control is assumed to be a dynamic process that should be influenced by variations in task difficulty. However, relatively little is known about how different experimental parameters affect the behavioural and neural underpinnings of this ability (Beste et al., 2010, Lindqvist and Thorell, 2009, Thorell et al., 2009). There are a number of key reasons why it is important to study the influence of task difficulty on inhibitory control. Firstly, from a clinical perspective, the nature of inhibition deficits can only be ascertained if the paradigms employed are sufficiently difficult to differentiate performance between clinical subjects and healthy controls (Beste et al., 2010, Lindqvist and Thorell, 2009). Further, variations in task difficulty, in and of themselves, have been linked to differences in neural activation, leading to inconsistencies in the Go/Nogo literature (for a meta-analysis see Simmonds et al., 2008). Baseline ERP data are required to clarify these effects. Finally, the possibility of developing targeted inhibition training paradigms as an adjunct to existing rehabilitation programmes may offer a potentially useful aid for individuals suffering from deficits in inhibitory control (for e.g. attention-deficit/hyperactivity disorder, ADHD; Johnstone et al., 2010, Thorell et al., 2009). Training outcomes in these studies may be enhanced if the approach taken is based on fundamental research into the optimal way to manipulate inhibition difficulty. Thus, studying how task difficulty influences inhibitory control is important from both a ‘pure science’ and applied perspective, and is the major aim of this study.

Previous research examining the influence of task difficulty on inhibition-related ERP components has been varied with respect to methodologies and findings. Jodo and Kayama (1992) manipulated task difficulty with reaction time deadline, asking one group of participants to respond within 300 ms of the Go signal, and another to respond within 500 ms. They reported an enhancement of the Nogo N2 only in the fast responders. Although this effect was interpreted as being due to increased inhibition difficulty, this was unable to be confirmed since no behavioural results for inhibitory performance were reported. In a subsequent investigation, Band et al. (2003) divided participants into one of two instructional conditions: a speed condition, where subjects were required to respond as fast as possible, and a balance condition, where speed as well as accuracy was emphasised. The speed of response was found to modulate both inhibitory performance and ERPs, with increased Nogo errors and Nogo N2 for the speed condition. In contrast to these reports, Smith et al. (2006), who separated participants into ‘fast’ and ‘slow’ responders via median split post-hoc, reported no differences for the N2.

Furthermore, despite clear effects being reported for the N2, the Go/Nogo literature examining the influence of task difficulty on the P3 is limited. Previous investigations have either not considered the P3 (Band et al., 2003, Jodo and Kayama, 1992), or have used a 50/50 Go/Nogo split (Jodo and Kayama, 1992, Smith et al., 2006) which may not reliably induce prepotent response inhibition, depending on the paradigm (e.g. Braver et al., 2001, Tekok-Kilic et al., 2001). Moreover, these studies have generally only employed two difficulty levels (i.e. low vs. high). Given that both theoretical viewpoints (e.g. Cognitive-energetic model; Sanders, 1983) and experimental findings (Wodka et al., 2009) have suggested performance improvements only during moderate rather than easy/hard difficulty levels, the use of the three task difficulty conditions in the present study allows examination of a range of effects, rather than simply assuming linear changes. Thus, one aim of this study was to extend previous research by clarifying the effect of task difficulty (as manipulated by reaction time deadline: RTD) on not only the N2, but also the P3, using a 70/30 Go/Nogo split and three difficulty conditions (Low, Medium and High).

Although the main focus of this study was the influence of task difficulty on inhibitory processing, the measurement of skin conductance level (SCL) — a well-established measure of central nervous system (CNS) arousal (Barry and Sokolov, 1993) — allows examination of the effect of arousal level on inhibitory performance and processing. A review of the literature suggests that arousal may amplify or improve task performance (for a discussion see VaezMousavi et al., 2007), which may be characterised by an inverted-U relationship, where moderate levels of physiological arousal result in optimal performance, with a deterioration in performance seen during low- or high-arousal levels (Yerkes and Dodson, 1908). Additionally, as initially proposed by Yerkes and Dodson (1908), optimal arousal levels may depend on the difficulty of a given task. In line with the findings of Yerkes and Dodson (1908) are results showing that inhibition performance was optimised only at moderate inter-stimulus intervals (ISIs; Wodka et al., 2009). Further work by Barry et al. (2007) has reported that increased arousal, via caffeine ingestion, resulted in not only increased SCL, but also concurrent improvements in Go/Nogo performance. However, findings from research using similar tasks have been mixed, showing no relationship between arousal and performance (Barry et al., 2005, VaezMousavi et al., 2009, VaezMousavi et al., 2007). The paucity of errors in the previous studies may help to explain these results, and as such, the manipulation of task difficulty would ensure greater errors and help to more thoroughly explore the arousal/performance link.

In sum, this study sought to extend previous research by examining the behavioural and neural effects of varying task difficulty, via RTD, on inhibitory processing. To this end, we used a modified version of the Go/Nogo task that required the inhibition of a prepotent response during three task difficulty conditions: Low (1000 ms), Medium (500 ms) and High (300 ms). As mentioned above, the Nogo N2 and Nogo P3 have been associated with different aspects of response inhibition so the ERP analyses focused on these components. While no specific predictions were made for the early ERP components, given the potential modulatory effects of task difficulty on early stimulus processing (e.g. Miller et al., 2011), any differences found would be explored. Moreover, participants provided perceived effort ratings and we recorded skin conductance to assess the contribution of arousal on performance and processing.