The interaction between stop signal inhibition and distractor interference in the flanker and Stroop task

Abstract

In the present study, two experiments were conducted to investigate the interaction between the behavioral inhibition, measured by the stop signal task, and distractor interference, measured by the flanker task and the Stroop task. In the first experiment, the stop signal task was combined with a flanker task. Analysis revealed that participants responded faster when the distractors were congruent to the target. Also, the data suggest that it is more difficult to suppress a reaction when the distractors were incongruent. Whether the incongruent distractor was part of the response set (i.e. the distractor could also be a target) or not, had no influence on stopping reactions. In the second experiment, the stop signal task was combined with a manual version of the Stroop task and the degree of compatibility was varied. Even though in the second experiment of the present study interference control is differently operationalized, similar results as in the first experiment were found, indicating that inhibition of motor responses is influenced by the presentation of distracting information that is not part of the response set.

Introduction

Inhibition has always been a very popular concept in psychology. Nevertheless, the relation between different kinds of inhibition is still poorly understood. In fact, inhibition used to be conceptualized as a unitary concept, but behavioral, neuropsychological and neurophysiological evidence suggests that a differentiation is more appropriate (e.g. Harnishfeger, 1995). In Nigg's (2000, p. 228) taxonomy for inhibitory systems in cognitive psychology there is a distinction between effortful and automatic inhibition, and both types are further subdivided into different inhibitory functions. In the present research, we will focus on the interaction between different forms of `effortful inhibition' and more specifically between (1) behavioral inhibition, measured by the stop signal task, and (2) interference control, measured by both a flanker task and a manual version of the Stroop task. Such an interaction could indicate that different forms of inhibition rely on a common mechanism.

The stop signal paradigm (Lappin & Eriksen, 1966; Logan, 1994; Logan & Cowan, 1984) provides a useful measurement of behavioral inhibition. In this task, participants have to execute a speeded choice reaction time (CRT) task. Infrequently (usually on 25% of the trials), a stop signal is presented. The stop signal tells the participants to suppress their response. On short stop signal delays (SSD; the interval between the presentation of the go signal and the stop signal), participants can easily suppress their response. By contrast, when the stop signal delay is long enough, participants will nearly always execute the response. Logan and Cowan (1984) and Logan, Cowan, and Davis (1984) explained those results by a race between two stochastically independent processes: a go process and a stop process. According to their horse-race model, if the stop process is completed before the go process, participants will inhibit their response (signal-inhibit trials). When the go-process on the contrary finishes before the stop-process, participants will respond (signal-respond trials). Based on the assumptions of the horse-race model, it is possible to estimate the covert latency of stopping: the stop signal reaction time (SSRT; for reviews see Logan, 1994, and Band, van der Molen, & Logan, 2003). The stop signal paradigm has already been used with various responses such as manual responses (see Logan, 1994, for a review), foot movements (De Jong, Coles, & Logan, 1995) and eye movements (Logan & Irwin, 2000). Several authors found age-related differences (e.g. Kramer, Humphrey, Larish, Logan, & Strayer, 1994; Ridderinkhof, Band, & Logan, 1999) and differences in clinical populations such as attention deficit hyperactivity disorder (ADHD; e.g. Jennings, van der Molen, Pelham, Brock, & Hoza, 1997; for a review see Nigg, 2001). Also the simulations of Band et al. (2003) have shown that this estimation can be used to discriminate between groups and conditions.

Until now, few have investigated the relations between the stop signal task and other inhibitory functions or tasks. Only three interactions have been investigated. First, Jennings and colleagues (Jennings, van der Molen, Brock, & Somsen, 1992) found that successful inhibition of motor responses slowed heart rate. The authors concluded that stop signal inhibition and cardiac inhibition may appeal on the same midbrain structures. Second, in several studies no difference between the inhibition of spatially compatible responses vs. spatially incompatible responses has been found (e.g. Logan, 1981). It has been argued that resolving interference of spatially incompatible responses, and stopping of behavior do not interact (Kornblum, Hasbroucq, & Osman, 1990). Logan and Irwin (2000) replicated the null effect for inhibiting manual responses to spatial compatible and incompatible stimuli. However, they did find an interaction between the speed of inhibiting saccadic responses and spatial S–R compatibility, suggesting a that eye and hand movements are inhibited by separate processes. Third, two studies found an interaction between the Eriksen flanker task (Eriksen & Eriksen, 1974; Eriksen & Schultz, 1979) and the stop signal task using different visual stimuli in the flanker task (Kramer et al., 1994; Ridderinkhof et al., 1999).

In the Eriksen flanker task, participants perform a speeded CRT task to target stimuli (usually letters) which are flanked by distractors. The distractors can be congruent (indicating the same response as the target), neutral (no response assignment) or incongruent (the target requires another response than the distractors). The common finding is that CRTs are larger when the flankers are incongruent. When flankers are neutral, CRTs may be larger than when flankers are congruent; this is usually interpreted as a facilitation effect. Several models have been proposed to explain such findings (e.g. Eriksen & Schultz, 1979; Miller, 1988; Ridderinkhof, 1997; Ridderinkhof, van der Molen, & Bashore, 1995). The general idea behind these models is that both flankers and target are processed. For example, Ridderinkhof (Ridderinkhof, 1997; Ridderinkhof et al., 1995) proposed that on the one hand the targets are processed via an attentive processing route with a target selection and stimulus–response translation. The flankers on the other hand are processed via a direct priming route: the congruent flankers activate the correct response and incongruent flankers prime an incorrect response. On condition that there is a close temporal overlap between the activation caused by the incongruent flankers and the response evoked by the target, responses are slowed down. Flowers (1990) manipulated the stimulus onset asynchrony (SOA) between flankers and target. With a simultaneous presentation he hardly found a facilitation effect of congruent flankers, whereas incongruent flankers interfered with the targets. However, when flankers are presented 200 ms earlier, targets with congruent flankers are processed faster and the effect of incongruent flankers disappeared. Interestingly, results of Kramer et al. (1994) and Ridderinkhof et al. (1999) indicate that the SSRT is affected in the same way as the CRT: stopping is slowed down when flankers are incongruent. In spite of the similarity of the results in both studies, the authors interpreted their results slightly differently. Kramer et al. (1994) stressed the number of responses activated and suggested that it is harder to suppress two responses, in case of incongruent flankers, than one response in case of congruent or neutral flankers. Ridderinkhof et al. (1999) interpreted their data more in terms of an interaction between inhibiting an incorrect response in the flanker task and the suppression of a motor response in the stop signal task. A similar interpretation for the data of Kramer et al. (1994) was provided by Logan (1994), stressing the interaction between the two types of inhibition. Also, Logan, Kantowitz, and Riegler (1986) (cited in Logan, 1994) found that the number of responses in the go-task had not much influence on the SSRTs of simple stopping, indicating that the interaction hypothesis of Ridderinkhof et al. (1999) may be more appropriate.

In the present study, it was our purpose to investigate whether the activation of incorrect responses is crucial for the findings in the latter studies by presenting distracting information that did not evoke responses. After all, notwithstanding differences in the interpretations of Kramer et al. (1994) and Ridderinkhof et al. (1999), both groups of authors stressed the fact that the observed interactions occur at the level of response sets.