Inhibition, response mode, and stimulus probability: a comparative event-related potential study

Abstract

Objectives: In the present study, effects of response mode (finger movement or counting) and stimulus probability on inhibitory processes were studied.

Methods: Electroencephalographic activity was registered in a visual go/nogo paradigm. Subjects either responded manually to go stimuli or counted silently the occurrence of each go stimulus in different conditions. In both response mode conditions, response probability was varied.

Results: For finger movement and counting, similar N2 and P3 go/nogo effects were found. The amplitude of the nogo N2 and nogo P3 varied as a negative function of nogo stimulus probability. The go P3 varied as a negative function of go stimulus probability. In the manual condition, however, the descending flank of the go N2 at anterior electrode sites was more negative in the 0.50go and 0.75go probability trials than in the 0.25go probability trials.

Conclusions: The results of the present study confirm the hypothesis that differences between go and nogo event-related potentials are not dependent on overt movement-related potentials. It could be speculated that the probability effect on the N2 amplitude in go trials in the manual condition has to be explained in terms of a modulation of the strength of motoric preparation processes varying as a positive function of the probability of the go stimulus.

Introduction

Several investigators have reported that event-related potentials (ERPs) elicited in visual go/nogo task studies show a frontocentral negative wave around 200–400 ms after stimulus onset (N2) and a more anteriorly distributed P3 component in nogo trials compared to go trials. In general, both effects have been interpreted as reflections (or an outcome) of inhibitory processes in the frontal cortex (Pfefferbaum et al., 1985, Kok, 1986, Pfefferbaum and Ford, 1988, Jodo and Kayama, 1992, Eimer, 1993, Kopp et al., 1996, Thorpe et al., 1996, Kiefer et al., 1998, Falkenstein et al., 1999, Filipović et al., 1999, Van't Ent and Apkarian, 1999, Tekok-Kilic et al., 2001).

One confounding variable in the comparison between go and nogo ERPs in most of these studies concerns the occurrence of movement-related potentials. These potentials might differentially overlap the go and nogo ERPs (Simson et al., 1977, Kok, 1986, Kopp et al., 1996). However, Pfefferbaum et al. (1985) showed a similar pattern of N2 and P3 go/nogo effects in a task condition where subjects had to count silently the go stimuli as that obtained in a task condition where subjects had to respond manually to the go stimuli.

The involvement of frontal areas in response inhibition is further corroborated by lesion studies (for example, Malloy et al., 1993, Iversen and Mishkin, 1970, Fuster, 1997) and several recent brain-imaging studies (Casey et al., 1997, Krams et al., 1998, Garavan et al., 1999, Konishi et al., 1999, Liddle et al., 2001). In addition, Gemba and Sasaki (1990) found, with depth electrodes in monkeys, bilateral prefrontal cortical activation starting at about 110 ms post-stimulus onset specific on nogo trials in a go/nogo reaction time hand-movement task. In a magnetoencephalographic (MEG) study with humans, Sasaki et al. (1993) also concluded that the frontal lobes are engaged in the suppression of inappropriate responses. Their conclusion was based on the basis of dipole analyses of the magnetic field specifically related to the nogo response around 135 ms. Furthermore, analyses of the P3 with a 3-dimensional source tomography method (LORETA) by Strik et al. (1998) revealed more pronounced activation in frontal areas on nogo trials than on go trials in a continuous performance task (CPT).

In several oddball paradigm studies the effects of stimulus probability on ERP waves have been investigated. One consistent finding in these studies concerns the inverse relationship between the (subjective) probability of a task-relevant event and the amplitude of the P3 wave of the ERP (for example, Courchesne et al., 1975, Duncan-Johnson and Donchin, 1977, Duncan-Johnson and Donchin, 1982, Johnson, 1989, Czigler et al., 1996, Polich et al., 1996, Polich and Bondurant, 1997, Polich and Margala, 1997, Spencer and Polich, 1999). According to the most influential current hypothesis about the psychological significance of the P3, the context updating hypothesis (Donchin, 1981, Donchin and Coles, 1988), the P3 reflects the updating of a proper representation of the environment in memory (see, however, also Verleger, 1988, Verleger, 1998).

In accordance with these oddball studies, data of Low and Miller (1999) do also suggest an enhancement of the P3 amplitude on go trials as an inverse function of response probability. However, in their study the P3 amplitude (at Pz) on nogo trials appeared to be unaffected by response probability. Unfortunately, they did not register ERP activity at anterior electrode sites. On the other hand, Pfefferbaum and Ford (1988) found a larger P3 on go as well as on nogo trials as an inverse function of probability of occurrence of these go and nogo trials, respectively. Decreasing the probability of nogo trials enhanced its central distribution.

Furthermore, Eimer (1993) found a more pronounced N2 amplitude enhancement by nogo stimuli with an a priori probability of 0.25 (experiment I) than by nogo stimuli with an a priori probability of 0.50 (experiment II). In this context, studies investigating effects of deviance are also of interest. One of the ERP components that has primarily been associated with effects of deviance is the fronto-central negativity (N2) starting around 200 ms (see, for example, Courchesne et al., 1975, Kenemans et al., 1989, Kenemans et al., 1992 for deviance effects in the visual modality; Näätänen, 1992 for a review of deviance effects in the auditory modality). This N2 deviance effect seems to occur independent of stimulus relevance. In that way, it could be considered an automatic process. Because the N2 go/nogo effect also occurs in experiments where the a priori probability of go and nogo stimuli was equated (Pfefferbaum et al., 1985, Kok, 1986, Jodo and Kayama, 1992, Eimer, 1993, Schröger, 1993, Thorpe et al., 1996, Falkenstein et al., 1999, Filipović et al., 1999, Van't Ent and Apkarian, 1999, Bruin et al., 2001, Tekok-Kilic et al., 2001), it can be concluded that response deviance in itself is not crucial for the N2 go/nogo effect. The N2 enhancements in, for example, Eimer's experiment I (see also Schröger, 1993), may thus be due to a combined effect of stimulus probability (deviance) and response assignment (inhibition).

The present study was designed to investigate effects of response mode and go/nogo stimulus probability on ERPs evoked in a visual go/nogo task paradigm. A thorough literature search did not yield any replication of Pfefferbaum et al.'s (1985) counting paradigm study in the visual modality. Therefore, because of its importance, our first goal was to replicate the finding of Pfefferbaum et al. (1985) with respect to the occurrence of N2 and P3 go/nogo effects in a count condition. Our subjects were presented a visual go or a nogo stimulus in each trial. In one condition, they had to respond to a go stimulus with either their right or left hand (manual condition), while in another condition they had to count covertly the number of presented go stimuli (count condition).

The second objective of this study was to examine effects of go/nogo stimulus probability on ERPs elicited in the manual and count condition of the go/nogo task. In each response mode condition we used 3 possible go/nogo stimulus probabilities (blocked design). Go stimuli were presented with a probability of 0.25, 0.50, or 0.75 (and vice versa, nogo stimuli were thus presented with a probability of 0.75, 0.50, or 0.25). It can be hypothesized that response preparation is stronger in trials in which the a priori probability of a go stimulus is high than in trials in which this probability is low. Therefore, we expected brain activity involved in inhibition in nogo trials to increase as a positive function of response probability. In other words, we expected in the manual and in the count conditions a decrease of the N2 and P3 go/nogo effect as a function of nogo stimulus probability.

On the other hand, as mentioned above, deviant stimuli evoke, in general, an enlarged N2 and P3 wave relative to frequent stimuli. We expected to see these effects in both go and nogo trials in both response mode conditions.

Lastly, we wanted to investigate (in a qualitative way) topographical differences in brain activity evoked during the processing of the different go and nogo stimuli. Such differences can point to a different involvement (in an absolute and/or relative sense) of brain areas in the processing of these stimuli. For example, activation of brain areas by nogo stimuli, but not by go stimuli, could possibly be ascribed to the involvement of inhibition processes in case of nogo stimuli. Therefore, ERPs were registered from 61 electrodes covering the whole head in order to construct precise topographical maps.