Response priming in a go/nogo task: do we have to explain the go/nogo N2 effect in terms of response activation instead of inhibition?

Abstract

Objectives: In the present study, we examined the effects of response priming on the event-related potentials (ERPs) evoked by target stimuli in a go/nogo task.

Methods: In each trial, subjects were presented a cue and a target stimulus. The cue informed subjects about the following target in that trial, and therefore, also about the kind of response (right-hand response, left-hand response, no overt response) potentially to be given in that trial.

Results: The traditional N2 and P3 go/nogo effects were replicated: the ERPs to nogo targets were negative compared to the ERPs evoked by go targets in the N2 latency range at frontal electrode sites, and the nogo P3s were more anteriorly distributed than the go P3s. Comparing the ERPs evoked by nogo targets, we found the P3, but not the N2, to be modulated by response priming.

Conclusions: These results seem to indicate that the P3, but not the N2, is associated with response inhibition, or with an evaluation/decision process with regard to the expected and/or given response. It could be speculated that the traditional go/nogo N2 effect has to be explained in terms of response activation instead of response inhibition.

Introduction

In reports of studies employing a visual go/nogo task usually two main differences in the event-related potentials (ERPs) between go (subject has to respond) and nogo (subject has to refrain from responding) trials are found. The first difference concerns a negative deflection in the ERPs evoked in nogo trials compared to go trials around 200–400 ms after stimulus onset. This N2 component is maximal at the frontocentral electrodes. The second difference concerns the more anterior topographical distribution of the P3 evoked by nogo stimuli compared to go stimuli, which usually evoke a P3 with a parietal maximum. Both effects have been interpreted as reflections (or an outcome) of inhibitory processes (Pfefferbaum et al., 1985, Kok, 1986, Pfefferbaum and Ford, 1988, Jodo and Kayama, 1992, Eimer, 1993, Kopp et al., 1996, Thorpe et al., 1996, Falkenstein et al., 1999, Filipović et al., 1999, Van ‘t Ent and Apkarian, 1999).

It has to be noted that the interpretation of the above-mentioned ERP effects in terms of inhibition processes is not undisputed. One confounding variable in the comparison between go and nogo trials concerns the movement-related brain activity. The P3 go/nogo effect could be attributed to an overlapping negativity due to the accompanying motor response required in go trials, which is withheld in nogo trials (Simson et al., 1977, Kok, 1986, Kopp et al., 1996). However, Pfefferbaum et al. (1985) and Bruin and Wijers (unpublished data) showed a similar pattern of results for the P3 go/nogo effect in a condition where subjects had to count covertly the go stimuli as that obtained in a condition where subjects had to respond manually to the go stimuli. This latter result indicates that the P3 go/nogo effect can not be solely explained by the overlay of confounding motor potentials in go trials.

Furthermore, Falkenstein et al. (1995) interpreted the absence of an N2 go/nogo effect with auditory stimuli in their go/nogo task study as evidence against the response inhibition hypothesis (see, however, also Falkenstein et al., 1999, for a more cautious interpretation of this finding). However, it is conceivable that there are differences in location and/or orientation of the corresponding source in the case of auditory nogo stimuli compared to visual nogo stimuli. Since in their study only a very limited electrode set-up (with only one frontal electrode, namely Fz) was used, it is very possible that the activation of a response inhibition mechanism in the N2 range to auditory stimuli simply has not been picked up. This latter argument is supported by, for example, an auditory study of Kiefer et al. (1998) in which there was a go/nogo effect in the time window of the N2 ERP component at electrodes placed at inferior frontotemporal brain region sites. Results of measurements in monkeys with depth electrodes also corroborate the hypothesis of different loci of the source of a nogo inhibition potential as a function of different sensory modalities (Gemba and Sasaki, 1990).

In general, it is assumed that the prefrontal cortex is involved in inhibition processes. However, there are some discrepancies in the exact location of this source in the prefrontal cortex between several studies. In the above-mentioned study of Gemba and Sasaki (1990) it was found that the principal sulcus in the prefrontal cortex (bilaterally) is activated during the occurrence of a potential specific to the nogo reaction. In a MEG study with humans, Sasaki et al. (1993) found a bilateral in- and outflow of magnetic fields also over the dorsolateral frontal parts of the head. Dipole analyses of these data suggested the presence of sources in the dorsolateral parts of the frontal lobes in both hemispheres, presumably in the prefrontal-premotor areas.

Lesion studies also suggest the importance of the prefrontal cortex in inhibition. In particular, the orbital frontal cortex has been named in these studies (e.g. Iversen and Mishkin, 1970, Malloy et al., 1993, Fuster, 1997). Recent neuroimaging studies provide further evidence for the role of the frontal lobe in inhibition processes. Konishi et al. (1999) found in their functional MRI study that nogo stimuli compared to go stimuli activated the posterior part of (mainly) the right inferior frontal sulcus. In a fMRI study of Garavan et al. (1999) inhibition-related activity was also predominantly right-lateralized, most pronounced in middle and inferior frontal gyri, the inferior parietal lobule, and the angular gyrus. Casey et al. (1997) found in nogo trials compared to go trials bilateral activation in predominantly inferior frontal, middle frontal, orbital frontal, and anterior cingulate gyri.

The present study was designed to investigate further response inhibition processes by priming differentially go/nogo responses. Our subjects were presented one of 3 different kinds of targets in each trial: a nogo target, a go target to be responded to with the left hand, or a go target to be responded to with the right hand. In each trial, a go/nogo target was preceded by one of 4 possible cues. The cue gave the subjects information about the kind of target that could be presented in that trial. One of the cues was always associated with the nogo target (specific nogo priming). Another cue could be followed by all 3 kinds of targets (nonspecific go priming). A third cue was always followed by a nogo target or a left-hand go target, while the fourth cue was associated with the nogo target and the right-hand go target (specific go priming).

The first aim of the present study was focused on the replication of traditionally found differences between ERPs evoked by go targets and ERPs evoked by nogo targets. We expected to replicate the common N2 and P3 go/nogo effects.

Our second question concerned the effect of priming go/nogo responses differentially on the ERPs evoked by the nogo target. We expected that subjects would prepare their response on the basis of the preliminary information as given by the cue in order to respond as fast as possible to the target if necessary. This could mean that the required response inhibition in the case of a nogo target would increase as a function of the strength/specificity of the priming of the go response. That is, we expected the strongest response inhibition after presentation of a nogo target in specific go priming trials and the weakest in specific nogo priming trials. On the other hand, it can also be conceived that inhibition occurs at different levels of information processing as a function of the strength/specificity of response priming. It can be hypothesized that inhibition occurs at a stadium further on in the information processing stream if there is beforehand more information about the response to be given. In other words, it could be that in the case of specific go priming inhibition occurs at the level of, for example, the motor cortex, while in the case of nonspecific go priming inhibition takes place at an earlier level, for example the pre-motor cortex or the supplementary motor area.

Kopp et al. (1996) have also investigated effects of response priming on inhibition processes. In their go/nogo task study subjects were presented flanker stimuli (primes) along with the targets. The relevant comparison concerned the nogo trials. In these trials nogo targets could be presented along with neutral (in their words: nonspecific) primes, or with specific response (that is, compatible with one of the two go responses) primes. It was found that the N2 but not the P3 amplitude on nogo targets was affected by response priming: the N2 was larger in the specific priming condition than in the nonspecific priming condition. Their conclusion was that the traditionally found N2 go/nogo effect, but not the P3 go/nogo effect, is associated with response inhibition. We wanted to investigate if we could find more evidence for this latter conclusion in our task set-up.

Our third question concerned the start of selective response activation to the targets. We expected faster selective response activation after go targets in specific go priming trials than after go targets in nonspecific go priming trials, because of the more specific information given by the cue about the response hand possibly needed. Furthermore, we wanted to know if we could also find evidence for selective response activation after nogo targets in specific go priming trials. In order to answer these questions we applied the lateralized readiness potential (LRP) technique (e.g. Kutas and Donchin, 1980, Smid, 1993). The LRP may be regarded as an index of hand-specific motor activation.

A fourth goal was 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 the case of nogo stimuli. Therefore, ERPs were registered from 61 electrodes covering the whole head in order to construct precise topographical maps.