Elsevier

Biological Psychology

Volume 58, Issue 3, December 2001, Pages 229-262
Biological Psychology

A psychophysiological analysis of inhibitory motor control in the stop-signal paradigm

https://doi.org/10.1016/S0301-0511(01)00117-XGet rights and content

Abstract

We examined two potential inhibitory mechanisms for stopping a motor response. Participants performed a standard visual two-choice task in which visual stop signals and no-go signals were presented on a small proportion of the trials. Psychophysiological measures were taken during task performance to examine the time course of response activation and inhibition. The results were consistent with a horse race model previously proposed to account for data obtained using a stop-signal paradigm. The pattern of psychophysiological responses was similar on stop-signal and no-go trials suggesting that the same mechanism may initiate inhibitory control in both situations. We found a distinct frontal brain wave suggesting that inhibitory motor control is instigated from the frontal cortex. The results are best explained in terms of a single, centrally located inhibition mechanism. Results are discussed in terms of current neurophysiological knowledge.

Introduction

Cognitive control mechanisms are needed to coordinate various cognitive processes involved in human task performance. Several models of human information processing accordingly posit cognitive (‘executive’) control. For instance, Shallice (Norman and Shallice, 1986, Shallice, 1994) hypothesized the existence of a Supervisory Attentional System (SAS), which governs the selection or suppression of schemata, which in turn control the elementary processing that takes place in the execution of a task. Likewise, Meyer and colleagues described an Executive-Process Interactive Control (EPIC) architecture of the human information processing system, in which a central cognitive processor controls perceptual and motor processors (Meyer et al., 1995, Meyer and Kieras, 1997).

Research on stopping motor responses provides a reasonably direct examination of cognitive control. Logan and colleagues initiated a program of research on inhibitory control that made extensive use of a stop-signal paradigm (see Logan, 1994, for an overview). The stop-signal paradigm employs a primary task, typically a visual choice reaction time task. While the participants are engaged in this task they are occasionally presented with a signal, usually a tone, shortly after the respond stimulus. This signal instructs them to withhold their response to the primary, choice task. The stop signal can be presented at various delays after the primary respond stimulus. The chance of stopping the response declines as the delay from the primary respond signal increases. Cognitive control is implicated because stopping is an internally generated act of control changing the current course of action to meet a new goal (Logan, 1994).

Logan and colleagues demonstrated that a horse race model fits the data from the stop-signal paradigm (e.g. Logan and Cowan, 1984). In the horse race model two sets of processes race for completion. The first set controls primary choice reaction time (RT) performance, and is thought to include the processing of the respond stimulus, response choice, and the preparation and execution of the appropriate response. It starts at the onset of the respond stimulus. The second set of processes, which starts at the presentation of the stop signal, controls inhibition, and is thought to consist of stop stimulus processing and response inhibition. These two sets of processes will be referred to as the respond process and the stopping process, respectively. The process that is completed first wins the race and determines whether a response occurs. If the respond process wins, a response is produced despite the stop signal, whereas the response is successfully inhibited when the stopping process wins the race. The horse race model also allows an estimate of inhibition time. Across a variety of tasks, inhibition time has been about 200 ms for young adult participants (Logan and Cowan, 1984).

The horse race model provides an excellent description of behavior in a variety of tasks involving response inhibition, but does not provide an extensive description of the respond and stop mechanisms. A global, perhaps unitary, inhibitory mechanism is suggested by the robustness of the estimates of stopping times across tasks which varied in instructional, stimulus, and response variables. Indeed, most results reported in the stopping literature can be described by a single global inhibitory mechanism (for reviews see Logan, 1994, Logan and Cowan, 1984). Empirical studies using both performance and psychophysiological strategies have, however, suggested the possibility of two separate stopping processes. A primary aim of the current experiment is to examine these putative processes and the evidence upon which they are based.

Two early studies suggested the existence of two stopping modes—a quick mode for inhibiting all responses, and a slow mode for inhibiting selectively. Riegler (1986; cited in Logan, 1994) presented two stop signals. In one condition, participants had to stop their response when either signal was presented. In another condition, they had to stop their response to only one signal, but not to the other. Inhibition times were longer in selective as compared to nonselective stopping. Another early study (Logan et al., 1986; cited in Logan, 1994) required participants to perform a two-choice and four-choice reaction in which different stimuli were mapped onto different key presses. Again, there were two conditions. In the stop-all condition, participants had to stop their response when the stop signal was presented, irrespective of which key had to be pressed. In the stop-selective task, however, they had to stop the response only when a key press with the right index finger was required; they responded normally whenever another key had to be pressed. Inhibition times were fast in the stop-all condition and hardly altered by whether two or four choices were required in the primary task. Inhibition times were substantially longer in the stop-selective condition than in the stop-all condition, and the difference between stop-all and stop-selective inhibition times was larger in the four-choice task than in the two-choice task. Logan et al. (1986) reasoned that selective stopping was slower than non-selective stopping because it required discrimination, and the duration of the discrimination process was longer the greater the number of alternatives. However, two inhibitory modes do not necessarily imply two inhibitory mechanisms. These results may also be interpreted by assuming a single mechanism that is differentially engaged in the two inhibitory modes. The required additional perceptual discrimination in the selective as compared to the non-selective inhibition mode might have prolonged processing prior to the engagement of the inhibitory mechanism.

Results obtained from stop-change tasks have also been used to distinguish between two inhibition modes. For instance, Logan and Burkell (1986) required participants to perform on three tasks; dual-task, stop-all, and stop-change. In the stop-change task, participants had to inhibit their response and then execute a different response when the second stimulus occurred after the respond signal. Inhibition times in the stop-change task were longer than in the stop-all task, suggesting two inhibitory modes. However, as for the results of selective inhibition discussed above, these findings may also be interpreted by assuming a single mechanism with different degrees of engagement.

De Jong and colleagues (De Jong et al., 1990, De Jong et al., 1995) related psychophysiological findings that suggested two inhibitory modes to the motor control literature. Bullock and Grossberg (1988) derived from this literature that there are two separate processes involved in the generation of movements. Central processes are concerned with the programming of structural movement parameters, such as their direction or amplitude. More peripherally operating processes were thought to generate a ‘GO’-signal, which scales (multiplies) the output of the central processes in order to produce the outflow of motor commands to the muscles. In this way, onset and speed of movements are controlled. Likewise, De Jong et al., 1990, De Jong et al., 1995 proposed that movements could be inhibited either by preventing the production of motor commands by the central mechanism, or by preventing the outflow of the motor commands by the peripheral mechanism. Because the ‘GO’-signal was thought to operate in a largely nonspecific way, scaling any motor commands from the central mechanism, De Jong et al. (1995) related the peripheral mechanism to the fast, nonselective mode of inhibition. By implication, the central mechanism was associated with the slow and selective mode.

De Jong et al. (1990) introduced a physiological criterion for the identification of the peripheral stopping process. They used measures of brain, muscle, and force activity to assess the time course of response activation and inhibition in a stop-all task. The lateralized readiness potential (LRP; for reviews see Coles, 1989, Eimer, 1998) reflects the response-specific involvement of the left and right motor cortices of the brain, and was employed by De Jong et al. (1990) to indicate the degree of central motor preparation induced by response processing prior to inhibition. They reasoned that the operation of the central inhibitory mechanism should affect central motor preparation, and therefore attenuate the LRP. Peripheral inhibition, on the other hand, was assumed to occur after central motor preparation, and therefore should leave the LRP intact. De Jong et al. (1990) defined successful inhibits as trials with a stop signal, but no muscle and force activity. The maximum LRP for these trials was below the maximum LRP for normal responding on trials without a stop signal, implicating the contribution of a central mechanism. Partial inhibits were trials on which muscle and force activity was present, but the force did not reach a preset criterion. The maximum LRP for these trials was also lower than the maximum for normal response trials, again indicating that the central inhibitory mechanism was involved. However, De Jong et al. (1990) doubted that the central mechanism alone could account for stopping on successful and partial inhibits. They reasoned that the inhibitory effect on the LRP must precede the inhibitory effects on muscle and force activity by at least the time required for the transmission of central commands to the peripheral motor system. Because the observed effects on muscle and force activity seemed to precede the effects predicted by transmission delays, they argued that the central mechanism alone could not account for stopping on partial inhibits. De Jong et al. (1990) also observed that the LRP exceeded a ‘criterion level’ associated with normal responding both on failed and on partial inhibit trials, even though the response was not completed in the latter case. The notion of a criterion level or threshold was based on the observation that LRP amplitude at the instant of responding is constant across conditions and reaction time bins, suggesting that responses are triggered when the criterion is exceeded (Gratton et al., 1988). Importantly, De Jong et al. (1990) found that the LRP frequently reached the criterion level even on trials resulting in completely successful muscle and force inhibition. If the criterion level of the LRP is indeed associated with triggering a response (Gratton et al., 1988), then the central signs of lateralization and triggering were present on numerous trials on which responses did not occur. This seemed to be strong evidence for a peripheral inhibitory mechanism. By implication, a below criterion LRP in combination with response inhibition defined a more central inhibition.

Some aspects of the data reported by De Jong et al. (1990) suggested that the peripheral mechanism might operate as an ‘emergency brake’, capable of intercepting responses that have escaped central inhibition. First, the maximum LRP on successful inhibit trials increased with stop-signal delay, presumably because of an increase in the number of trials with an above-criterion LRP that contributed to the average. The number of trials on which the peripheral mechanism was required to prevent the response seemed to increase with stop-signal delay. Second, the percentage of partial responses also increased with stop-signal delay. De Jong et al. (1990) reported 13.8, 21.0, and 36.3% of partial responses on early, middle, and late stop-signal delays, respectively. It seems that, although the majority of responses were inhibited by the central mechanism, the proposed peripheral mechanism might be increasingly involved as the respond process has proceeded.

Further evidence in favor of two inhibitory mechanisms came from a study by De Jong et al. (1995). They directly compared stop-all, stop-change and stop-selective conditions. The peripheral mechanism was expected to be involved only in the stop-all condition, hence De Jong et al. (1995) expected the LRP to exceed the threshold associated with normal responding. This was indeed found to be the case, replicating their earlier findings (De Jong et al., 1990). Successful inhibition in the stop-change and stop-selective conditions, by contrast, was thought to be mediated by the central mechanism, hence the LRP was expected to remain sub-threshold. The LRP in the stop-change condition indeed remained below the criterion level, and this finding provided support for the two-mechanism hypothesis. Contrary to their expectations, however, the LRP in the stop-selective condition exceeded the threshold associated with normal responding. This finding, and the results of an additional experiment, led them to conclude that the peripheral mechanism was indeed involved in selective stopping.

A second physiological index, known to be influenced by non-cortical centers, also added to the description of the peripheral mechanism. Jennings et al. (1992) observed cardiac slowing on partial and full inhibit trials. Because cardiac slowing can be initiated by midbrain centers, they suggested that it might reflect the actions of the peripheral mechanism. It remains uncertain, however, whether cardiac slowing can be viewed exclusively as reflecting the peripheral mechanism, most notably because central structures may contribute to cardiac slowing (e.g. Skinner, 1991). In the data reported by Jennings et al. (1992), cardiac slowing did not discriminate between full and partial inhibits, but full and partial inhibits are thought to differ in the involvement of the peripheral mechanism (De Jong et al., 1990). Thus, although the cardiac evidence was interpreted in terms of the peripheral mechanism, it did not provide a demonstration of this mechanism.

In summary, two stopping mechanisms have been suggested but not wholly established. A central or cortical mechanism is seen as relatively slow, but selective. A peripheral mechanism is seen as fast, but global. Differences in inhibition times across paradigms, and differences in the amount of cardiac deceleration between successful and unsuccessful inhibition, are compatible with the distinction between the two mechanisms, but also with a single mechanism. The distinction between the two mechanisms is mainly based on differences in a hypothetical triggering mechanism keyed to LRP amplitude. Experimental conditions that are characterized by a LRP below a ‘respond threshold’ are assumed to involve only central inhibition, whereas conditions exhibiting a supra-threshold LRP are thought to involve a more peripheral mechanism in addition to the central mechanism.

Section snippets

The present research

On balance, it seems that the notion of two inhibitory mechanisms provides an attractive account of the data but is far from firmly established. After reviewing the stopping literature, Logan (1994) concluded that “. . . the evidence for central and peripheral inhibitory mechanisms is scant and depends as much on argument than on fact” (p. 206). The timing arguments of De Jong et al. (1990) depend on very small differences and assumptions generalized from different subjects, tasks, and paradigms.

Participants

Ten right-handed participants, five men and five women, with ages ranging from 19 to 28 years (mean 22.2 years) participated in the experiment. They were all healthy, non-smokers, and had normal or corrected-to-normal vision and hearing. They were paid for the completion of the experiment.

Experimental task and procedure

The experiment was carried out in a dimly lit, sound attenuating, electrically shielded chamber. The participants were each seated in a comfortable reclining chair with supports for hands, arms, and legs. The

Task performance

Table 1 presents an overview of RTs and the probabilities of inhibiting, responding, or producing a partial response. Task performance was fast and accurate. The proportion of correct responses on no-signal (go) trials and correct inhibitions on no-go trials were both high. The mean RT on go trials was reasonably fast (351 ms), hence participants did not seem to delay their responses to increase the chance of withholding the response when a stop signal occurred. The focus on fast responding

Further discussion and conclusions

The results of the present experiment replicate and extend the findings reported by De Jong et al. (1990). In both experiments participants were able to withhold responses within 200 ms after stop-signal presentation, while primary task performance remained fast and accurate. In fact, behavioral analyses in both experiments indicated that stop-signal presentation did not affect primary task performance in any way. In both studies the behavioral results fit the horse race model of inhibitory

Acknowledgements

This study was supported by the Netherlands Organization for Scientific Research (NWO; grant 575-63-082B) and by the Netherlands Institute for Advanced Studies in the Humanities and Social Sciences (KNAW). The invaluable assistance of Wery van den Wildenberg in data collection is gratefully acknowledged.

References (59)

  • P. Romaiguère et al.

    Motor cortex involvement during choice reaction time: a transcranial magnetic stimulation study in man

    Brain Research

    (1997)
  • K. Sasaki et al.

    Suppression of visually initiated hand movement by stimulation of the prefrontal cortex in the monkey

    Brain Research

    (1989)
  • K. Sasaki et al.

    No-go activity in the frontal association cortex of human participants

    Neuroscience Research

    (1993)
  • M.M.C. Van den Berg-Lenssen et al.

    Correction of ocular artifacts in EEGs using an autoregressive model to describe the EEG; A pilot study

    Electroencephalography and Clinical Neurophysiology

    (1989)
  • Band, G.P.H., 1997. Preparation, adjustment, and inhibition of responses. Unpublished doctoral dissertation, University...
  • D. Bullock et al.

    Neural dynamics of planned arm movements: emergent invariants and speed-accuracy properties during trajectory formation

    Psychological Review

    (1988)
  • D. Bullock et al.

    Adaptive neural networks for control of movement trajectories invariant under speed and force rescaling

    Human Movement Science

    (1991)
  • R.E. Burke

    Motor units: anatomy, physiology, and functional organization

  • L.G. Carlton et al.

    Reaction time and response dynamics

    Quarterly Journal of Experimental Psychology

    (1987)
  • M.G.H. Coles

    Modern mind-brain reading: psychophysiology, physiology, and cognition

    Psychophysiology

    (1989)
  • L. Deecke

    Bereitschaftspotential as an indicator of movement preparation in supplementary motor area and motor cortex

  • R. De Jong et al.

    In search of the point of no return: the control of response processes

    Journal of Experimental Psychology: Human Perception and Performance

    (1990)
  • R. De Jong et al.

    Strategies and mechanisms in nonselective and selective inhibitory motor control

    Journal of Experimental Psychology: Human Perception and Performance

    (1995)
  • C.J. De Luca

    The use of surface electromyography in biomechanics

    Journal of Applied Biomechanics

    (1997)
  • J.E. Desmedt et al.

    Voluntary motor commands in human ballistic movements

    Annals of Neurology

    (1979)
  • M. Eimer

    The lateralized readiness potential as an on-line measure of central response activation processes

    Behavior Research Methods, Instruments, Computers

    (1998)
  • M. Fabiani et al.

    Definition, identification, and reliability of measurement of the P300 component of the event-related brain potential

  • J.M. Fuster

    The Prefrontal Cortex. Anatomy, Physiology, and Neuropsychology of the Frontal Lobe

    (1997)
  • G. Goldberg

    Supplementary motor area structure and function: review and hypotheses

    Behavioral and Brain Sciences

    (1985)
  • Cited by (231)

    View all citing articles on Scopus
    View full text