Specific proactive and generic reactive inhibition

Highlights

• Inhibition of on-going response tendencies involves a proactive top-down mechanism.

• This proactive mechanism is logically tractable within a computational model.

• Inhibition of on-going response tendencies also involves a reactive mechanism.

• The reactive mechanism is generically invoked by diverse salient events.

• The pro-active mechanism depends on ‘rareness detectors’ in sensory cortex.

Abstract

Inhibition concerns the capacity to suppress on-going response tendencies. Patient data and results from neuro-imaging and magnetic-stimulation studies point to a proactive mechanism involving top-down control signals that potentiate inhibitory sensory-motor connections, depending on whether possibly necessary inhibition is anticipated or not. The proactive mechanism is manifest in stronger sensory-cortex responses to stop signals yielding successful inhibition, observed as a modulation of short-latency human evoked potentials (N1) which may overlap with generic mechanisms for infrequent-event detection. A second, reactive, mechanism would be much more independent of the specific inhibition context, and generalize to situations in which behavioral interrupt is not dictated by task demands but invoked by the salience of task-irrelevant but potentially distracting events. The reactive mechanism is visible in a longer-latency human event-related potential termed frontal P3 (fP3) which is elicited by (successful) stop stimuli and most likely originates from dorsal-medial prefrontal cortex (preSMA), and is dissociated from the proactive mechanism pharmacologically and by individual differences. Implications may arise for more personalized treatments of disorders such as ADHD.

Introduction

Deficits of inhibitory control have been described as key to various pathologies such as schizophrenia, substance abuse, obsessive-compulsive disorder, and especially attention-deficit hyperactivity disorder (ADHD) (Oosterlaan et al., 1998, Lijffijt et al., 2005, Chambers et al., 2009, Bari and Robbins, 2013). It has been argued that inhibitory control can be exerted either proactively or reactively (Cai et al., 2011, Majid et al., 2013). Proactivity refers to a top-down control signal from one brain region (e.g., association cortex) conveying inhibition in another (e.g., motor cortex). Reactivity refers to a more bottom-up activation of neuronal circuits that in turn send an inhibitory signal to another neuronal ensemble (e.g., the motor system). More specifically, proactivity has been associated with selective inhibition, i.e., the suppression of certain response tendencies whereas others are left untouched (Cai et al., 2011, Majid et al., 2013). One major tenet of this review is that such proactivity is also instrumental for global, “all-or-none” inhibition.

A popular neurocognitive task to capture inhibitory control is the so-called stop-signal task (SST). Especially with respect to disorders such as ADHD, deficits in stopping performance probably constitute the best documented neurocognitive impairment. In addition, the SST has high face validity to the extent that it by its very nature captures the capacity to inhibit response-selection and -execution processes that have already been initiated. In the present review it is argued that enriching the SST with brain-activity measures with high temporal resolution enables a direct view on, and distinction between, proactive and reactive mechanisms of stopping. Furthermore, it is proposed that the reactive mechanism transcends the SST and manifests in experimental paradigms that have been devised from perspectives rather different from that of inhibition.

Briefly, the traditional stop task entails participants engaging in a choice-reaction-time task, in which the choice-RT stimuli are occasionally followed by a stop signal. This stop signal is presented within a few hundred milliseconds after the go-RT stimulus and conveys the instruction to withhold the impending behavioral reaction to the go stimulus (Logan and Cowan, 1984). It has been argued that this traditional SST is relatively simple in its task demands and therefore of limited value to in-depth understanding of cognitive and affective processes and especially their pathological aberrations (Aron, 2011). Specifically, the SST would tax downright ‘motor inhibition’ but not control of affectively or motivationally driven impulses, or inhibitory-control functions that are selective for certain response tendencies but leave others untouched. Moreover, the traditional SST would not allow for explicitly addressing the proactive ‘top-down’ component of inhibitory control. Zandbelt and Vink (2010) proposed that manipulation of the probability of stop-signal occurrence be used to track proactive mechanisms. However, as also reported by others (Lansbergen et al., 2007), such manipulations affect go RT rather than stopping performance proper (i.e., the stop-signal reaction time or SSRT, explained further below). As will be argued in this paper, a proactive mechanism specific to the stopping process is directly related to variation in SSRT.

It is further argued that the traditional stop task version does in fact contain a proactive component. Furthermore, downright motor inhibition may in fact be central to pathology, if not for understanding its nature, then at least for clues to treatment. With respect to understanding pathology as such, we have already noted the robust and very well documented motor-inhibition deficits in widely prevalent disorders such as ADHD. With respect to treatment options, consider the case of addiction to, for example, nicotine. Struggling to quit smoking can be easily seen as containing a substantial motor-inhibition component. Given the appropriate cues and the craving for nicotine they elicit, all efforts are needed to suppress the behavioral tendencies that are instrumental in acquiring the nicotine, such as fetching a cigarette from whatever source. It is also conceivable that long-term successful inhibition of these instrumental motor acts eventually helps in weakening the impact of the cues and the ensuing craving, and in turn the impulses to yield to the instrumental motor act, as in extinction of an instrumentally conditioned response. Thus, although motor inhibition may not be the core problem, addressing it in the context of treatment may eventually ameliorate the core problem.

The present review is about non-selective or global stopping. It will be argued that proactivity can be an integral feature to global stopping and is not specific for selective stopping. This necessary dissociation between proactivity and selectivity was also emphasized by Schall and Godlove (2012, p. 1016). These authors identify as one issue in need of clarification, that proactivity versus reactivity is often confounded with selective (associated with proactivity) versus global (associated with reactivity) stopping. Of note, it has recently been demonstrated that inferences from selective-stopping paradigms are quite complex because of issues with higher order moments of the RT-go and failed-stop distributions (Xu et al., 2015).

In sum, motor inhibition may still be a useful tool in understanding pathology as well as inspiring new treatments. In addition, the traditional SST may still be highly valuable to provide insights in both ‘higher’ (e.g., proactive top-down) and lower-level (reactive, more bottom-up driven) mechanisms of inhibition. The remainder of this article discusses these principles and applications in more detail.

It is first specified how a proactive component in the traditional SST can in fact be addressed. Specifically, it would be reflected in the modulation of an obligatory electrocortical response to the stop signal (Section 2.1). This modulation is dependent on top-down signals from either inferior-frontal or superior-frontal cortex (Section 2.2). It is further argued that this modulated response to the stop signal, or the ‘stop-N1’, is followed closely in time by another inhibition-related electrocortical process (‘stop P3’), that is very probably generated in dorsal-medial regions, perhaps the superior frontal gyrus (SFG) or the pre-supplementary motor cortex (preSMA; Section 4.1). If any electrocortical response to the stop signal, the stop P3 may be considered as reactive. Dissociations between stop N1 and stop P3 are noted with respect to the effects of methylphenidate and differences between samples from different populations (Section 4.2). We further argue that the stop P3 reflects a non-specific behavior-interrupt signal that is elicited in a wide variety of circumstances, for example also by distracting but salient stimuli that have no relevance for the current task (Section 5). Finally, a further characterization is proposed of the stop N1, in that it is postulated to be equivalent with so-called transient- or ‘rarity’ detectors in sensory cortex, such as the auditory N1 evoked potentials and the visual rareness-related negativity (Section 6).