Visual salience of the stop signal affects the neuronal dynamics of controlled inhibition

Abstract

The voluntary control of movement is often tested by using the countermanding, or stop-signal task that sporadically requires the suppression of a movement in response to an incoming stop-signal. Neurophysiological recordings in monkeys engaged in the countermanding task have shown that dorsal premotor cortex (PMd) is implicated in movement control. An open question is whether and how the perceptual demands inherent the stop-signal affects inhibitory performance and their underlying neuronal correlates. To this aim we recorded multi-unit activity (MUA) from the PMd of two male monkeys performing a countermanding task in which the salience of the stop-signals was modulated. Consistently to what has been observed in humans, we found that less salient stimuli worsened the inhibitory performance. At the neuronal level, these behavioral results were subtended by the following modulations: when the stop-signal was not noticeable compared to the salient condition the preparatory neuronal activity in PMd started to be affected later and with a less sharp dynamic. This neuronal pattern is probably the consequence of a less efficient inhibitory command useful to interrupt the neural dynamic that supports movement generation in PMd.

Figure 1

Countermanding task (multi-stop-signal version) description and behavioural results. (a) Every trial started with the simultaneous appearance of the central target (large red circle) and of the cue signal (small red circle). Monkeys touched the central target (Hold time, variable duration). After, the peripheral target appeared and the Delay period started. Then the change in colour of the cue signal was used as Go, instructing for a reaching movement towards the peripheral target. In no-signal trials the monkeys were rewarded upon the touch of the peripheral target. In stop-signal trials the monkeys had to refrain from moving to get the reward (signal-inhibit trials); otherwise, if a movement was made, the reward was not delivered (signal-respond trials). One out of three different stop-signals (a further change in colour of the cue) could unpredictably and equally probable appears (Go to Stop transition: easy, medium, hard). The white halo around either the central or the peripheral target was used as feedback of touch for the monkey. (b) Schematic of the race model to illustrate the two processes racing toward the threshold in stop-signal trials. The go process is shown as mean (green line) and the possible range (corresponding to the full distribution of RT). The stop processes are indicated separately for the three conditions assuming a slope effect. (c) Effects of the stop-signal salience modulation on inhibitory performance: Left column: Inhibition functions obtained for the two monkeys (Monkey P: circles; Monkey PIC: squares) in the fixed SSDs session; Right column: SSRTs values for each monkey and condition (see Behavioral Results and Table 1 for details). RT, reaction time. SSRT, stop signal reaction time. SSD, stop signal delay.

Figure 2

Neuronal modulation in signal-inhibit trials. (a) For the easy condition, activity comparison between signal-inhibit trials and latency-matched no-signal trials (single movement direction; population data). Modulation is evident before the end of the SSRT. Data are relative to a single recording session (Monkey P, n = 28 electrodes; monkey PIC n = 6 electrodes). (b) Latencies of the divergences between signal-inhibit and no-signal trials (mean ± SE) relative to the finish time of the stop process (SSRT) (monkey P, n = 58; monkey PIC n = 12).

Figure 3

Effect of salience conditions on MUA modulation in signal-inhibit trials. Neuronal activity is aligned to the Stop signal and it is represented for a single channel (a) and at the population level for a single session (b) separately for each monkey (monkey P, for the population n = 28; monkey PIC, for the population n = 6). (c) Average ( ± SE) onset times and slopes of the neuronal modulations across all sessions and monkeys.

Figure 4

Comparison between no-signal and signal-respond trials. (a) Population average for all conditions (latency-matched no-signal trials; hard/medium/easy Stop signal for the signal-respond trials) and sessions (Monkey P, tracking SSD; Monkey PIC, fixed SSD). The grey areas show the ‘before detach’ epoch of analysis. (b) Average activities (mean ± SE) in the ‘before detach’ epochs for data in a.