Evidence for distinct roles for basal ganglia and SMA in automatic and unconscious inhibition of voluntary actions

Introduction: Although previous research highlighted the importance of automatic and unconscious inhibition in motor control, the neural correlates of such processes remain unclear. Basal ganglia dysfunctions have long been associated with impairment in automatic motor control. In addition, Sumner & al. (2007) suggested a key role of the medial frontal cortex by administrating a masked priming task (e.g., Eimer & Schlaghecken 2003) to a patient with a small lesion restricted to the supplementary motor area (SMA)., Here, we used fMRI in normal subjects to better delineate the respective roles of SMA and basal ganglia in automatic and unconscious motor inhibition.

Methods: We used event-related BOLD fMRI at 3T to record brain activity in 26 healthy volunteers (22 ± 2 years) as they performed the subliminal masked priming task. In this visuomotor task, participants are asked to make speeded button presses with the left or right hand following leftward or rightward pointing arrows, which are preceded by masked prime arrows.
Here, two experimental variables were manipulated: the interval between the mask and the target (SOA: 0,100,150,200 or 250 ms) and the prime/target direction (compatible or incompatible). Imaging data processing and analysis were performed using SPM8b.
Results: using Repeated Measures ANOVA of behavioral data (global interaction SOA*compatibility, p<0.0000001), we replicated the masked priming effects showing faster reaction times (i.e., motor response facilitation) in compatible than incompatible trials at 0-SOA (positive compatibility effect: diff = 21 ms, linear contrast : p<0.0000001) and the reverse (negative compatibility effect) at 100 (diff = -12 ms, p= 0.01) and 150-SOA (diff= -12 ms, p= 0.008) suggesting motor response inhibition. At 200 & 250 SOA, we no longer found significant compatibility effects (p>0.05)

By applying a similar statistical model to imaging data, we observed a stronger activity in the in several regions, the SMA (p<0.001, uncorrected), caudate (p=0.002, uncorrected) and thalamus (p<0.001, uncorrected) showing stronger activity in compatible than incompatible trials at 100 and 150-SOA, as compared with 0-SOA. Moreover, the differential activity in the SMA was correlated with the negative compatibility effect (p= 0.01). When testing for a main effect of SOAs we didn’t find a differential activation of the SMA, but a stronger deactivation of the caudate (p=0.009, uncorrected) and the thalamus (p=0.007, uncorrected) at 100-150 SOA (inhibition conditions) compared to 0-SOA (facilitation condition).

In a prime identification task administered after the fMRI experiment, subjects’ performance was at chance levels for primes displayed for 17 ms as in the main study, suggesting that the prime was not consciously perceived.

Conclusions: These new findings suggest that automatic and unconscious inhibition of an activated motor response is mediated by the basal ganglia whereas medial frontal regions seem to be more implicated in the control of response conflict related to inhibition.