Does observation of postural imbalance induce a postural reaction?

Abstract

Background: Several studies bring evidence that action observation elicits contagious responses during social interactions. However automatic imitative tendencies are generally inhibited and it remains unclear in which conditions mere action observation triggers motor behaviours. In this study, we addressed the question of contagious postural responses when observing human imbalance.

Methodology/principal findings: We recorded participants' body sway while they observed a fixation cross (control condition), an upright point-light display of a gymnast balancing on a rope, and the same point-light display presented upside down. Our results showed that, when the upright stimulus was displayed prior to the inverted one, centre of pressure area and antero-posterior path length were significantly greater in the upright condition compared to the control and upside down conditions.

Conclusions/significance: These results demonstrate a contagious postural reaction suggesting a partial inefficiency of inhibitory processes. Further, kinematic information was sufficient to trigger this reaction. The difference recorded between the upright and upside down conditions indicates that the contagion effect was dependent on the integration of gravity constraints by body kinematics. Interestingly, the postural response was sensitive to habituation, and seemed to disappear when the observer was previously shown an inverted display. The motor contagion recorded here is consistent with previous work showing vegetative output during observation of an effortful movement and could indicate that lower level control facilitates contagion effects.

Figure 1. The three stimulus conditions.

The Control stimulus was a little white cross presented against a black background. The Upright stimulus was a point-light display of a gymnast, oriented upright, trying to keep his balance on a rope (which was not represented). The biological humanoid was constituted of 23 azure dots (3 for the head, 2 for the pelvis, 1 for each hand, 2 for each foot and 12 for the main joints) presented against a black background. The Inverted stimulus was the same as the Upright one, but rotated 180° around the barycentre of the dots. All stimuli lasted 11s. All displays were shown in reverse contrast and the underlying skeleton, indicated here by the dashed lines, was not represented.

Figure 2. Representative excursions and mean lengths of CoP displacement in the two axes.

Panels A and B show CoP excursions of a representative subject from group Up-first in the three stimulus conditions in the medio-lateral (M-L) and antero-posterior (A-P) directions. Panels A' and B' represent mean lengths of CoP displacement in both axes for group Up-first. Panels C, D, C', D' represent the corresponding data for group Inv-first. Bars indicate standard deviations; * and *** mean significant differences between conditions (respectively p<0.05 and p<0.001; ANOVA for repeated measures). Con.  =  Control condition, Up.  =  Upright condition, Inv.  =  Inverted condition. As can be seen for group Up-first, the length of A-P CoP excursions was greater in the Upright condition compared to the two other conditions.

Figure 3. Area of CoP excursions in the three stimulus conditions.

Panel A shows statokinesigrams of the CoP for a typical subject from group Up-first. Panel B shows the mean area of CoP displacement (area of the confidence ellipse including 90% of the distribution of the CoP) in the three stimulus conditions for group Up-first. Panels C and D illustrate the corresponding data for group Inv-first. Bars indicate standard deviations, * and *** mean significant differences between conditions (respectively p<0.05 and p<0.001; ANOVA for repeated measures). As can be seen for group Up-first, the mean area of CoP displacement was greater in the Upright condition compared to the two other conditions.

Figure 4. Mean lengths and area of CoP excursions: comparison between groups Up-first and Inv-first.

Panels A, B and C respectively show mean medio-lateral length, antero-posterior length and area covered by CoP excursions in the Control and Upright conditions for Group Up-first, and Control and Inverted conditions for group Inv-first. Bars indicate standard deviations, ** and *** mean significant differences between conditions (respectively p<0.01 and p<0.001). Con.  =  Control condition, Up.  =  Upright condition, Inv.  =  Inverted condition. As can be seen, A-P CoP displacement and CoP area were significantly higher in the Upright condition compared to the Control condition for Group Up-first. No significant difference was detected between the Inverted and Control conditions for group Inv-first.