Figure 4. Maximum grip aperture and Maximum grip aperture variance in the two groups during Precise grasping.

The upper panel (A) illustrates the four-level Session×Action-type×Movement-type×Group significant interaction shown by the general ANOVA on Maximum grip aperture (MaxAp). It indicates that, only in the MG, MaxAp of Precise grasping changed over sessions according to Action-type; indeed, only in this group, MaxAp in Complementary trials increased in Session 2 with respect to Session 1 (p = .006), so that the two Action-types (complementary/imitative), that were identical at the beginning of the experiment (p = .4), diverged in Session 2 (p = .001). These results suggest that in the MG interference effects, due to the observation of an incongruent movement performed by the partner, increased over time. The lower panel (B) illustrates the Session×Interaction-type×Movement-type×Group significant interaction emerged from the general ANOVA on Maximum grip aperture variance (Var_MaxAp). The grip aperture variance in Precise grasping significantly decreased in NG while it significantly increased in MG throughout sessions. These results suggest that while individuals in the NG learned how to coordinate without being influenced by the partner's movement, participants in the MG became more mutually responsive over time. This can be considered an index of the enhancement of reciprocal responsiveness between partners in the MG, in terms of both involuntary mimicry and movement corrections. The fact that these effects were found in Precise grasping only is likely to be due to the more sensitive feature of this movement-type to action-goals. Error bars indicate s.e.m. (*) p<.05, (**) p<.01, (***) p<.001.