Left extrastriate body area is sensitive to the meaning of symbolic gesture: evidence from fMRI repetition suppression

Functional magnetic resonance imaging (fMRI) adaptation (a.k.a. repetition suppression) paradigm was used to test if semantic information contained in object-related (transitive) pantomimes and communicative (intransitive) gestures is represented differently in the occipito-temporal cortex. Participants watched 2.75 s back-to-back videos where the meaning of gesture was either repeated or changed. The just observed (typically second) gesture was then imitated. To maintain participants’ attention, some trials contained a single video. fMRI adaptation –signal decreases– for watching both movement categories were observed particularly in the lateral occipital cortex, including the extrastriate body area (EBA). Yet, intransitive (vs. transitive) gesture specific repetition suppression was found mainly in the left rostral EBA and caudal middle temporal gyrus- the rEBA/cMTG complex. Repetition enhancement (signal increase) was revealed in the precuneus. While the whole brain and region-of-interest analyses indicate that the precuneus is involved only in visuospatial action processing for later imitation, the common EBA repetition suppression discloses sensitivity to the meaning of symbolic gesture, namely the “semantic what” of actions. Moreover, the rEBA/cMTG suppression reveals greater selectivity for conventionalized communicative gesture. Thus, fMRI adaptation shows higher-order functions of EBA, its role in the semantic network, and indicates that its functional repertoire is wider than previously thought.


Appendix: Video stimulus lists
Average peak coordinates (in MNI space), and their peak values (Z statistics) in all major functional areas and regions identified in this study are listed.

Supplemental ROI analyses
In the subdivision that belongs to the rostral EBA (rEBA; see Supplemental Figure 1A, upper inset on the right), a 2 (gesture: transitive, intransitive) by 3 (context: first gesture, second same, second different) repeated-measures ANOVA revealed a significant main effect of gesture (F (1,11) = 14.332; p < 0.05; p η 2 = 0.56; alpha = 0.93), such that the signal associated with watching intransitive gestures was significantly weaker than for transitive gestures.
There was also a trend towards a main effect of context (F (2,22) = 2.782; p = 0.084; p η 2 = 0.202), but signal decreases across the two categories for watching the repeated gestures were not substantial enough. The gesture by context interaction was not significant, either (F (2,22) = 1.457; p = 0.254). Yet, both the whole brain (voxelwise) analysis and the inspection of the the upper inset in Fig. 1A clearly indicates that rEBA reveals different adaptation for the two gesture categories. Indeed, as shown by the a priori t-test, the signal decreases for the same back-to-back intransitive and transitive gestures were significantly different (t (11) = -2.751; p < 0.05). In the subdivision that belongs to the caudal MTG (cMTG ROI; see the lower inset of Supplemental Figure 1A), a repeated-measures ANOVA also showed a significant main effect of gesture (F (1,11) = 8.324; p < 0.05; p η 2 = 0.431; alpha = 0.748) and again the signal associated with watching intransitive gestures was weaker than for transitive pantomimes. A main effect of context (F (2,22) = 2.503; p = 0.105) and an interaction were not significant (F (2,22) = 2.108; p = 0.145). As before, such an ANOVA was not particularly informative because a mere inspection of the graph (the lower inset on the right) indicates that the signals associated with watching the first gestures from both categories did not differ. This is indeed the case (t (11) = 1.54; p = 0.150) Even more importantly, and consistent with the whole brain analysis, the a priori t-test also revealed that the decreases of signal for the same back-to-back intransitive gestures were significantly greater (t (11) = -2.473; p < 0.05).
Supplemental Figure 1B, shows all the major subdivisions of the occipito-temporal cortex.
In the left anterodorsal Precuneus ROI (see Supplemental Figure 1C, upper inset on the left), a repeated-measures ANOVA again revealed a main effect of gesture (F (1,11) = 6.789; p < 0.05; p η 2 = 0.382; alpha = 0.661) in which, consistent with the voxelwise analysis, the signal associated with watching transitive gestures was significantly higher. A main effect context was also significant (F (2,22)  = -2.251; p < 0.05). Only the medial surface of the left hemisphere is presented here. The actual cluster of significant difference which was revealed in a direct contrast (i.e., intransitive vs. transitive imitation) in the whole brain analysis is shown in a schematic form. The independent bilateral paracingulate gyrus ROI was obtained from the Harvard-Oxford atlas implemented in the FSL package. The probabilistic map was thresholded at 30% of its lower probability tail.