Investigating the Neural Basis of Theta Burst Stimulation to Premotor Cortex on Emotional Vocalization Perception: A Combined TMS-fMRI Study

Abstract

Previous studies have established a role for premotor cortex in the processing of auditory emotional vocalizations. Inhibitory continuous theta burst transcranial magnetic stimulation (cTBS) applied to right premotor cortex selectively increases the reaction time to a same-different task, implying a causal role for right ventral premotor cortex (PMv) in the processing of emotional sounds. However, little is known about the functional networks to which PMv contribute across the cortical hemispheres. In light of these data, the present study aimed to investigate how and where in the brain cTBS affects activity during the processing of auditory emotional vocalizations. Using functional neuroimaging, we report that inhibitory cTBS applied to the right premotor cortex (compared to vertex control site) results in three distinct response profiles: following stimulation of PMv, widespread frontoparietal cortices, including a site close to the target site, and parahippocampal gyrus displayed an increase in activity, whereas the reverse response profile was apparent in a set of midline structures and right IFG. A third response profile was seen in left supramarginal gyrus in which activity was greater post-stimulation at both stimulation sites. Finally, whilst previous studies have shown a condition specific behavioral effect following cTBS to premotor cortex, we did not find a condition specific neural change in BOLD response. These data demonstrate a complex relationship between cTBS and activity in widespread neural networks and are discussed in relation to both emotional processing and the neural basis of cTBS.

Keywords: cTBS; emotional vocalization; emotions; fMRI; functional magnetic resonance imaging; premotor cortex; transcranial magnetic stimulation.

FIGURE 1

Responses during passive perception of emotional vocalizations in dorsolateral temporal cortices. Bold responses to emotional sounds compared to spectrally rotated versions were seen in middle and posterior superior temporal gyri in both hemispheres for all categories of emotional sound, largely lying within cortical regions revealed by the comparison of all sounds compared to rest (depicted by solid red lines). In most cases responses were more distributed in the left hemisphere (A–D). Perception of triumph and fear sounds was associated with an additional peak in the left inferior frontal gyrus (B,D). Thresholded at p < 0.001 uncorrected with a cluster extent threshold of 30 voxels.

FIGURE 2

Continuous TBS to premotor cortex increases BOLD responses in frontoparietal cortices compared to vertex stimulation. In order to look at the effect of cTBS to right premotor cortex on perception of sounds, compared to stimulation at the vertex, individual t-tests for the comparison of each condition post-stimulation > pre-stimulation were calculated. These t-tests, which reflect the difference between pre- and post-stimulation were then entered into a 2 × 6 ANOVA, whereby the two factors were ‘Site’ (premotor and vertex) and ‘Condition’ (Amusement, Disgust, Fear Triumph, Rotated, Rest). These t contrasts reflect the difference between pre- and post-stimulation, and such that effects of the Site and Condition factors in fact reflect the interaction of these factors with the effect of Session (pre- vs. post-stimulation). A ‘Site’ × ‘Session’ interaction was observed in inferior frontal gyri, left middle frontal gyrus and right insula cortex and finally in right precentral gyrus. A second set of clusters lay in bilateral inferior parietal cortex and postcentral gyrus on the right, left supplementary motor area extending into superior frontal gyrus. Finally a set of midline clusters were seen in right middle cingulate cortex, precuneus and parahippocampal gyrus. (A) A t-test revealed greater activity for the comparison of PMv post-stimulation compared to PMv pre-stimulation, than vertex post-stimulation compared to vertex pre-stimulation, in left inferior frontal gyrus, left supplementary motor area, cerebellar vermis, right parahippocampal gyrus, bilateral inferior parietal lobe, bilateral superior and right middle frontal gyri and right postcentral gyrus (B). The opposite comparison was associated with activity in left hippocampus, right middle cingulate cortex, right precuneus, left supramarginal gyrus, right supplementary motor area, right inferior frontal gyrus and right rolandic operculum (C).

FIGURE 3

Region of interest analysis reveals three distinct response profiles in response to cTBS to premotor cortex. In order to look at the response profiles of the regions revealed by the ‘Site’ × ‘Session’ interaction, mean parameter estimates were extracted from each cluster and are displayed in (A–K) where positive values indicate greater activity post-stimulation compared to pre-stimulation (i.e., increased activity following cTBS) and negative values indicate a reduction in activity following cTBS. This approach revealed one network of regions that show increased levels of activity following stimulation to right PMv including supplementary motor area, inferior parietal cortex, bilateral inferior frontal gyrus, superior frontal gyrus and post central gyrus lying close to the PMv target site (graphs A–F). In contrast to this, two distinct regions of the precuneus, cingulate cortex and right IFG displayed a reduction in activity following stimulation to premotor cortex (graphs H–K) and less activity following premotor stimulation compared to following vertex stimulation. A third and final response profile was observed in a single cluster in left supramarginal gyrus whereby activity was greater post-stimulation at both sites, compared to pre-stimulation (G).

FIGURE 4

BOLD responses at the site of PMv stimulation, and the left homolog, were not significantly changed following cTBS. No significant whole brain effects were seen at the exact site of simulation or the in the corresponding site in the left hemisphere. In order to look at BOLD responses within this region further, a 6 mm spherical regions of interest were created around the coordinates the target site [54 –2 44] and the left homolog [–54 –2 44]. Mean parameter estimates were extracted and are plotted in the lower panel. The pattern of responses is not consistent across the emotional sound categories and an ANOVA indicated that there were no significant differences between any of the conditions post-stimulation compared to pre-stimulation following premotor or vertex stimulation in either site (here, a value of zero indicates no change in activity levels between pre- and post-stimulation, whereas positive values indicate greater post-stimulation activity than pre-stimulation).