Combined TMS-fMRI reveals behavior-dependent network effects of right temporoparietal junction neurostimulation in an attentional belief updating task
Combined TMS-fMRI reveals behavior-dependent network effects of right temporoparietal junction neurostimulation in an attentional belief updating task
Updating beliefs after unexpected events is fundamental for an optimal adaptation to the environment. Previous findings suggested a causal involvement of the right temporoparietal junction (rTPJ) in belief updating in an attention task.
We combined offline continuous theta-burst stimulation (cTBS) over rTPJ with functional magnetic resonance imaging (fMRI) to investigate local and remote stimulation effects within the attention and salience networks. In a sham-controlled, within-subject crossover design, 25 participants performed an attentional cueing task during fMRI with true or false information about cue predictability (see Fig. 1). By estimating learning rates from response times, we characterized participants’ belief updating. Model-derived cue predictability entered the fMRI analysis as a parametric regressor to identify the neural correlates of updating.
Figure 1. Experimental paradigm and illustration of the modeling approach. (A) Participants underwent two experimental sessions preceded by a preparatory session. Active cTBS and sham stimulation were administered before the belief updating task in two separate sessions. The bold font highlights the data analyzed in the present work. (B) Timeline of the experimental paradigm for a validly cued trial. At the beginning of each block, the %CV was shown. The information about %CV could be true or false, but participants were not aware whether the information was true or false in a given block. The initial %CV information was used as prior before the observation of the first trial in the modeling approach. In each trial, participants indicated whether the upper or lower corner of the target was missing. (C) Example of one participant’s trial-by-trial changes in the probability estimate that the cue will be valid (νˆ) as derived from the learning model in a block with a false 90%CV prior after cTBS or sham stimulation. Green dots indicate valid trials (outcome = 1) and invalid trials (outcome = 0) in this experimental block. (D) Observed and predicted patterns of response speed (RS; inverse RT) costs from the learning model. RS costs were calculated by subtracting RSs of invalid trials from valid trials and are shown in relation to the participants’ estimated probability that the cue will be valid in an upcoming trial (νˆ), binned in intervals of 0.2.
rTPJ-cTBS effects showed high interindividual variability. The expected learning rate reduction with false cue predictability information by cTBS was only observed in participants showing higher updating in false than in true blocks after sham. cTBS modulated the neural signatures of belief updating, both in rTPJ and in nodes of the attention and salience networks. The interindividual variability of the behavioral cTBS effect was related to differential activity and rTPJ connectivity of the right anterior insula (see Fig. 2). These results demonstrate a crucial interaction between ventral attention and salience networks for belief updating.
Figure 2. fMRI results of the whole-brain regression and PPI analyses. (A) The activity of the right anterior insula was differentially modulated by cTBS depending on the prior as well as on the participants’ individual updating pattern. Correlation plots show the pattern of BOLD modulation in the right insula (rINS) for the stimulation∗validity interaction, separately for true and false priors. Negative scores indicate a reduction of the BOLD modulation (difference in parametric modulation by νˆ of valid and invalid trials) by the model-derived cue predictability after cTBS whereas positive scores indicate an increase of such a modulation after cTBS. The blue circle indicates the coordinate of the right insula in the ROI analysis. (B) Connectivity between rTPJ and rINS. Correlation plots show the modulation of connectivity (difference in parametric modulation by νˆ of valid and invalid trials) for the stimulation∗validity interaction, separately for true and false priors. Negative scores indicate a reduction of the modulation of connectivity (difference in parametric modulation by νˆ of valid and invalid trials) by the model-derived cue predictability after cTBS, whereas positive scores indicate an increase of such a modulation after cTBS.
Publication:
Mengotti, P., Käsbauer, A.-S., Fink, G. R., & Vossel, S. (2022). Combined TMS-fMRI Reveals Behavior-Dependent Network Effects of Right Temporoparietal Junction Neurostimulation in an Attentional Belief Updating Task. Cerebral Cortex. doi: 10.1093/cercor/bhab511
