Risk preferences causally rely on parietal magnitude representations: Evidence from combined TMS-fMRI

Risk preferences have traditionally been considered as stable traits that reflect subjective-valuation processes in prefrontal areas. More recently, however, it has been suggested that risk preferences may also be shaped by how choice problems are perceived and processed in perceptual brain regions. Specifically, the acuity of the parietal approximate number system (ANS), which encodes payoff magnitudes for different choice options, has been shown to correlate with both risk preferences and choice consistency. However, this correlational relationship leaves open the question whether parietal magnitude representations in fact causally underlie choice. Here, we provide direct evidence for such a key causal role of parietal magnitude representations in economic choice, using continuous theta-burst transcranial magnetic stimulation (cTBS), combined with functional MRI (fMRI) and numerical population receptive field (nPRF) modeling. Our stimulation protocol targeted numerosity-tuned regions in the right parietal cortex, identified through nPRF modeling of individual fMRI data (n=35; within-subject design). The stimulation successfully perturbed neural processing, as evidenced by decreased amplitude of numerical magnitude-tuned responses and less accurate multivariate decoding of presented magnitudes from unseen data that were not used for model fitting. In line with a perceptual account of risky choice, the reduction in neural information capacity was also reflected in noisier behavioral responses. Moreover, a computational cognitive model fitted to choice behavior revealed that perturbing the ANS specifically increased the noisiness of small-magnitude representations. This perturbation made small magnitudes to be perceived as larger than they actually are, leading to more risk-seeking behavior. Finally, individual estimates of the cTBS effect on cognitive noise correlated with the corresponding decrease in amplitude of numerical magnitude-tuned fMRI responses, further solidifying the role of the parietal ANS in economic choice. In conclusion, our study demonstrates that the precision of parietal magnitude representations causally influences economic decision-making, with noisier encoding promoting biased risk-taking as formalized in recent perceptual models of risky choice.