Single-Trial Event-Related Potential Correlates of Belief Updating

Progress in understanding the neural bases of cognition requires computational frameworks. Bennett et al. provide an important example for Bayesian belief updating (BBU). They describe BBU as the transformation of prior beliefs into posterior beliefs after new information is observed, and they report that BBU is encoded in anterior P3 (P3a) amplitude variability. Bennett et al. describe their P3a finding as if it was novel, citing two previous publications as formulating the hypothesis that P3 amplitude reflects a BBU mechanism (Kopp, 2008; Mars et al., 2008). Kopp (2008) provides theoretical arguments. Mars et al. (2008) found that predictive surprise was encoded in the parietal P3b, leading these authors to speculate that the P3a may encode BBU, but they left their speculation unexamined. Bennett et al. shortly mention another recently published study: ’The observed results are broadly consistent with recent research investigating Bayesian single-trial properties of the P3 in a prediction task without reinforcement (Kolossa et al., 2015).’ (p. 12). However, they do not provide detailed descriptions of the reported findings: This study evidently showed that P3a amplitude variability can be best accounted for by BBU, that P3b amplitude variability can be best accounted for by predictive surprise (see also Kolossa et al., 2013), and that posterior slow wave amplitude variability can be best accounted for by Bayesian prediction updating. Our previous work was inadequately cited by Bennett et al., leaving the unjustified impression that their “P3a encodes BBU” finding represents a first-time scientific discovery through neglecting our study’s veritable impact.Show Less