Congrats to Kobe for getting accepted in Elife!
In this study, we put forward the hypothesis that the sense of confidence controls how much evidence we gather before making a new choice. The underlying assumption, which is well supported empirically, is that most decisions are based on the protracted accumulation of noisy “evidence”; the agent commits to a choice when a certain level of evidence supporting that particular choice has been accumulated (Bogacz et al, 2006; Gold & Shadlen, 2007). To test this hypothesis, we used an established algorithmic model of the above accumulation-to-bound process for two-choice tasks, the drift diffusion model (Ratcliff & McKoon, 2008). We fit this model to choice and reaction time data of three different perceptual choice tasks (n = 67 people in total), each of which required participants to first report a choice, followed by subjective confidence ratings, and no trial-by-trial feedback (i.e. participants never knew for sure whether they were correct or wrong on the previous trial). As predicted, the estimated separation between decisions bounds depended on the level of confidence they reported about the decision they had made on the previous trial (see panel B above). We also unraveled a neurophysiological signature of this online adjustment process. We namely found that a centro-parietal EEG signal serves as a sensitive indicator of subjective confidence, and it also seems to be used for the online adjustment of speed-accuracy-tradeoff.
- Desender, K., Boldt, A., Verguts, T., & Donner, T. H. (2019). Confidence predicts speed-accuracy tradeoff for subsequent decisions. ELife, 8, e43499. https://doi.org/10.7554/eLife.43499