Purchase this article with an account.
Dobromir Rahnev, Brian Maniscalco, Hakwan Lau; Direct injection of neural noise leads to double dissociation between accuracy and confidence. Journal of Vision 2012;12(9):617. doi: https://doi.org/10.1167/12.9.617.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
The relationship between accuracy and confidence in psychophysical tasks has traditionally been assumed to be mainly positive, i.e. the two typically increase or decrease together. However, recently we showed that spatial attention can lead to dissociations between visual sensitivity and confidence ratings (Rahnev et al., 2011, Nature Neuroscience). The data were explained by a computational model in which lack of attention increased the variability of the internal distributions. In order to test this model, we directly injected neural noise by applying transcranial magnetic stimulation (TMS) to the occipital cortex. We expected TMS to increase the variability of the internal response and, in a similar fashion as in inattention, lead to decreased accuracy but increased confidence. In our task, subjects discriminated between left- and right-tilted bars and simultaneously received single TMS pulses at SOA of 100 ms. Compared to TMS applied to a control site (vertex), occipital TMS led to significantly lower accuracy (p = .007) but significantly higher confidence (p = .01). This striking double dissociation provides strong support of our variance-based model. To confirm that the increased variability of the signal was indeed the key reason for the dissociation, we performed a formal model comparison analysis that used information theoretic methods. The analysis confirmed that our model was preferred over models in which TMS changed the signal strength or "dual-channel" models that postulate two processing streams. These results imply a central role of neural variability in perceptual decisions (Cohen and Maunsell, 2009) and suggest that confidence ratings may be produced by a highly suboptimal decision mechanism.
Meeting abstract presented at VSS 2012
This PDF is available to Subscribers Only