Purchase this article with an account.
Lleymi J. Martinez Arango, Olenka Graham Castaneda, Jason Samaha; Electrophysiological Signatures of Confidence in a Perceptual Decision under Performance-controlled Conditions. Journal of Vision 2020;20(11):1323. doi: https://doi.org/10.1167/jov.20.11.1323.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
Perceptual decisions are thought to arise from the accumulation of sensory evidence towards a decision boundary. Human and animal experiments have found that the rate of evidence accumulation - which reflects the quality of sensory evidence - is informative of accuracy, reaction time, and confidence. However, many experiments have shown that subjective confidence judgments are influenced by factors other than the quality of evidence, and that, accordingly, confidence and performance (accuracy) are readily dissociable. This raises the question of whether neural signatures of evidence accumulation track objective performance or subjective confidence. Observers discriminated the net direction of motion of a random dot kinematogram (RDK) while EEG was recorded. The stimulus was an overlap of two RDKs, one with evidence in favor of a correct decision (“positive evidence”; PE) and the other with a lower amount of counterevidence (“negative evidence”; NE). In separate trials, we increased both the amount of PE and NE by an individually-titrated level that should produce no significant change in discrimination accuracy. According to prior work, higher confidence ratings (made concomitantly with the choice) occur when PE and NE are increased, allowing us to dissociate neural signatures of confidence and accuracy. We observed signals over central parietal electrodes that obey the principles of evidence accumulation. EEG activity beginning around 200 ms post-stimulus increased at a rate that predicts reaction times. Namely, trials with faster responses were associated with a steeper increase in voltage. This activity also tracked confidence ratings, as sharper increases in activity correlated with higher confidence. Interestingly, the same component showed steeper increases in voltage when comparing high to low PE trials, which were associated with higher confidence, but no change in accuracy. This finding suggests signatures of evidence accumulation may underlie the formation of our subjective sense of confidence even when accuracy is controlled.
This PDF is available to Subscribers Only