Skip to main content
. Author manuscript; available in PMC: 2011 Jun 11.
Published in final edited form as: Neuroimage. 2010 Apr 18;56(2):601–615. doi: 10.1016/j.neuroimage.2010.04.036

Fig. 9. Conventional vs. model-based decoding performance (dataset 2).

Fig. 9

The diagram contrasts conventional decoding (blue) with model-based decoding (green) in terms of overall classification accuracies obtained on each auditory mismatch dataset. Model-based accuracies tend to be lower than conventional accuracies, but they remain significantly above chance in 2 out of 3 cases (59.7% and 54.1%, p < 0.05 each). All results are given in terms of balanced accuracies (see Section 2.2) along with 95% confidence intervals of the generalization performance.