Skip to main content
. 2023 Oct 16;18(10):e0292808. doi: 10.1371/journal.pone.0292808

Fig 3. Assessing goodness-of-fit and fitted conditional responses of predictive GLME models.

Fig 3

(A) The absolute difference between the real and fitted conditional responses gave another metric for evaluating how well the model represented the binary responses. The SFG had the lowest error at the onset of the stimulus and the end of the stimulus. (B) The fitted conditional responses versus the real VAS responses for the MTG during the 0:500 ms window show a proportional relationship. Not many real VAS responses were greater than five. (C) The fitted conditional responses versus the real VAS responses for the IFG during the 250:750 ms window show a proportional relationship.