Extended Data Fig. 6: Neural–behavioral logistic model.
From: Behavioral read-out from population value signals in primate orbitofrontal cortex
Logistic regression was used to explain the choice of first vs. second offers with neural and behavioral variables described in equation (4). (a) Boxplots of regression estimates for n = 32 sessions. Thick lines and boxes show the median and IQR; whisker minima/maxima are up to three times the IQR. Estimates are in arbitrary units. Significance indicators indicate difference from zero by two-sided t-test. (b) Estimates from neural signal variables in panel (a) (full model) were no different from estimates from a model containing only the first four terms in equation (4) (minimal model). (c) Estimates from models fit using only trials with value difference of 1 drop compared to estimates fit using only trials with equal-value offers (dark vs. light bars, n = 32 sessions each). Significance indicators give difference between dark and light bar in each pair by paired, two-sided t-test. No significant differences were found after p-value correction for multiple comparisons. Note that an estimate for ‘Chosen value’ was not obtained for ‘value difference 0’, because this variable is constant within this group of trials.
