Comparison of abstract decision encoding in the monkey 3 prefrontal cortex ,

decision encoding. Two types of measurements were used for this evaluation. 439 First, we compared the decision SLR coefficients of decision neurons across the three 440 areas (ad, Figure 7A). Moreover, we related the strength of the decision processing of 441 these brain areas to their encoding strength of stimulus intensity, by intensity selective 442 neurons (aint, Figure 7B). We calculated a three-way ANOVA for the absolute values 443 of the SLR coefficients with the following factors: neuron type (intensityor decision444 coding), analysis phase (stimulus or delay phase), and recording area (PFC, preSMA, 445 CMAr). This analysis showed a significant effect of recording area (PFC: ad = 0.22; 446 preSMA ad = 0.27; CMAr ad = 0.24; p < 0.05, ANOVA). Post-hoc tests revealed that 447 preSMA showed significantly higher decision encoding strength than PFC (p < 0.01, 448 Wilcoxon test, Bonferroni corrected). No differences of decision encoding strength 449 were found between the analysis phases (p > 0.05, ANOVA). In all three areas the 450 encoding strength of the abstract decision (ad = 0.24) was significantly larger 451 compared to the encoding strength of stimulus intensity (aint = 0.04; p < 0.05, 452 ANOVA). 453 454 As a second measurement of the encoding strength, we used the choice probability 455 indices (sc) derived from salient and threshold trials for all recording areas during the 456 stimulus and delay phase, respectively. For a comparison of absolute differences of 457 choice probability indices, a chance discrimination level 0.5 was used (sc = | choice 458 probability – 0.5 | ). A three-way ANOVA with factors trial type (salient/threshold) × 459 trial phase (stimulus/delay phase) × recording area (preSMA, CMAr, PFC) showed a 460 significant effect of the recording area (Figure 13C, PFC: sc = 0.11; preSMA 461