Model comparison for the effect of 5-HTR blockade on error rate in the reward-size task (for Fig 4).

a(cond) a(mk), a(cond_mk), e(cond), e(mk), and e(cond_mk) indicate the random effects of blocking 5-HTR on parameters a and e, respectively. The probability distribution of the random effects were as follows: a_cond, a_mk, and a_{cond_mk} were ~N(0,σ²_cond), ~N(0,σ²_mk), and ~N(0,σ²_{cond_mk}), respectively. (a’_cond, e’_cond), (a’_mk, e’_mk), and (a’_{cond_mk}, e’_{cond_mk}) were ~biNorm(0, ∑_cond), ~biNorm(0, ∑_mk), and ~biNorm(0, ∑_{cond_mk}), respectively. E, error rate; cond, treatment condition (antagonist or control); mk, monkey; cond_mk, interaction of treatment condition and monkey. BIC is a relative measure of quality for the models (#1–30). ΔBIC denotes the difference from the minimum BIC.

(DOCX)