Neural population dynamics underlying expected value computation

Abstract

21Computation of expected values, i.e., probability times magnitude, seems to be a 22 dynamic integrative process performed in the brain for efficient economic behavior. 23 However, neural dynamics underlying this computation remain largely unknown. 24 We examined (1) whether four core reward-related regions detect and integrate 25 the probability and magnitude cued by numerical symbols and (2) whether these 26 regions have different dynamics in the integrative process. Extractions of 27 mechanistic structure of neural population signal demonstrated that expected-28 value signals simultaneously arose in central part of orbitofrontal cortex (cOFC, 29 area 13m) and ventral striatum (VS). These expected-value signals were incredibly 30 stable in contrast to weak and unstable signals in dorsal striatum and medial OFC. 31Notably, temporal dynamics of these stable expected-value signals were 32 unambiguously distinct: sharp and gradual signal evolutions in cOFC and VS, 33 respectively. These intimate dynamics suggest that cOFC and VS compute the 34 expected-values with unique time constants, as distinct, partially overlapping 35 processes. 36 37 Impact Statement 38 Our empirical study on neural population dynamics suggests that the orbitofrontal 39 cortex and ventral striatum co-operate on expected-value computation with unique 40 time constants, as distinct, partially overlapping processes.41 42 107 1 where 1 indicates a 100% chance). After a 2.5 second delay, the visual pie-chart 108 disappeared, and a reward outcome was provided to the monkeys with the 109 indicated amount and probability of reward, unless no reward was given. Under110 6 this experimental condition, the expected values of rewards are defined as the 111 probability multiplied by the magnitude cued by the numerical symbols (Table. S1). 112To examine whether the monkeys accurately perceived the expected values 113 from the numerical symbols for probability and magnitude, we applied a choice 114 task to the monkeys (Fig. 1B). The two monkeys exhibited a near-efficient 115 performance in selecting a larger expected value option among two alternatives 116 during choice trials (Fig. 1C). We examined which of the following three behavioral 117 models best described the monkey's behavior: model 1 (M1), monkeys make 118 choices based on the number of pies; model 2 (M2), monkeys make choices based 119 on the probability and magnitude; and model 3 (M3), monkeys make choices based 120 on the expected values (Table. S1). Comparisons of the model performances 121 based on Akaike's Information Criterion (AIC) and Bayesian Information Criterion 122 (BIC) (Burnham & Anderson, 2004) revealed that the model 3 best explained the 123 monkey's behavior as indicated by the smallest AIC and BIC values (Fig. S1). The 124 model 3 consistently showed the highest pseudo r-squared values in each monkey 125 (Fig. 1D). These results indicated that the monkeys utilize the expected values 126 estimated from the numerical symbols for probability and magnitude. 127 128Neural pop...

A neuronal prospect theory model in the brain reward circuitry

“…The percentage of the right target choices was estimated from the pooled choice data for all combinations of the expected values of the left and right target options. This result has been reported previously (Yamada et al, 2021).…”

“…One possible information processing for this input-output mapping can be achieved by neural population dynamics (Chen and Stuphorn, 2015; Gardner et al, 2019; Yoo and Hayden, 2020), in which some subclusters of neurons can process information moment-by-moment as a dynamical system. Stable neural population dynamics in the VS and cOFC were indeed observed in contrast to the fluctuating signals in the DS population (Yamada et al, 2021), which may reflect some differences in distributed coding.…”

“…We estimated the monkeys’ subjective valuations of risky rewards using a gambling task (Figure 1A) (Yamada et al, 2021) similar to those used with human subjects in economics (Hey and Orme, 1994). In the choice trials, monkeys chose between two options that offered an amount of liquid reward with some probability.…”

“…Gray indicates activity showing the negative regression coefficients for probability and magnitude (P-M- type). (A)–(C) and (G) have been previously published in Yamada et al, 2021.…”

“…Baseline firing rates during 1 s before the appearance of central fixation targets were also compared with a statistical significance level of P < 0.05. These basic firing properties have been described in Yamada et al, 2021.…”