Mice can count and optimize count-based decisions

Abstract

Previous studies showed that rats and pigeons can count their responses, and the resultant count-based judgments exhibit the scalar property (also known as Weber's Law), a psychophysical property that also characterizes intervaltiming behavior. Animals were found to take a nearly normative account of these well-established endogenous uncertainty characteristics in their time-based decision-making. On the other hand, no study has yet tested the implications of scalar property of numerosity representations for reward-rate maximization in count-based decision-making. The current study tested mice on a task that required them to press one lever for a minimum number of times before pressing the second lever to collect the armed reward (fixed consecutive number schedule, FCN). Fewer than necessary number of responses reset the response count without reinforcement, whereas emitting responses at least for the minimum number of times reset the response counter with reinforcement. Each mouse was tested with three different FCN schedules (FCN10, FCN20, FCN40). The number of responses emitted on the first lever before pressing the second lever constituted the main unit of analysis. Our findings for the first time showed that mice count their responses with scalar property. We then defined the reward-rate maximizing numerical decision strategies in this task based on the subject-based estimates of the endogenous counting uncertainty. Our results showed that mice learn to maximize the reward-rate by incorporating the uncertainty in their numerosity judgments into their count-based decisions. Our findings extend the scope of optimal temporal risk-assessment to the domain of count-based decision-making.

An Accumulation-of-Evidence Task Using Visual Pulses for Mice Navigating in Virtual Reality

“…Specifically, our stimulus period durations were longer than when the benefits of prolonged stimulus saturate (Gold and Shadlen, 2007 ; Kiani et al, 2008 ; Brunton et al, 2013 ). On the other hand, the finding that behavior in our task was influenced by the number but not duration of cues is consistent with multiple previous studies of counting in rodents (Mechner, 1958 ; Fernandes and Church, 1982 ; Gallistel and Gelman, 2000 ; Çavdaroglu and Balci, 2016 ). Counting is thought to be carried out as a magnitude-estimation process that displays the property of scalar variability, i.e., the noise (standard deviation) in estimates scales linearly with count/magnitude (Fechner, 1966 ; Gallistel and Gelman, 2000 ).…”

An accumulation-of-evidence task using visual pulses for mice navigating in virtual reality

“…Accordingly, following a recent report in rats (Scott et al 2015) , we show through modeling that noise in the mice's estimates of the number of towers in our task scales in a way that is compatible with the phenomenon of scalar variability. We extend previous findings by showing that, in addition to selfgenerated lever presses (Çavdaroğlu and Balcı 2016) , mice can accumulate visual stimuli in the context of a perceptual decisionmaking task.…”

Numerical error monitoring

“…A contemporary line of evidence shows that humans and animals can optimize their quantitative and perceptual decisions by taking near normative account of their level of endogenous timing and counting uncertainty (Balcı et al, 2011;Çavdaroğlu & Balcı, 2016). Our results, coupled with the earlier results on temporal error-monitoring, suggest that humans and animals might adapt their decision according to their estimate regarding the level of their uncertainty.…”

“…An important feature of these scenarios is that each magnitude estimation is subject to internal sources of uncertainty leading to substantial trial-to-trial variability in behavior and characterizing every magnitude-based decision as decisions made under uncertainty. A recent line of research (e.g., Çavdaroğlu, Zeki, & Balcı, 2014;Çavdaroğlu & Balcı, 2016) has addressed the importance of the abovementioned subjective timing and counting uncertainty for decisionmaking by formulating the dependence of reward-ratemaximizing decisions on the level of uncertainty. The results of these studies showed that humans and rodents can nearly optimize their quantitative decisions by integrating the level of their subjective timing and numerical uncertainty into these decisions (Çavdaroğlu & Balcı, 2016;Çavdaroğlu, Zeki, & Balcı, 2014).…”

“…A recent line of research (e.g., Çavdaroğlu, Zeki, & Balcı, 2014;Çavdaroğlu & Balcı, 2016) has addressed the importance of the abovementioned subjective timing and counting uncertainty for decisionmaking by formulating the dependence of reward-ratemaximizing decisions on the level of uncertainty. The results of these studies showed that humans and rodents can nearly optimize their quantitative decisions by integrating the level of their subjective timing and numerical uncertainty into these decisions (Çavdaroğlu & Balcı, 2016;Çavdaroğlu, Zeki, & Balcı, 2014). These observations suggest that internal uncertainty about magnitude estimates can be adaptively integrated into the decision process as a biasing signal.…”