Skip to main content
PMC home page
PLOS Biology logoLink to PLOS Biology
. 2020 Aug 24;18(8):e3000841. doi: 10.1371/journal.pbio.3000841

Choosing fast and simply: Construction of preferences by starlings through parallel option valuation

Tiago Monteiro 1,*, Marco Vasconcelos 2,*, Alex Kacelnik 1,*
Editor: Alexandre Pouget3
PMCID: PMC7480835  PMID: 32833962

Abstract

The integration of normative and descriptive analyses of decision processes in humans struggles with the fact that measuring preferences by different procedures yields different rankings and that humans appear irrationally impulsive (namely, show maladaptive preference for immediacy). Failure of procedure invariance has led to the widespread hypothesis that preferences are constructed “on the spot” by cognitive evaluations performed at choice time, implying that choices should take extra time in order to perform the necessary comparisons. We examine this issue in experiments with starlings (Sturnus vulgaris) and show that integrating normative and descriptive arguments is possible and may help reinterpreting human decision results. Our main findings are that (1) ranking alternatives through direct rating (response time) accurately predicts preference in choice, overcoming failures of procedure invariance; (2) preference is not constructed at choice time nor does it involve extra time (we show that the opposite is true); and (3) starlings’ choices are not irrationally impulsive but are instead directly interpretable in terms of profitability ranking. Like all nonhuman research, our protocols examine decisions by experience rather than by description, and hence support the conjecture that irrationalities that prevail in research with humans may not be observed in decisions by experience protocols.

Introduction

The distinction between 2 kinds of processes underlying human choice [1]—a fast one (System 1), presumably shared with other species, and a slower one (System 2), involving cognitive evaluation of alternatives at the time of choice—is widely accepted [2–6] but has not yet been integrated with decision processes in other animals.

Something similar occurs with the failures of procedure invariance detected when human preferences are measured by direct rating (viz “willingness to pay”) or by choice [7], with measures of impulsiveness [8–11] (defined as a stronger preference for smaller–sooner over larger–later rewards than explained by an optimality criterion such as rate of gains over time) and with the distinction between “decisions from description” and “decisions from experience” [12, 13]. Failures of procedure invariance are critical, because they are used to support the widespread notion that preference does not exist outside choice contexts but is constructed “on the spot” through comparative evaluations at choice time [7].

We believe that all these issues are deeply interconnected and that investigating decision-making in other species with reference to hypotheses developed for humans, and vice versa, should help to develop an integrated and more general theory of decision-making.

We contribute to this integration and present critical experiments with starlings (S. vulgaris) as research species. Our main experimental findings are that (1) ranking of a measure of direct rating (defined as any graded response to individual options outside a choice context; in humans, a common direct rating metric is “willingness to pay,” and in starlings, we use response time) accurately predicts preference in choice, overcoming reported failures of procedure invariance and claims that the valuations underlying preferences at the time of choice do not exist outside choice contexts; (2) response times in choices are shorter than when only 1 option is present, contradicting the hypothesis that a cognitive comparative evaluation occurs on the spot at choice times; and (3) our subjects’ choices are not irrationally impulsive but instead are directly interpretable in terms of profitability ranking. Like all nonhuman research, our protocols examine decisions by experience rather than by description and hence, support the conjecture that irrationalities that prevail in results of human decision-making (including failures of procedure invariance) may be endemic to the so-called System 2 peculiar to our species. For an in-depth review of the empirical significance of using experience or description in human decision research see [13].

Our results confirm hypothetical models of choice that originate in foraging research. In nature, the probability (per unit of time) of encountering 2 or more independently distributed prey types simultaneously reflects the product of their probabilities, meaning that simultaneous choice must be comparatively rare. This might not be strictly so when the distribution of prey types is not independent (as when preying upon gregarious prey), but even then, the ability to select between multiple profitable prey can build upon efficient hunting of prey found on their own, whereas the reverse is not true. For these reasons, we expect stronger selection to act adaptively in sequential than in simultaneous encounters, and the psychological machinery evolved to act efficiently in sequential encounters to be biologically more significant than specific adaptations for simultaneous choice. This rationale underlies the sequential choice model [14, 15] (SCM, see S1 Text for mathematical details and numerical simulations). The SCM can be construed as a race (inhibition-free) drift-diffusion decision model (DDDM, [3, 16–20]) but with the special feature of predicting choice as a consequence of mechanisms driving action in sequential encounters. It postulates that when facing a response opportunity, the subject undergoes an accumulation process leading to action when a threshold is crossed, with the rate of accumulation and the threshold determining the distribution of response times for each prey type, depending on its profitability relative to the background rate of consumption. A difference with other DDDMs is that in the SCM the hypothetical accumulation is not interpreted as a gathering of evidence in favor of one option over the alternative but a build-up leading to action toward a stimulus regardless of whether it is alone or in the presence of alternatives. Following classical optimal foraging theory [21–25], we define profitability of a prey type as the ratio of gain (typically net energy content) to involvement time (time to pursue and handling each item), excluding time costs common to all prey types, such as travel or searching periods. Based on empirical evidence, the interval between encountering a prey type and attacking it (which has no direct interpretation in foraging theory) is modeled as being variable, with distributions’ parameters depending on both the prey type’s profitability and the rate of gain in the background: relatively richer prey are attacked faster. Critically, the SCM postulates that when more than 1 potential prey is present, a “choice” results from the independent process for each of the options present, without any “on the spot” comparison, without building evidence about relative values as assumed by classical DDDMs, and without construction of preference at the time of choice.

A consequence of these assumptions is that response times toward a given stimulus when picked out of a choice should be shorter than when the same stimulus is encountered alone (i.e., if prey types A and B are encountered simultaneously, and A is attacked, observed response time should be shorter than when A is encountered in the absence of an alternative). Further, the shortening effect should be stronger for leaner, infrequently chosen stimuli. These predictions result from assuming cross-censorship between independent processes for each of the options present and contradict the hypothesis that preference results from a comparative evaluation at choice time and hence, involves extra time. The mathematical justification for these properties is detailed in S1 Text.

We trained starlings (Fig 1a) in an environment in which 6 stimuli, each associated with a given profitability ratio (Fig 1b), were randomly encountered, either sequentially or in binary pairings requiring a choice. Notice that we make a nomenclature distinction between “profitability ratio” and “profitability.” We define profitability ratio as particular combinations of an amount and delay to food (thus, 1 pellet/10 seconds and 2 pellets/20 seconds are 2 different profitability ratios), and profitability as the quotient between them (0.1 pellets/second in both cases; this is the variable used in classical foraging theory [25]). This distinction is necessary because profitability is scale free, but options with the same profitability (as in symbols joined by lines in Fig 1b) can impact preferences if they differ in profitability ratio, and the distinction is useful to discuss impulsivity. We used 6 profitability ratios that allow the following comparisons: 3 had equal delay but differed in amount, 3 had equal amount but differed in delay, and 3 pairs had equal profitability but differed in both amount and delay. In addition to preference in binary choices, we recorded response times in single-option encounters and in the 30 types of binary-choice outcomes (6 options allow for 15 possible pairings, and in each pairing, the subject could choose either of the 2 options). We also interspersed infrequent unrewarded probe trials to assess the animals’ sensitivity to the delay associated with each stimulus. These probe trials (Fig 1c, key in colored rectangles in Fig 1b) showed that the birds learned these delays and that their timing was not distorted by differences in amount.

Fig 1. Starlings were trained in a multialternative environment.

Fig 1

Open in a new tab

a. European starling (S. vulgaris, [photo credit: Dick Daniels, http://carolinabirds.org/]/CC BY-SA https://creativecommons.org/licenses/by-sa/3.0]). b. Alternatives present in the experimental environment. Each symbol depicts a profitability ratio (a specific combination of amount and delay) and each shading, a profitability (a given ratio of amount to delay in rewards/second). Lines join stimuli with equal profitability. Within each profitability pair, the larger symbol indicates the Larger–Later profitability ratio and the smaller symbol the Smaller–Sooner one. c. Rate of responding in probe trials (mean ± SEM, n = 9 birds). All data included in this figure can be found in the S1 Data file.

As assumed by the SCM, higher profitability was associated with shorter response times in single-option trials and with higher preference in choices (Fig 2a). Moreover, the strength of preference for each option was modulated by the profitability of the alternative option (Fig 2b). A unique, specific prediction of the SCM is that response times in choices should be shorter than in single-option trials, with the effect being stronger for leaner alternatives (see S1 Text). To test for this, Fig 3a shows the difference in response time for each stimulus type when picked out of a pair compared with the same stimulus type in single-option trials. Average response times were shorter or equal in choices than in single-option encounters (Fig 3a, bottom row; Fig 3b shows further quantitative details), as predicted by the SCM, and the effect was stronger for the least-preferred alternatives. The figure also shows that this population result reflects the data from individual subjects. A bootstrapped analysis confirmed that these results, even at the individual level, were not due to chance (Fig 3c and 3d, see Analyses section for details).

Fig 2. Profitability impacts response times and choices.

Fig 2

Open in a new tab

a. Preference in choice trials (mean proportion picked when presented paired with any of the other 5 profitability ratios) versus response time in single-option trials (means ± SEM, n = 9 birds). Inset shows individual data. Symbols match panel in Fig 1b. The dashed line shows a Deming regression (Estimated model: y = −0.0072x + 0.8259, [slope = 95% CI −0.004 to −0.0104]). b. Mean proportion of choices in every binary-choice combination (n = 9 birds). All data included in this figure can be found in the S1 Data file.

Fig 3. RTs are shorter in choices.

Fig 3

Open in a new tab

a. ΔRT toward each profitability ratio between single-option and choice trials, averaged against the other 5 combinations; bottom row shows means across subjects. White cells are tied cases in which differences were lower than 1s in either direction. b. Number of nontied cases (out of 54) as a function of increasing minimal absolute response time difference (black) and corresponding binomial p-value (gray). c. Mean differences of individual mean response times in single-option trials minus individual mean response times in choice trials (n = 9, mean ± SEM). Inset shows individual animal means, with symbols laterally displaced for visualization purposes. (same data as bottom row in a.) d. Same as in c. for bootstrapped differences. A repeated-measures ANOVA, with option (6 levels) and difference as factors (2 levels), and mean ΔRTs as dependent variable, showed a significant effect of difference, only (Greenhouse–Geisser adjusted F(1,8) = 7.307, p = 0.027, η2p = 0.660). All data included in this figure can be found in the S1 Data file. RT, response time; ΔRT, difference in response time.

The common assumption that decision-makers show an impulsive bias to prefer smaller–sooner rewards over equally profitable larger–later ones is explored in Fig 4a. In the pairwise comparisons in which the profitability of the options was equated, the birds, in fact, showed the opposite: They preferred the larger–later alternative over the smaller–sooner one. This result contradicts prevalent ideas on impulsivity [9] but is consistent with reward rate maximization with partial account of the common time intervals [26–29] (Fig 4b).

Fig 4. When profitability is equated, starlings prefer larger–later alternatives.

Fig 4

Open in a new tab

a. Proportion larger–later alternative chosen against the equal profitability smaller–sooner combination (mean ± SEM, n = 9 birds). Dots show individual animals, laterally displaced for visualization purposes. Two-tailed t-tests against 50%, smallest t(8) = 2.912, *p < 0.05, **p < 0.01, smallest Cohen’s d = 2.05, effect size = [0.71, 0.96]. b. Long-term rate of returns R in which the time in the ITI is given the same decision weight as the delay between response and outcome (R = amount/(ITI + delay); equivalent to the slope of the dashed lines) and profitability as in Fig 1b (slope of solid lines). Inset shows R2 of linear regressions between observed preference ranking (mean across animals, n = 9) and long-term estimated rate of returns when the common ITI is given a discounted weight (R = amount / (w*ITI + delay), computed with increasing w); note that the strongest predictive power is for a long-term rate computation that includes an 8% weight of the ITI. All data included in this figure can be found in the S1 Data file. ITI, intertrial-interval.

In summary, we show procedure invariance between direct rating and choice when using response time as a direct rating metric, and we find that the SCM is appropriate to describe the choice process. This contrasts with the failure of procedure invariance characteristic of human experiments, but this does not necessarily imply species’ differences. It is worth recalling that the discovery of failures of procedure invariance in humans are typical of experiments based on description rather than experience, and such experiments are not possible in nonhumans. It is thus possible that such failures are peculiar of human protocols using description rather than experience. Whether this or genuine species differences in decision mechanisms are responsible for the contrast between our results and those reported in humans remains to be determined. These possibilities can be differentiated by replicating our procedures with human subjects. If humans behave in such protocols as starlings do, the problem becomes mapping the significance of the finding in human experience. It is possible that the processes we report, and the foraging rationale that explains them, are relevant in humans, but only for problems recruiting the so-called System 1 of human decision-making and that they are consistent with the evidence for different neural representation of the values of alternatives and the background [30, 31] but are not relevant to the reasoning-based decision process that prevails when System 2 is in action.

Materials and methods

Ethics statement

All experiments were approved by the Department of Zoology Ethical Committee, University of Oxford, and were carried out in accordance with the current laws of the United Kingdom. Animals were obtained under English Nature license No. 20010082 and were cared in accordance with the University of Oxford’s “gold standard” animal care guidelines. Maintenance and experimental protocols adhered to the Guidelines for the Use of Animals in Research from the Association for the Study of Animal Behaviour/Animal Behavior Society [32].

On completion, the birds were reintroduced into communal aviaries and eventually released into the wild.

Subjects

Twelve adult wild-caught European starlings (S. vulgaris) participated in the current experiment. All subjects were naive to experimentation. Prior to testing, birds were housed communally in a large outdoor aviary (3 meters high × 3.2 meters wide × 4 meters long) where they received ad libitum food (a mixture of turkey crumbs and mealworms (Tenebrio sp.) and had continuous access to fresh drinking and bathing water.

During the course of the experiment, starlings were moved in pairs to indoor cages, which were visually but not acoustically isolated. Indoor temperatures ranged from 16 °C to 18 °C, and lights followed a 12:12 dark schedule with daylight at 07h00, and gradual transitions at dawn and dusk. Starlings were rewarded with precision pellets throughout the experimental session and provided with 4 hours (13h00–17h00) of free access to a mixture of turkey crumbs and Orlux Remiline universal granules, mealworms, and social interaction with the respective pair. This regime maintains starlings at approximately 90% of their free feeding weight. Birds had continuous access to drinking water and received bathing trays every day.

Apparatus

Each indoor cage (1,350 mm × 784 mm× 800 mm [l × w × h]) was composed of 2 units, vertically mounted (800 mm each). Each unit included 2 experimental areas separated by a common middle section. Each experimental area had a panel attached 100 mm above the floor (S1a Fig). The panel was 400-mm tall with 3 sections: a middle subpanel, facing the cage (115-mm wide), and 2 side subpanels (same length) attached to the cage at a 120° angle from the center subpanel. The middle subpanel had 1 response key in the center (110 mm from the bottom) and the food hopper (25 mm from the bottom), which was connected to the pellet dispenser (Campden Instruments) containing 20-mg precision dustless pellets. Each side subpanel had 1 response key in the center (110 mm from the bottom). Every key was composed of a 16-LED lights matrix that could display 16 different symbols in 7 possible colors (S1b Fig). Trials were governed by custom software running on the Microsoft Windows operating system and attached to an Animal Behaviour Environment Test System (Campden Instruments) via WhiskerServer [33].

Pretraining

Following magazine training, in the first phase of pretraining, the birds were trained to peck a flashing white center key (700 milliseconds ON, 300 milliseconds OFF). A single peck to this key delivered 4 precision pellets, extinguished the keylight, and initiated a 120-second intertrial-interval (ITI). Failing to peck the key within 30 seconds from the onset of flashing resulted in the same events as a peck. Training thus combined autoshaping with an operant response. Once birds were reliably pecking the center key, the lateral keys were introduced. Trials again began with the flashing of a white center key. A peck to this key turned its light off and started flashing a white lateral key. A peck to the flashing lateral key turned it off and produced 2 food pellets followed by a 90-second ITI. During this phase, failing to peck the center key within 30 seconds from trial onset resulted in a 10-second period with the light switched off, followed by the ITI. Failing to peck the flashing side key during the same time period was followed by food 75% of the time. This program was maintained 5 hours per day until the birds reliably pecked the center and side keys on at least 80% of the trials.

Next, birds received response-initiated fixed interval (FI) training on the side keys. In this phase, 2 pecks to the side key were required to deliver food. The first peck turned the symbol steadily on and began a delay, the duration of which was progressively increased from 1 to 20 seconds. The first peck after the delay elapsed extinguished the keylight, delivered 2 food pellets, and initiated a 45-second ITI. Once birds were reliably obtaining food on FI 20-second trials, the experimental sessions began.

For experimental sessions, a different set of symbols/colors were used for each bird.

General procedure

Starlings were trained in an environment consisting of 6 different options composed by a different amount/delay to food combinations (2 pellets in 5 seconds, 4 pellets in 10 seconds, 1 pellet in 5 seconds, 2 pellets in 10 seconds, 1 pellet in 10 seconds, and 2 pellets in 20 seconds); hence, each option signaled 1 out of 3 possible profitability ratios (0.4, 0.2, and 0.1 pellets per second, Fig 1b).

Each training session lasted a maximum of 5 hours and was composed of 252 trials (240 single-option trials, 40 per option, 20 in each possible location; and 12 probe trials, 2 per option, 1 in each possible location; see details next).

Birds were moved to a testing phase after completing 15 days with more than 75% completion of daily trials or after 20 days of 60% of daily completion, whichever came first. In testing, single-option trials were interspersed with binary-choice trials between all possible combinations of different options. Testing sessions were also 252 trials and/or 5 hours long (120 single-option trials, 20 per option, 10 in each possible location; 120 binary-choice trials, 8 per pairing, 4 in each possible location; and 12 probe trials, 2 per option, 1 in each possible location). Testing ended after 2 blocks of 3 consecutive days with no trend in preferences (and SD < 0.15) or after 15 days, whichever came first.

Trial types

Single-option trials began with the flashing illumination of the center attention key. A peck to this flashing key turned the key off and began flashing (700 milliseconds ON, 300 milliseconds OFF) 1 of the 2 side keys (randomly assigned). A peck to the illuminated flashing side key turned the key steadily on and initiated its programmed delay. Once this interval elapsed, the first peck to the key turned the keylight off and delivered the predefined number of pellets of reward at a rate of 200 milliseconds per pellet, after which a 45-second ITI was initiated. Not pecking the keys at any stage (flashing or steady keys) meant not moving forward to the next trial.

Probe trials were as single-option trials, but once pecked, a given symbol remained on for 3 times its typical delay. At the conclusion of this interval, the light extinguished without reward or requiring a peck, followed by the ITI [34, 35].

Choice trials offered individuals simultaneous choices between combinations of pairs of training stimuli, chosen among 15 possible binary-combinations of the 6 alternatives. Trials again began with a flashing center attention key. A peck to this key turned it off and began flashing the 2 side keys (each one displaying a different symbol). The location of presentation was also randomized. A peck to one of the flashing keys turned the selected key on and initiated its programmed contingencies and turned the unselected key off. Thus, the bird committed itself to a given option with a single peck. Absence of pecking produced the same consequences as in single-option trials.

Data selection

Only the last 6 testing days (i.e., for preferences and response times) were considered in the analysis. Three birds were excluded from all analyses after repeatedly failing to reach the experimental criteria (see General procedure).

Analyses

All analyses were run in custom MATLAB R2019a software (Mathworks), available upon request.

Deming regression analysis in Fig 1d uses code from [36], retrieved from MATLAB Central File Exchange. (https://www.mathworks.com/matlabcentral/fileexchange/33484-linear-deming-regression).

We are interested in the difference in response times for each option when picked up of a choice as opposed to when the option is the only one available. The “construction of preference” hypothesis assumes that choices should take time, so reaction times should be longer, whereas the SCM predicts that reaction times in choices should be shorter. As Fig 3a shows, reaction times were predominantly shorter in choices, and the present analysis aims at testing whether the observed results could be outcome of chance. To this effect, we performed a bootstrap analysis (Fig 3c and 3d) in which we considered the null hypothesis that both single-option and choice reaction times are sampled from the same distribution, namely, that there is no systematic difference between them. Under this null hypothesis, for each animal and option, all response times collected during this experiment (i.e., from single-option and choice trials) come from the same underlying distribution. If that is the case, any difference between average response times in single-option versus choice trials (i.e., ΔRT) is simply due to chance.

To test the null hypothesis, we generated subject-specific null distributions of response times. For each animal and option, we pooled response times for single-option and choice trials, so that each animal had 6 pools of response times, one per option. Then, we simulated 105 experiments per option, where in each iteration, we generated n choice response times (where n was the animal’s real number of choices for that option), drawn with replacement from the response time pool corresponding to that option. Next, we calculated the average for each of these simulated distributions and subtracted these averages from the corresponding real (observed) mean single-option response time (i.e., bootstrapped difference).

For example, bird number 1 gave us 77 response time values for option 2/20 in single-option trials and 46 response times from its choices of option 2/20 in choice trials. So we formed a single pool of 77+46 response times for that subject with its real response times. In each simulated iteration, we randomly selected 46 response time values from its pool, computed the average (n = number of simulations) and subtracted this mean from the mean of observed reaction times in that subject’s choices of option 2/20 − ΔRT.

Finally, we compared the simulated ΔRT with the observed ΔRT using a repeated-measures ANOVA.

Supporting information

S1 Text. The SCM.

Computational description, analytical implementation and numerical simulations. SCM, sequential choice model.

(PDF)

Click here for additional data file. (373.7KB, pdf)
S1 Fig

a. Schematics of the operant panels (adapted with permission from [15], APA). b. Close-up photographs of the operant panels displaying different symbols (top) and colors (bottom).

(TIF)

Click here for additional data file. (3.6MB, tif)
S1 Data. An Excel file with all data used in the main figures, with each tab corresponding to a figure panel, from main Figs 1 to 4, respectively.

Tabs from the same figure are labelled with the same color.

(XLSX)

Click here for additional data file. (36.3KB, xlsx)

Acknowledgments

We thank Simon and Jan Parry for trapping our participants. We would like to thank Miguel Santos, Andreia Madalena, Filipe Rodrigues, Catarina Albergaria, Bruno Cruz, Sofia Soares, Andrés Ojeda, Mark Walton, and Alasdair Houston for feedback on earlier versions of the manuscript.

Abbreviations

DDDM

drift-diffusion decision model

SCM

sequential choice model

Data Availability

All data can be found in the S1 Data file.

Funding Statement

This research was supported by Grant BB/G007144/1 from the UK Biotechnology and Biological Sciences Research Council to AK. TM was supported by a doctoral grant from the Portuguese Foundation for Science and Technology (SFRH/BD/43385/2008) and a Pembroke College Graduate Scholarship. MV was supported by the Portuguese Foundation for Science and Technology (UIDB/04810/2020). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

References

  • 1.Kahneman D. Thinking, Fast and Slow. Penguin; UK; 2012. 499 p. [Google Scholar]
  • 2.Rangel A, Camerer C, Montague PR. A framework for studying the neurobiology of value-based decision making. Nat Rev Neurosci. 2008. July;9(7):545–56. 10.1038/nrn2357 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 3.Bogacz R, Brown E, Moehlis J, Holmes P, Cohen JD. The physics of optimal decision making: A formal analysis of models of performance in two-alternative forced-choice tasks. Psychological Review. 2006;113(4):700–65. 10.1037/0033-295X.113.4.700 [DOI] [PubMed] [Google Scholar]
  • 4.Oud B, Krajbich I, Miller K, Cheong JH, Botvinick M, Fehr E. Irrational time allocation in decision-making. Proc Biol. 2016. January 13;283(1822). 10.1098/rspb.2015.1439 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 5.Sugrue LP, Corrado GS, Newsome WT. Choosing the greater of two goods: neural currencies for valuation and decision making. Nat Rev Neurosci. 2005. May;6(5):363–75. 10.1038/nrn1666 [DOI] [PubMed] [Google Scholar]
  • 6.Levy DJ, Glimcher PW. The root of all value: a neural common currency for choice. Curr Opin Neurobiol. 2012. December;22(6):1027–38. 10.1016/j.conb.2012.06.001 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 7.Slovic P. The construction of preference. American Psychologist. 1995;50(5): 364–71. 10.1037/0003-066x.50.5.364 [DOI] [Google Scholar]
  • 8.Stephens DW, Anderson D. The adaptive value of preference for immediacy: when shortsighted rules have farsighted consequences. Behavioral Ecology. 2001;12(3): 330–9. 10.1093/beheco/12.3.330 [DOI] [Google Scholar]
  • 9.Frederick S, Loewenstein G, O’donoghue T. Time discounting and time preference: a critical review. Journal of Economic Literature. 2002;40(2):351–401. 10.1257/jel.40.2.351 [DOI] [Google Scholar]
  • 10.Fehr E. The economics of impatience. Nature. 2002. January 17;415(6869):269–72. 10.1038/415269a [DOI] [PubMed] [Google Scholar]
  • 11.Kacelnik A. The evolution of patience In: Loewenstein G., Read D., Baumeister R., editor. Time and decision: Economic and psychological perspectives on intertemporal choice. 2003. p. 115–38. [Google Scholar]
  • 12.Ludvig EA, Spetch ML. Of black swans and tossed coins: is the description-experience gap in risky choice limited to rare events? PLoS ONE. 2011. June 1;6(6):e20262 10.1371/journal.pone.0020262 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 13.Hertwig R, Erev I. The description-experience gap in risky choice. Trends Cogn Sci. 2009. December;13(12):517–23. 10.1016/j.tics.2009.09.004 [DOI] [PubMed] [Google Scholar]
  • 14.Kacelnik A, Vasconcelos M, Monteiro T, Aw J. Darwin’s “tug-of-war” vs. starlings’ “horse-racing”: how adaptations for sequential encounters drive simultaneous choice. Behavioral Ecology and Sociobiology. 2011;65:547–58. 10.1007/s00265-010-1101-2 [DOI] [Google Scholar]
  • 15.Shapiro MS, Siller S, Kacelnik A. Simultaneous and sequential choice as a function of reward delay and magnitude: normative, descriptive and process-based models tested in the European starling (Sturnus vulgaris). J Exp Psychol Anim Behav Process. 2008. January;34(1):75–93. 10.1037/0097-7403.34.1.75 [DOI] [PubMed] [Google Scholar]
  • 16.Ratcliff R, Smith PL, Brown SD, McKoon G. Diffusion Decision Model: Current Issues and History. Trends in Cognitive Sciences. 2016;20:260–81. 10.1016/j.tics.2016.01.007 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17.Ratcliff R, McKoon G. The diffusion decision model: theory and data for two-choice decision tasks. Neural Comput. 2008. April;20(4):873–922. 10.1162/neco.2008.12-06-420 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18.Ratcliff R. A theory of memory retrieval. Psychological Review. 1978;85(2):59–108. 10.1037/0033-295x.85.2.59 [DOI] [Google Scholar]
  • 19.Brunton BW, Botvinick MM, Brody CD. Rats and humans can optimally accumulate evidence for decision-making. Science. 2013;340(6128):95–8. 10.1126/science.1233912 [DOI] [PubMed] [Google Scholar]
  • 20.Drugowitsch J, Mendonça AG, Mainen ZF, Pouget A. Learning optimal decisions with confidence. Proc Natl Acad Sci U S A. 2019. December 3;116(49):24872–80. 10.1073/pnas.1906787116 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21.Charnov EL. Optimal Foraging: Attack Strategy of a Mantid. Vol. 110, The American Naturalist. 1976;110(971):141–51. 10.1086/283054 [DOI] [Google Scholar]
  • 22.Charnov EL. Optimal foraging, the marginal value theorem. Theoretical population biology. 1976;9(2):129–36. 10.1016/0040-5809(76)90040-x [DOI] [PubMed] [Google Scholar]
  • 23.Cowie RJ. Optimal foraging in great tits (Parus major). Nature. 1977;268:137–9. 10.1038/268137a0 [DOI] [Google Scholar]
  • 24.Kacelnik A. Central Place Foraging in Starlings (Sturnus vulgaris). I. Patch residence time. The Journal of Animal Ecology. 1984;53(1):283–299. 10.2307/4357 [DOI] [Google Scholar]
  • 25.Stephens DW, Krebs JR. Foraging Theory. Princeton University Press; 1986. 247 p. [Google Scholar]
  • 26.Bateson M, Kacelnik A. Rate currencies and the foraging starling: the fallacy of the averages revisited. Behavioral Ecology. 1996;70:341–52. 10.1093/beheco/7.3.341 [DOI] [Google Scholar]
  • 27.Mazur JE. Hyperbolic value addition and general models of animal choice. Psychol Rev. 2001. January;108(1):96–112. 10.1037/0033-295x.108.1.96 [DOI] [PubMed] [Google Scholar]
  • 28.Hayden BY. Time discounting and time preference in animals: A critical review. Psychon Bull Rev. 2016. February;23(1):39–53. 10.3758/s13423-015-0879-3 [DOI] [PubMed] [Google Scholar]
  • 29.Drugowitsch J, Moreno-Bote R, Churchland AK, Shadlen MN, Pouget A. The cost of accumulating evidence in perceptual decision making. J Neurosci. 2012. March 14;32(11):3612–28. 10.1523/JNEUROSCI.4010-11.2012 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 30.Kolling N, Behrens TEJ, Mars RB, Rushworth MFS. Neural mechanisms of foraging. Science. 2012;336:95–8. 10.1126/science.1216930 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 31.Hayden B, Moreno-Bote R. A neuronal theory of sequential economic choice. Brain and Neuroscience Advances. 2018;2:1–15. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 32.Guidelines for the treatment of animals in behavioural research and teaching. Animal Behaviour. 2006;71:245–53. 10.1016/j.anbehav.2005.10.001 [DOI] [PubMed] [Google Scholar]
  • 33.Cardinal RN, Aitken MRF. Whisker: a client-server high-performance multimedia research control system. Behav Res Methods. 2010. November;42(4):1059–71. 10.3758/BRM.42.4.1059 [DOI] [PubMed] [Google Scholar]
  • 34.Roberts S. Isolation of an internal clock. Journal of Experimental Psychology: Animal Behavior Processes. 1981;7:242–68. 10.1037//0097-7403.7.3.242 [DOI] [PubMed] [Google Scholar]
  • 35.Catania AC, Reynolds GS. A quantitative analysis of the responding maintained by interval schedules of reinforcement. Journal of the Experimental Analysis of Behavior. 1968;11:327–83. 10.1901/jeab.1968.11-s327 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 36.James Hall (2020). Linear Deming Regression. (https://www.mathworks.com/matlabcentral/fileexchange/33484-linear-deming-regression), MATLAB Central File Exchange.

Decision Letter 0

Gabriel Gasque

3 Feb 2020

Dear Dr Monteiro,

Thank you for submitting your manuscript entitled "Thinking fast and simple: how decisions are taken" for consideration as a Short Report by PLOS Biology. Please accept my apologies for the delay in sending this decision to you. We were interested in your study, and thus, sought advice from an Academic Editor with relevant expertise. With that advice now in hand, I am writing to let you know that we would like to send your submission out for external peer review.

However, please note that we would need to be persuaded by the reviewers that the findings are as novel and that they extend to human decision-making as directly as you have suggested for further consideration.

In addition, before we can send your manuscript to reviewers, we need you to complete your submission by providing the metadata that is required for full assessment. To this end, please login to Editorial Manager where you will find the paper in the 'Submissions Needing Revisions' folder on your homepage. Please click 'Revise Submission' from the Action Links and complete all additional questions in the submission questionnaire.

Please re-submit your manuscript within two working days, i.e. by Feb 05 2020 11:59PM.

Login to Editorial Manager here: https://www.editorialmanager.com/pbiology

During resubmission, you will be invited to opt-in to posting your pre-review manuscript as a bioRxiv preprint. Visit http://journals.plos.org/plosbiology/s/preprints for full details. If you consent to posting your current manuscript as a preprint, please upload a single Preprint PDF when you re-submit.

Once your full submission is complete, your paper will undergo a series of checks in preparation for peer review. Once your manuscript has passed all checks it will be sent out for review.

Feel free to email us at plosbiology@plos.org if you have any queries relating to your submission.

Kind regards,

Gabriel Gasque, Ph.D.,

Senior Editor

PLOS Biology

Decision Letter 1

Gabriel Gasque

19 Mar 2020

Dear Dr Monteiro,

Thank you very much for submitting your manuscript "Thinking fast and simple: how decisions are taken" for consideration as a Short Report at PLOS Biology. Your manuscript has been evaluated by the PLOS Biology editors, by an Academic Editor with relevant expertise, and by four independent reviewers. Please accept my apologies for the delay in communicating the decision below to you.

In light of the reviews (below), we will not be able to accept the current version of the manuscript, but we would welcome re-submission of a much-revised version that takes into account the reviewers' comments. We cannot make any decision about publication until we have seen the revised manuscript and your response to the reviewers' comments. Your revised manuscript is also likely to be sent for further evaluation by the reviewers.

We expect to receive your revised manuscript within 2 months.

Please email us (plosbiology@plos.org) if you have any questions or concerns, or would like to request an extension. At this stage, your manuscript remains formally under active consideration at our journal; please notify us by email if you do not intend to submit a revision so that we may end consideration of the manuscript at PLOS Biology.

**IMPORTANT - SUBMITTING YOUR REVISION**

Your revisions should address the specific points made by each reviewer. As a Short Report, please keep the number of main figures down to four. Please also submit the following files along with your revised manuscript:

1. A 'Response to Reviewers' file - this should detail your responses to the editorial requests, present a point-by-point response to all of the reviewers' comments, and indicate the changes made to the manuscript.

*NOTE: In your point by point response to the reviewers, please provide the full context of each review. Do not selectively quote paragraphs or sentences to reply to. The entire set of reviewer comments should be present in full and each specific point should be responded to individually, point by point.

You should also cite any additional relevant literature that has been published since the original submission and mention any additional citations in your response.

2. In addition to a clean copy of the manuscript, please also upload a 'track-changes' version of your manuscript that specifies the edits made. This should be uploaded as a "Related" file type.

*Re-submission Checklist*

When you are ready to resubmit your revised manuscript, please refer to this re-submission checklist: https://plos.io/Biology_Checklist

To submit a revised version of your manuscript, please go to https://www.editorialmanager.com/pbiology/ and log in as an Author. Click the link labelled 'Submissions Needing Revision' where you will find your submission record.

Please make sure to read the following important policies and guidelines while preparing your revision:

*Published Peer Review*

Please note while forming your response, if your article is accepted, you may have the opportunity to make the peer review history publicly available. The record will include editor decision letters (with reviews) and your responses to reviewer comments. If eligible, we will contact you to opt in or out. Please see here for more details:

https://blogs.plos.org/plos/2019/05/plos-journals-now-open-for-published-peer-review/

*PLOS Data Policy*

Please note that as a condition of publication PLOS' data policy (http://journals.plos.org/plosbiology/s/data-availability) requires that you make available all data used to draw the conclusions arrived at in your manuscript. If you have not already done so, you must include any data used in your manuscript either in appropriate repositories, within the body of the manuscript, or as supporting information (N.B. this includes any numerical values that were used to generate graphs, histograms etc.). For an example see here: http://www.plosbiology.org/article/info%3Adoi%2F10.1371%2Fjournal.pbio.1001908#s5

*Blot and Gel Data Policy*

We require the original, uncropped and minimally adjusted images supporting all blot and gel results reported in an article's figures or Supporting Information files. We will require these files before a manuscript can be accepted so please prepare them now, if you have not already uploaded them. Please carefully read our guidelines for how to prepare and upload this data: https://journals.plos.org/plosbiology/s/figures#loc-blot-and-gel-reporting-requirements

*Protocols deposition*

To enhance the reproducibility of your results, we recommend that if applicable you deposit your laboratory protocols in protocols.io, where a protocol can be assigned its own identifier (DOI) such that it can be cited independently in the future. For instructions see: https://journals.plos.org/plosbiology/s/submission-guidelines#loc-materials-and-methods

Thank you again for your submission to our journal. We hope that our editorial process has been constructive thus far, and we welcome your feedback at any time. Please don't hesitate to contact us if you have any questions or comments.

Sincerely,

Gabriel Gasque, Ph.D.,

Senior Editor

PLOS Biology

*****************************************************

REVIEWS:

Reviewer #1: Monteiro et all provide a very interesting set of results regarding the nature and mechanisms for decision making by using birds as an animal model. They find that choices are not made on the spot but rather reflect some intrinsic preferences. Additionally, they find that decision making in their task does not involve direct comparison between alternatives, and rather the behavior can be explained by a model where options are selected randomly one from each other. In general, the paper is well written, although I think that the claims about generality to other conditions could be toned down slightly for a more balanced manuscript.

The description of the SC Model is difficult to follow. It is described in previous literature, but I think that a simplified mathematical description of it in the SI will be valuable for the readers. In addition, one could use that opportunity to develop the basic predictions and their underlying intuitions.

Indeed, the prediction that RTs should be shorter for pair vs single decisions is a very strong one. However, I have a number of concerns related to its experimental testing. First, the prediction is qualitative, rather than quantitative. However, the SC model should predict by how much one expects to reduce RT. I think that making a semi-quantitative analysis of the model and the data would be important. Otherwise, it is unclear whether short RT times in pair vs single choices provides really support in favor of the SCM and against the DDM, or whether it provides support for a different interpretation of the results (that I develop below).

I am concerned that the shorter RTs observed in the pair choice condition is due to some attentional effects that do not have anything to do -a priori- with the choice mechanics. In the experimental setup, in pair conditions there is more physical stimulation than in single conditions -there are two choices and the amount of light over the display, if I understand well, is also larger. It can happen that more stimulation can increase attention and therefore reducing RTs. I wonder whether if one can control from arousal/attentional effects by controlling for differences in low level features of the stimulus (such as total brightness of the choices between single and pair conditions), or whether actually the authors have already taken care of this.

I have some comments about Figure 1. I would propose using the same color code in panels b and c. Right now it is quite confusing using colors for one panel, and gray levels for the other. In panel e, it is difficult to see the large differences in RT. I think it would be better to plot RT for single vs RT for pairs, and see that this line has lower slope than one, or something on those lines.

Minor comments:

-I don't see how reference 6 first into supporting Systems 1 vs Systems 2 hypothesis. Also I would say that this distinction is largely controversial, at least many people think that System 1 is not sloppy and rather follows the rule of automatic Bayesian inference.

-In the main text the definition of profitability is not very transparent. I would define the mathematical terms in the text by using symbols, and just spell out the definition of profitability using an equation.

-In the list references 26-28 for reward maximization in sequential decision making, Drugowitsch et al, Journal of Neuroscience, 2012 could be a missing relevant reference.

Reviewer #2: This paper addresses the interesting topic of whether a non-human animal will show processes similar to the fast and slow decision processes famously reported for humans. As discussed, humans have been found to show failures of procedural invariances in their decision making and have been suggested to make different decisions depending on which system is activated.

In this research, starlings learned six options that differed in amount and delay to food. Latency to peck on single option trials was predictive of preference on binary choice trials -showing procedural invariance - and latencies were shorter on choice trials than when a single option was presented - inconsistent with a deliberation process on choice trials. The birds also did not show an irrational tendency to choose smaller sooner rewards.

The topic is important, the results are clear, and there are interesting messages in the paper, for example that simultaneous decisions are likely rare and less important in nature. However, I have several problems with the current manuscript, some of which likely stem from the short report format that does not provide space to clarify and qualify statements.

For example, "irrational impulsivity" is introduced in the abstract as a human problem. The end of the abstract suggests that irrationality in humans could be due to the use of description. The authors may be just referring to procedural invariance here but since impulsivity is labelled as irrational this is certainly not clear. I think a reader who doesn't know the literature would conclude from reading the paper that impulsivity is a human problem and that these results suggest that since starlings don't show it in an experience-based tasks that it may be because of the use of described problems for humans. What is not discussed is that other animals also show impulsivity and humans do too in experience-based tasks (e.g. Jimura et al., 2009). So I think more needs to be said about how and why the starlings appear to differ from not only humans on described choices but also other animals and humans on experience-based tasks.

I am also not clear on what aspect of these results are novel. The main findings stated in the abstract are: "(1) ranking alternatives through direct rating (response time) accurately predicts preference in choice, overcoming failures of procedural invariance; (2) preference is not constructed at choice time nor does it involve time (we show that the opposite is true); and (3) starlings' choices are not irrationally impulsive, but instead directly interpretable in terms of profitability ranking ". Certainly, the first two of these have been shown and argued in some of the authors' previous work. Thus I think that the one-sentence conclusion: "The hypothesis that preferences are built at choice time is contradicted by data on starlings' decisions by experience" has been reached previously. If not, then it needs to be explained how this goes beyond what was done before to reach this conclusion. For the third finding I would like to know more about how it differs from Shapiro et al., 2008.

I suspect that these concerns are because the authors did not have the space to properly discuss the research novelty and implications but I don't think it should be published until it is better explained.

Reviewer #3: This paper presents an analysis of choice and response time in Starlings under different delays and magnitudes. I think these data are valuable, but the paper is confusingly written, making it difficult for me to assess its contribution. My impression is that the authors are making very strong claims about the nature of decision making beyond what the data support.

Major comments:

The exposition could use some work. A number of concepts are introduced in the first paragraph (procedural invariance, decisions from description vs experience, willingness to pay) without explanation. I think the authors should introduce these concepts more systematically.

p. 3: "A consequence of these assumptions is that response times towards a given stimulus when picked out of a choice should be shorter than when the same stimulus is encountered alone." I must confess that, based on the description of the model in the preceding paragraph, I don't see where this prediction comes from.

p. 3: what is "cross-censorship"?

p. 4: "In the pairwise comparisons where the profitability of the options was equated, the birds in fact showed the opposite: they preferred the larger-later alternative over the smaller-sooner one. This result contradicts prevalent ideas on impulsivity but supports reward rate maximization with partial account of the common time intervals." I see a few problems with this statement. First, a preference for larger-later does not mean that the animals not impulsive (they could still have a time preference expressed as a discount function), although it's not clear from this statement what exactly the alternative account is or what it predicts. Second, the reward rate maximization account is not explained enough to understand what it does or doesn't predict relative to other models. If one is going to discuss computational models in a paper, it would be very useful to actually simulate the models and fit them to data, compare the data to other models, and so on.

Showing procedural invariance using a response time measure does not necessarily contradict the results with humans where choice is the dependent measure. In any case, I think the onus is on the authors to demonstrate why you get procedural invariance in one situation/species and not another. This has to be more than just hand-waving about System 1 vs. System 2.

I found it curious that the authors contrast their work with "human experiments, which are mostly based on description rather than experience" (p. 5). But they neglect to mention that there is a now enormous human experimental literature on the description-experience gap. I see only one paper from this literature (Ludvig & Spetch, 2011) cited here. The influential review by Hertwig & Erev (2009) has already been cited over 600 times.

I think it is problematic to call response time a "rating". This is meant to draw an analogy to human desirability ratings, but it's not clear to me that these are measuring the same thing.

The finding that response time is predictive of choice is not new, at least in the human choice literature. Indeed, this is one of the key phenomena motivating sequential sampling models of choice (some of which the authors cite).

There is only one statistical analysis reported in the paper, in the caption of Figure 1, and the results are quite weak: only one comparison yields a p-value below 0.05. This does not lend strong support to the authors' claims.

Minor comments:

p. 2: "maybe endemic" -> "may be endemic"

Sometimes "procedure invariance" is used instead of "procedural invariance"

Caption of Fig 1: I find it misleading to label p<0.1 with an asterisk.

Reviewer #4: This interesting paper reports the results of a food choice experiment in Starlings. The starting point of the paper is the question whether preferences are constructed on the spot during decision making or whether they represent exist some basic approach tendencies that also guide binary choices. If the former is true, then choices should take extra time to perform compared to single-item approach decisions. However, the experiments show that this is not the case. Binary choices can be well predicted using the response times of single-item approach decisions and choices in general reflect profitability rankings of the options. The authors conclude that "irrationalities that prevail in research with humans may not show in decisions by experience protocols».

There is a lot to like about this paper. It formally tests decision theories developed for human choice in an animal population that is not routinely tested and comes up with clean, somewhat surprising results. The methods are all fine and the results are definitively thought-provoking and should be published in some form. However, I do not fully agree with the framing and interpretation of the results and think the authors would need to change the relevant parts of the manuscript so that this is not misleading.

1) The authors claim their results show that decisions differ between experience-based protocols and protocols by description, and that human research would need to test experience-based protocols to establish whether all previous results supporting preference construction really reflect this methodical difference. The implication of this statement is that findings of human research reflect, at least in parts, linguistic decoding of option descriptions. However, this is misleading. There are numerous experience-based protocols in human research by now, for instance experiments where participants pick between depicted food items and get to actually eat one of them. I cannot see how these protocols reflect language processing any more than the disk-pecking protocol applied here. The authors would need to take this into account when comparing experiments conducted in humans and animals.

2) The authors give very little space to discussion of other possible explanations for their findings and their divergence to studies in humans. They do mention possible interspecies differences at least briefly, but they do not discuss whether their findings may be specific to pecking for food. It seems a bit of a stretch to me to draw conclusions based on the pecking protocol about e.g. human decisions where to go on holiday. The manuscript would benefit if the authors could be more precise which types of choices they talk about and which ones are probably fundamentally different and thus not affected by the present findings.

3) The authors give very little space to the explanation of their SCM model and its predictions. In particular, the sentence that the reaction time effects result from cross-censoring of choice alternatives will be next to impossible to understand for naïve readers. The authors should at least provide references to papers explaining this model and its implications in detail when describing their predictions.

4) The authors claim that animals are very unlikely to encounter several prey at the same time. This is obviously not true for animals who hunt animals in herds and need to decide which animal to go for.

5) The authors interpret their findings in terms of the involvement of system 1 and system 2, and claim that system 2 is language based. Again, this statement is misleading. There is no direct link between system 2 and language, and it is very well possible to ponder visually-presented choice options without involving language.

Decision Letter 2

Gabriel Gasque

2 Jul 2020

Dear Dr Monteiro,

Thank you for submitting your revised Short Report entitled "Thinking fast and simple: how decisions are taken" for publication in PLOS Biology. I have now obtained advice from the original reviewers and have discussed their comments with the Academic Editor.

Based on the reviews, we will probably accept this manuscript for publication, assuming that you will modify the manuscript to address the remaining points raised by reviewer 4. Regarding the title, the editors would like you to consider a more instructive one; one that conveys the central message. We suggest either “Fast decisions are not made at choice time but by ranking alternatives through direct rating” or “Starlings do not make fast decisions at choice time but rank alternatives through direct rating”, but we’d be happy to consider alternatives. We would also like you to consider moving some of the supporting figures to the main text. You have space (up to four main figures), and it would make the data more accesible.

In addition to a clean copy of the manuscript, please also upload a 'track-changes' version of your manuscript that specifies the edits made. This should be uploaded as a "Related" file type.

Please also make sure to address the data and other policy-related requests noted at the end of this email.

We expect to receive your revised manuscript within two weeks.

In addition to the remaining revisions and before we will be able to formally accept your manuscript and consider it "in press", we also need to ensure that your article conforms to our guidelines. A member of our team will be in touch shortly with a set of requests. As we can't proceed until these requirements are met, your swift response will help prevent delays to publication.

*Copyediting*

Upon acceptance of your article, your final files will be copyedited and typeset into the final PDF. While you will have an opportunity to review these files as proofs, PLOS will only permit corrections to spelling or significant scientific errors. Therefore, please take this final revision time to assess and make any remaining major changes to your manuscript.

NOTE: If Supporting Information files are included with your article, note that these are not copyedited and will be published as they are submitted. Please ensure that these files are legible and of high quality (at least 300 dpi) in an easily accessible file format. For this reason, please be aware that any references listed in an SI file will not be indexed. For more information, see our Supporting Information guidelines:

https://journals.plos.org/plosbiology/s/supporting-information

*Published Peer Review History*

Please note that you may have the opportunity to make the peer review history publicly available. The record will include editor decision letters (with reviews) and your responses to reviewer comments. If eligible, we will contact you to opt in or out. Please see here for more details:

https://blogs.plos.org/plos/2019/05/plos-journals-now-open-for-published-peer-review/

*Early Version*

Please note that an uncorrected proof of your manuscript will be published online ahead of the final version, unless you opted out when submitting your manuscript. If, for any reason, you do not want an earlier version of your manuscript published online, uncheck the box. Should you, your institution's press office or the journal office choose to press release your paper, you will automatically be opted out of early publication. We ask that you notify us as soon as possible if you or your institution is planning to press release the article.

*Protocols deposition*

To enhance the reproducibility of your results, we recommend that if applicable you deposit your laboratory protocols in protocols.io, where a protocol can be assigned its own identifier (DOI) such that it can be cited independently in the future. For instructions see: https://journals.plos.org/plosbiology/s/submission-guidelines#loc-materials-and-methods

*Submitting Your Revision*

To submit your revision, please go to https://www.editorialmanager.com/pbiology/ and log in as an Author. Click the link labelled 'Submissions Needing Revision' to find your submission record. Your revised submission must include a cover letter, a Response to Reviewers file that provides a detailed response to the reviewers' comments (if applicable), and a track-changes file indicating any changes that you have made to the manuscript.

Please do not hesitate to contact me should you have any questions.

Sincerely,

Gabriel Gasque, Ph.D.,

Senior Editor

PLOS Biology

------------------------------------------------------------------------

ETHICS STATEMENT:

-- Please include the full name of the IACUC/ethics committee that reviewed and approved the animal care and use protocol/permit/project license. Please also include an approval number.

------------------------------------------------------------------------

DATA POLICY:

-- For figure 1d, please include all replicates AND the way in which the plotted mean and errors were derived (it should not present only the mean/average values).

-- Please also ensure that each figure legend in your manuscript include information on where the underlying data can be found (S1 Data) and ensure your supplemental data file has a legend.

------------------------------------------------------------------------

Reviewer remarks:

Reviewer #1: The authors have addressed appropriately all my comments.

Reviewer #2: I think the edits and additional information provided in this revision make the points much clearer and I recommend publication of this version.

Reviewer #3: I am satisfied by the response and revised manuscript.

Reviewer #4: The authors have been very responsive to my comments and have edited their manuscript thoroughly to address the issues I raised. I find the paper much improved now and think it will be of interest to the field (and understandable also for readers who are not experts in decision theory). Moreover, the claims are formulated much more cautiously now and are supported by the data.

I have two minor comments that the authors could address in the final revision:

1) While it is certainly true that the SCM differs from classical DDMs in terms of its predictions, the same may not be true for the classical race models that are also being used to model value-based choices and the associated confidence in humans. It would be good if the authors could quickly mention this and briefly discuss to what degree their SCM differs from these classical race models.

2) While the manuscript is written much more cautiously now and provides a more balanced perspective, the same cannot be said of its title. "How decisions are taken" implies that the manuscript will describe the one unifying theory of how all decisions are taken, which is clearly not the case. The authors acknowledge this themselves in their responses to my previous comments. I would therefore recommend that the authors change the title to a somewhat less grandiose version that more accurately describes the scope of their manuscript. Perhaps just "Thinking fast and simple" will do and raise the readers interest?

Decision Letter 3

Gabriel Gasque

31 Jul 2020

Dear Dr Monteiro,

On behalf of my colleagues and the Academic Editor, Alexandre Pouget, I am pleased to inform you that we will be delighted to publish your Short Reports in PLOS Biology.

The files will now enter our production system. You will receive a copyedited version of the manuscript, along with your figures for a final review. You will be given two business days to review and approve the copyedit. Then, within a week, you will receive a PDF proof of your typeset article. You will have two days to review the PDF and make any final corrections. If there is a chance that you'll be unavailable during the copy editing/proof review period, please provide us with contact details of one of the other authors whom you nominate to handle these stages on your behalf. This will ensure that any requested corrections reach the production department in time for publication.

Early Version

The version of your manuscript submitted at the copyedit stage will be posted online ahead of the final proof version, unless you have already opted out of the process. The date of the early version will be your article's publication date. The final article will be published to the same URL, and all versions of the paper will be accessible to readers.

PRESS

We frequently collaborate with press offices. If your institution or institutions have a press office, please notify them about your upcoming paper at this point, to enable them to help maximise its impact. If the press office is planning to promote your findings, we would be grateful if they could coordinate with biologypress@plos.org. If you have not yet opted out of the early version process, we ask that you notify us immediately of any press plans so that we may do so on your behalf.

We also ask that you take this opportunity to read our Embargo Policy regarding the discussion, promotion and media coverage of work that is yet to be published by PLOS. As your manuscript is not yet published, it is bound by the conditions of our Embargo Policy. Please be aware that this policy is in place both to ensure that any press coverage of your article is fully substantiated and to provide a direct link between such coverage and the published work. For full details of our Embargo Policy, please visit http://www.plos.org/about/media-inquiries/embargo-policy/.

Thank you again for submitting your manuscript to PLOS Biology and for your support of Open Access publishing. Please do not hesitate to contact me if I can provide any assistance during the production process.

Kind regards,

Vita Usova

Publication Editor,

PLOS Biology

on behalf of

Gabriel Gasque,

Senior Editor

PLOS Biology

Associated Data

    This section collects any data citations, data availability statements, or supplementary materials included in this article.

    Supplementary Materials

    S1 Text. The SCM.

    Computational description, analytical implementation and numerical simulations. SCM, sequential choice model.

    (PDF)

    Click here for additional data file. (373.7KB, pdf)
    S1 Fig

    a. Schematics of the operant panels (adapted with permission from [15], APA). b. Close-up photographs of the operant panels displaying different symbols (top) and colors (bottom).

    (TIF)

    Click here for additional data file. (3.6MB, tif)
    S1 Data. An Excel file with all data used in the main figures, with each tab corresponding to a figure panel, from main Figs 1 to 4, respectively.

    Tabs from the same figure are labelled with the same color.

    (XLSX)

    Click here for additional data file. (36.3KB, xlsx)
    Attachment

    Submitted filename: Response to Reviewers.pdf

    Click here for additional data file. (306.6KB, pdf)
    Attachment

    Submitted filename: Cover letter - revision2.docx

    Click here for additional data file. (14.7KB, docx)

    Data Availability Statement

    All data can be found in the S1 Data file.

    Articles from PLoS Biology are provided here courtesy of PLOS

    RESOURCES