Investigating the decision-making mechanisms underlying choice between drug and nondrug rewards is essential to understand how their alterations can contribute to substance use disorders. Over the past decade, a growing research effort has been made to model this choice situation in animals, particularly in rodents (Lüscher et al., 2020; Ahmed, 2018a; Ahmed, 2018b). Overall, when given a choice, most rats prefer the nondrug alternative (e.g., sweet water, food pellets or social interaction) over potent drugs of abuse, such as cocaine or heroin (Cantin et al., 2010; Lenoir et al., 2013a; Lenoir et al., 2007; Caprioli et al., 2017; Caprioli et al., 2015; Venniro et al., 2018; Vandaele et al., 2016). Only few individual rats prefer the drug. This pattern of individual preferences is observed even after extended drug use and regardless of the drug dose available (Cantin et al., 2010; Lenoir et al., 2007). We recently found that, after some amount of training in our procedure, preference may not necessarily involve a comparison and deliberation over options’ value, but could instead rely on more automatic processes such as habits (Vandaele et al., 2020; Vandaele et al., 2019a). However, investigating habitual control in a drug choice setting is particularly challenging, mainly because there is no effective method of drug reward devaluation in animals, particularly for drugs administered intravenously. An alternative approach, sufficiently versatile to probe the decision-making mechanisms underlying individual preferences in a drug choice setting, is thus needed.

Previous research aimed at testing the validity of different models of animal choice, including behavioral ecology-inspired models, have proposed that the decision-making mechanisms underlying choice should be, in theory, inferable from a detailed analysis of response latencies (Kacelnik et al., 2011; Monteiro et al., 2020; Shapiro et al., 2008; Vasconcelos et al., 2010). One such model, the deliberative choice model (DCM), makes the assumption that deliberation requires time. Indeed, although deliberative behavior is highly flexible in adapting to new environmental contingencies, weighting the pros and cons of multiple choice outcomes to select the most valuable option requires higher cognitive demand and is therefore time-consuming (Keramati et al., 2011; Wickelgren, 1977; Gold and Shadlen, 2007). Specifically, if choice behavior involves a deliberation over the values of the different reward options, one would expect an increase in response latency during choice trials in comparison to sampling trials where the different options are presented separately and successively.

The sequential choice model (SCM) makes a prediction in direct opposition to the DCM. This model explains choice without postulating any explicit deliberation and/or valuation processes. Specifically, when facing different options, animals would consider each option sequentially, not simultaneously, and would decide whether to accept or reject it with no consideration of the other option available. It is hypothesized that such sequential decision process has evolved as an adaptation to the natural ‘reward ecology’ of most animals. Unlike in laboratory choice settings, successive encounters with different rewards are the rule in natural environments while simultaneous encounters are the exception (Charnov, 1976; Lea, 1982). Based on this assumption, the SCM proposes that there would be no genuine decision and that preference would be the result of a race-like competition between independent responses. Briefly, during choice, each option would automatically elicit a specific response with a certain latency: the shorter the latency, the more likely the corresponding response will win the race and thus, the more likely the animal will prefer that option. Importantly, due to this race-like selection process, the SCM uniquely predicts that the choice latencies should be shorter, not longer, than the sampling latencies (see Figure 1—figure supplement 1 for additional information).

Thus, the DCM and SCM models make opposite testable predictions about the differences between choice and sampling latencies (i.e., longer or shorter) (Table 1). Interestingly, however, both models make one general common prediction regarding the relationship between sampling latencies and preference, though for different reasons. Both models predict that individuals that respond faster for one option relative to its alternative during sampling trials should also choose the former more frequently than the later during choice trials. In other terms, there should be a positive correlation between individual sampling latencies and individual preferences during choice. According to the SCM, the nature of this relationship should also be predictive since response latencies play a causal role in the establishment of preference in this model (Shapiro et al., 2008; Vasconcelos et al., 2010). In the DCM, sampling latencies rather represent another measure of the options' relative values.

The goal of the present study is to test these predictions by conducting a systematic retrospective analysis of all cocaine versus saccharin choice experiments that have been conducted in rats in our laboratory over the past 12 years. This allowed us to probe the decision-making mechanisms underlying choice between drug and nondrug rewards as a function of prior training (limited or extended).

The aim of this study was to determine the decision-making processes underlying choice by comparing sampling and choice response latencies, after limited or extended training. An increase in latencies during choice is predicted from the DCM whereas the SCM predicts a shortening of latencies during choice compared to sampling. Here we tested these different predictions in a systematic retrospective analysis of all the choice experiments conducted in the laboratory over the past 12 years. Our analysis shows a lengthening of choice latencies after limited training, suggesting the involvement of a deliberative process. However, when prior training is included before choice testing, we observed similar choice and sampling latencies on saccharin trials and a shortening of choice latencies on cocaine trials. These results are partially consistent with the SCM, but inconsistent with the DCM. Finally, additional analyses in two independent experiments assessing habitual control of choice behavior confirm that the engagement of deliberative or valuation processes can be inferred from the comparison of sampling and choice response latencies.

In agreement with previous experiments (Monteiro et al., 2020; Shapiro et al., 2008; Vasconcelos et al., 2010; Freidin et al., 2009; Vasconcelos et al., 2013; Vasconcelos et al., 2015; Ojeda et al., 2018), results from our analysis validated in both experimental sets (with or without prior training) the general common prediction that the relative response latency during sampling trials should correlate with preference during choice trials. Indeed, the parameters of relative latency were both strongly correlated with the percentage of cocaine choices. Furthermore, in agreement with the SCM, which assumes that the relative sampling latency causally determines preference, sampling latencies could predict whether rats preferred cocaine, saccharin, or were indifferent. However, these findings are not sufficient to favor the SCM over the DCM, which also predicts a relationship between sampling latencies and preference. Indeed, according to the DCM, options are compared and selected based on their relative value (Kacelnik et al., 2011; Rangel et al., 2008; Rangel and Hare, 2010) and numerous studies have reported that latencies are inversely related to options’ relative value (Shapiro et al., 2008; Bateson and Kacelnik, 1995; Killeen and Hall, 2001; Lagorio and Hackenberg, 2012; Shull et al., 1990). More generally, it was suggested that response latencies could be used as a sensitive metric for response strength and subjective value (Shapiro et al., 2008; Bateson and Kacelnik, 1995; Killeen and Hall, 2001; Lagorio and Hackenberg, 2012; Shull et al., 1990). The analysis and comparison of sampling and choice latencies are therefore needed to arbitrate between the DCM and SCM models.

Contrary to all prior studies testing predictions of the SCM (Kacelnik et al., 2011; Monteiro et al., 2020; Shapiro et al., 2008; Vasconcelos et al., 2010; Freidin et al., 2009; Vasconcelos et al., 2013; Vasconcelos et al., 2015; Ojeda et al., 2018; Macías et al., 2021), analysis of experiments without instrumental training prior to choice testing reveals a lengthening of choice latencies compared to sampling latencies. These results are consistent with the DCM which assumes that processing of information during simultaneous encounters would increase the response latency compared to sequential encounters, due to the time cost of the evaluation process (Kacelnik et al., 2011; Keramati et al., 2011; Gold and Shadlen, 2007; Rangel et al., 2008; Rangel and Hare, 2010; Table 1). The discrepancy with prior findings could result from the relatively long training in instrumental tasks involving discrete trial designs, shown to promote rapid habitual learning (Vandaele et al., 2017; Vandaele et al., 2019b), or from species differences (Kacelnik et al., 2011; Monteiro et al., 2020; Shapiro et al., 2008; Vasconcelos et al., 2010; Freidin et al., 2009; Vasconcelos et al., 2013; Vasconcelos et al., 2015; Macías et al., 2021). Indeed, in this analysis, the lengthening of latencies was observed in naïve rats trained for a very limited number of sessions while most studies investigating SCM predictions were conducted in European starlings or pigeons (but see Ojeda et al., 2018), after extensive instrumental training.

In the data set with prior instrumental training, the shortening of cocaine response latencies during choice is consistent with the SCM. Indeed, the race-like competition process implies that only the shortest latencies should be expressed during choice trials, the selection of one option excluding the opportunity to select the alternative option (Figure 1—figure supplement 1). This cross-censorship between distributions of latencies results in a shortening of choice latencies. This prediction was validated for cocaine but not saccharin trials. The negative result for saccharin trials is however consistent with prior studies testing SCM predictions (Monteiro et al., 2020; Shapiro et al., 2008; Vasconcelos et al., 2010; Vasconcelos et al., 2013; Vasconcelos et al., 2015; Ojeda et al., 2018; Macías et al., 2021; Aw et al., 2012). Indeed, none of these studies but one (Ojeda et al., 2018) has reported a shortening of choice latencies for the preferred option. Demonstrating this decrease in latency is difficult because when animals strongly prefer one option, the distribution of latencies for that option is minimally censored (Figure 1—figure supplement 1). Furthermore, the high relative value of this option results in short latencies during sampling, limiting the room for further decrease. Analysis of prior studies conducted in our laboratory indicates that response latencies cannot go below 3 s, likely representing the reaction time.

The floor effect for saccharin response latencies cannot be excluded, but is also likely to occur (Macías et al., 2021). In the absence of floor effect, how to explain the similarity between saccharin sampling and choice latencies while there is a shortening of choice latencies on cocaine trial? A less parsimonious hypothesis would be that a shortening of latencies predicted from the SCM is masked by a lengthening of response latencies due to the involvement of deliberative processes. However, this hypothesis is not supported by findings from our devaluation experiments showing that preference for the nondrug reward rapidly becomes habitual (Vandaele et al., 2020; Vandaele et al., 2019a). In the confirmatory analysis conducted on these experiments, we systematically found similar sampling and choice latencies when responding for the nondrug reward was under habitual control, as indicated by preference insensitivity to devaluation. Although inconsistent with the SCM, this result also invalidates the DCM and confirm that the absence of deliberative or valuation processes can be inferred from the comparison of sampling and choice response latencies. Interestingly, in the second experiment involving choice between water and cocaine, the increased sensitivity to water devaluation was associated with a lengthening of choice latencies compared to sampling latencies. This finding is consistent with the DCM and suggests that across devaluation training, rats are learning to re-engage deliberative processes during choice. Overall, these findings suggest that the DCM and SCM may not be general models of choice, as initially formulated, but could be dynamically engaged to control choice behavior across early and extended training, or following changes in outcomes value or instrumental contingencies.

Although cocaine was the preferred option in CP rats, we were able to observe a shortening of latencies at choice following extended training. Interestingly, this difference was mainly driven by the long sampling latencies since choice latencies were more homogenous between rewards and groups of rats. This finding is in agreement with the SCM, which assumes that while sampling latencies for each option are expressed without cross-censorship from the alternative option, choice latencies are censored by the selection of the alternative option. In CP and IND rats, cocaine sampling latencies are quite long, a finding consistent with prior studies from our lab that remains to be elucidated. However, during choice trials, selection of saccharin censors longer cocaine response latencies, thereby biasing the distribution of cocaine choice latencies in favor of the shortest latencies, more similar to saccharin latencies. Interestingly the gap between cocaine choice and sampling latencies was more pronounced in IND rats compared to CP rats. This result further supports the SCM, which predicts that trimming of the leftward side of the distribution of latencies would be more pronounced when the overlap between response distributions is stronger (Kacelnik et al., 2011; Shapiro et al., 2008; Figure 1—figure supplement 1). Since rats are indifferent, the drug and nondrug rewards are closer in value and selected with more comparable response latencies during sampling. The stronger cross-censorship between distributions of latencies during choice could have increased the shortening of cocaine choice latencies in this subgroup of rats. However, at indifference, the SCM predicts a shortening of choice latencies for both options (Macías et al., 2021). The similarity between saccharin sampling and choice latencies in IND rats is thus unexpected. Given the already low value of latencies during sampling (mean: 4.7 ± 0.49 s; median: 3.97 s), a floor effect cannot be excluded.

Like habit, the automatic selection process of option based on the response latency implied by the SCM is oblivious to the options’ value. The SCM assumes that during sequential encounters, each option is processed independently and choice results from a race between the mechanisms generating response latencies for each available option (Kacelnik et al., 2011). Thus, although the SCM proposes a mechanism for the selection of responses based on their latency, this model makes no assumption on the valuation process generating distributions of latencies. It is tempting to speculate that the habitual system would process the action value (or ‘action policy’) while the SCM offers a mechanism by which habit translates in preference in a choice setting between two simultaneous options. This assumption is in agreement with a recent theory suggesting that decision-making does not necessarily require computation of the economic value by the brain (Hayden and Niv, 2020). Overall, our analysis suggests that after extended training, responses for the drug and the nondrug rewards are selected via a race-like competition mechanism, without deliberation over options’ value, not unlike the habitual preference demonstrated in previous studies (Vandaele et al., 2020; Vandaele et al., 2019a). Numerous studies show a shift in behavioral control from goal-directed to habitual decision-making processes across extended training (Holland, 2004; Adams, 1982; Dickinson and Weiskrantz, 1985). Thus, our results suggest that rats first engage a goal-directed deliberative strategy with computation and comparison of the options value but, after more extended training, adopt a SCM-like response selection process while their behavior becomes habitual. Interestingly, one would expect similar sampling and choice response latencies for a behavior performed automatically under habitual control and without representation of the options’ value. This is in fact what we systematically observed when comparing saccharin latencies in IND and SP rats, after some amount of prior instrumental training. Further research is needed to determine whether comparable sampling and choice latencies could be equated with habit in a choice setting.

A key question raised by our findings is whether the differences between sampling and choice latencies plausibly reflect operation of different decision-making mechanisms (race-like versus deliberative model) or different operations of the same decision-making mechanism. Likewise, it would be interesting to determine whether decision-making mechanisms or operations differ in the groups without and with prior training, when the behavior is presumably goal-directed and habitual, respectively. The interplay between goal-directed and habitual behavior is often described as dichotomous (i.e. based on separate decision-making mechanisms) but there is now a consensus on the need to consider it as a gradient (Schreiner et al., 2020), which may involve a single decision-making mechanism with variations in the underlying operations. The diffusion decision model (DDM) or the linear ballistic accumulator (LBA) are both derived from models of human choice reaction time (Ratcliff et al., 2016; Brown and Heathcote, 2008), and could be applied to our data sets to address this issue. A single decision-making model (LBA or DDM) could explain response latencies in both data sets (with and without prior training) with variations in how the model is operating between sampling and choice and/or between groups, for instance with a change in the starting point of the decision, in the rate of evidence accumulation, or in the threshold at which evidence is translated into choice behavior. Alternatively, behavior in groups with and without prior training could be better explained by different models, indicating that behavior is driven by distinct decision-making mechanisms. Future studies will aim at developing formal modeling of choice behavior to better delineate the decision-making mechanisms or parameters differing between sampling versus choice, limited versus extended training, or as a function of individual preference.

To conclude, this systematic analysis has begun to probe the decision-making processes underlying choice between sweet water and cocaine in rats. It shows that when simultaneously facing these two options, rats first engage a deliberative strategy but after more extended training, rats adopt an automated SCM-like response selection strategy, in the absence of deliberation. Clearly, more research is needed to dissect the precise decision-making mechanisms underlying choice between drug and nondrug rewards in our discrete-trial choice procedure. However, these results suggest that the comparison of sampling and choice latencies could be used to infer the involvement of deliberative processes during choice.

Data tables and Python codes used to generate figures and conduct statistical analysis have been deposited on Open Science Framework: DOI https://doi.org/10.17605/OSF.IO/BFZJY.

1. 1. Vandaele Y
   2. Lenoir M
   3. Vouillac-Mendoza C
   4. Guillem K
   5. Ahmed SH

   (2021) Open Science Framework

   ID 10.17605/OSF.IO/BFZJY. Probing the decision-making mechanisms underlying choice between drug and nondrug rewards in rats.

   https://osf.io/bfzjy/

1. 1. Adams CD

   (1982) Variations in the sensitivity of instrumental responding to reinforcer devaluation

   The Quarterly Journal of Experimental Psychology Section B 34:77–98.

   https://doi.org/10.1080/14640748208400878

   • Google Scholar
2. 1. Ahmed SH

   (2012) The science of making drug-addicted animals

   Neuroscience 211:107–125.

   https://doi.org/10.1016/j.neuroscience.2011.08.014

   • PubMed
   • Google Scholar
3. 1. Ahmed SH

   (2018a) Individual decision-making in the causal pathway to addiction: contributions and limitations of rodent models

   Pharmacology Biochemistry and Behavior 164:22–31.

   https://doi.org/10.1016/j.pbb.2017.07.005

   • PubMed
   • Google Scholar
4. 1. Ahmed SH

   (2018b) Trying to make sense of rodents' drug choice behavior

   Progress in Neuro-Psychopharmacology and Biological Psychiatry 87:3–10.

   https://doi.org/10.1016/j.pnpbp.2017.09.027

   • PubMed
   • Google Scholar
5. 1. Aw J
   2. Monteiro T
   3. Vasconcelos M
   4. Kacelnik A

   (2012) Cognitive mechanisms of risky choice: is there an evaluation cost?

   Behavioural Processes 89:95–103.

   https://doi.org/10.1016/j.beproc.2011.09.007

   • PubMed
   • Google Scholar
6. 1. Bateson M
   2. Kacelnik A

   (1995) Preferences for fixed and variable food sources: variability in amount and delay

   Journal of the Experimental Analysis of Behavior 63:313–329.

   https://doi.org/10.1901/jeab.1995.63-313

   • PubMed
   • Google Scholar
7. 1. Brown SD
   2. Heathcote A

   (2008) The simplest complete model of choice response time: linear ballistic accumulation

   Cognitive Psychology 57:153–178.

   https://doi.org/10.1016/j.cogpsych.2007.12.002

   • PubMed
   • Google Scholar
8. 1. Cantin L
   2. Lenoir M
   3. Augier E
   4. Vanhille N
   5. Dubreucq S
   6. Serre F
   7. Vouillac C
   8. Ahmed SH

   (2010) Cocaine is low on the value ladder of rats: possible evidence for resilience to addiction

   PLOS ONE 5:e11592.

   https://doi.org/10.1371/journal.pone.0011592

   • PubMed
   • Google Scholar
9. 1. Caprioli D
   2. Zeric T
   3. Thorndike EB
   4. Venniro M

   (2015) Persistent palatable food preference in rats with a history of limited and extended access to methamphetamine self-administration

   Addiction Biology 20:913–926.

   https://doi.org/10.1111/adb.12220

   • PubMed
   • Google Scholar
10. 1. Caprioli D
    2. Venniro M
    3. Zhang M
    4. Bossert JM
    5. Warren BL
    6. Hope BT
    7. Shaham Y

    (2017) Role of dorsomedial striatum neuronal ensembles in incubation of methamphetamine craving after voluntary abstinence

    The Journal of Neuroscience 37:1014–1027.

    https://doi.org/10.1523/JNEUROSCI.3091-16.2016

    • PubMed
    • Google Scholar
11. 1. Charnov EL

    (1976) Optimal foraging, the marginal value theorem

    Theoretical Population Biology 9:129–136.

    https://doi.org/10.1016/0040-5809(76)90040-X

    • PubMed
    • Google Scholar
12. 1. Dickinson A
    2. Weiskrantz L

    (1985) Actions and habits: the development of behavioural autonomy

    Philosophical Transactions of the Royal Society B 308:67–78.

    https://doi.org/10.1098/RSTB.1985.0010

    • Google Scholar
13. 1. Freese L
    2. Durand A
    3. Guillem K
    4. Ahmed SH

    (2018) Pre-trial cocaine biases choice toward cocaine through suppression of the nondrug option

    Pharmacology Biochemistry and Behavior 173:65–73.

    https://doi.org/10.1016/j.pbb.2018.07.010

    • PubMed
    • Google Scholar
14. 1. Freidin E
    2. Aw J
    3. Kacelnik A

    (2009) Sequential and simultaneous choices: testing the diet selection and sequential choice models

    Behavioural Processes 80:218–223.

    https://doi.org/10.1016/j.beproc.2008.12.001

    • PubMed
    • Google Scholar
15. 1. Gold JI
    2. Shadlen MN

    (2007) The neural basis of decision making

    Annual Review of Neuroscience 30:535–574.

    https://doi.org/10.1146/annurev.neuro.29.051605.113038

    • PubMed
    • Google Scholar
16. Preprint

    1. Hayden B
    2. Niv Y

    (2020) The case against economic values in the brain

    PsyArXiv.

    https://psyarxiv.com/7hgup/download?format=pdf

    • Google Scholar
17. 1. Holland PC

    (2004) Relations between Pavlovian-Instrumental transfer and reinforcer devaluation

    Journal of Experimental Psychology: Animal Behavior Processes 30:104–117.

    https://doi.org/10.1037/0097-7403.30.2.104

    • Google Scholar
18. 1. Kacelnik A
    2. Vasconcelos M
    3. Monteiro T
    4. Aw J

    (2011) Darwin’s “tug-of-war” vs. starlings’ “horse-racing”: how adaptations for sequential encounters drive simultaneous choice

    Behavioral Ecology and Sociobiology 65:547–558.

    https://doi.org/10.1007/s00265-010-1101-2

    • Google Scholar
19. 1. Keramati M
    2. Dezfouli A
    3. Piray P

    (2011) Speed/accuracy trade-off between the habitual and the goal-directed processes

    PLOS Computational Biology 7:e1002055.

    https://doi.org/10.1371/journal.pcbi.1002055

    • PubMed
    • Google Scholar
20. 1. Killeen PR
    2. Hall SS

    (2001) The principal components of response strength

    Journal of the Experimental Analysis of Behavior 75:111–134.

    https://doi.org/10.1901/jeab.2001.75-111

    • PubMed
    • Google Scholar
21. 1. Lagorio CH
    2. Hackenberg TD

    (2012) Risky choice in pigeons: preference for amount variability using a token-reinforcement system

    Journal of the Experimental Analysis of Behavior 98:139–154.

    https://doi.org/10.1901/jeab.2012.98-139

    • PubMed
    • Google Scholar
22. Book

    1. Lea SEG

    (1982)

    The mechanisms of optimality in foraging

    In: Commons JE, Nevin J. A, Rachlin H, editors. Quantitative Analyses of Behavior. Erlbaum. pp. 169–188.

    • Google Scholar
23. 1. Lenoir M
    2. Serre F
    3. Cantin L
    4. Ahmed SH

    (2007) Intense sweetness surpasses cocaine reward

    PLOS ONE 2:e698.

    https://doi.org/10.1371/journal.pone.0000698

    • PubMed
    • Google Scholar
24. 1. Lenoir M
    2. Cantin L
    3. Vanhille N
    4. Serre F
    5. Ahmed SH

    (2013a) Extended heroin access increases heroin choices over a potent nondrug alternative

    Neuropsychopharmacology 38:1209–1220.

    https://doi.org/10.1038/npp.2013.17

    • PubMed
    • Google Scholar
25. 1. Lenoir M
    2. Augier E
    3. Vouillac C
    4. Ahmed SH

    (2013b) A choice-based screening method for compulsive drug users in rats

    Current Protocols in Neuroscience 9:44.

    https://doi.org/10.1002/0471142301.ns0944s64

    • PubMed
    • Google Scholar
26. 1. Lüscher C
    2. Robbins TW
    3. Everitt BJ

    (2020) The transition to compulsion in addiction

    Nature Reviews Neuroscience 21:247–263.

    https://doi.org/10.1038/s41583-020-0289-z

    • PubMed
    • Google Scholar
27. 1. Macías A
    2. González VV
    3. Machado A
    4. Vasconcelos M

    (2021) The functional equivalence of two variants of the suboptimal choice task: choice proportion and response latency as measures of value

    Animal Cognition 24:85–98.

    https://doi.org/10.1007/s10071-020-01418-8

    • PubMed
    • Google Scholar
28. 1. Monteiro T
    2. Vasconcelos M
    3. Kacelnik A

    (2020) Choosing fast and simply: construction of preferences by starlings through parallel option valuation

    PLOS Biology 18:e3000841.

    https://doi.org/10.1371/journal.pbio.3000841

    • PubMed
    • Google Scholar
29. 1. Ojeda A
    2. Murphy RA
    3. Kacelnik A

    (2018) Paradoxical choice in rats: subjective valuation and mechanism of choice

    Behavioural Processes 152:73–80.

    https://doi.org/10.1016/j.beproc.2018.03.024

    • PubMed
    • Google Scholar
30. 1. Rangel A
    2. Camerer C
    3. Montague PR

    (2008) A framework for studying the neurobiology of value-based decision making

    Nature Reviews Neuroscience 9:545–556.

    https://doi.org/10.1038/nrn2357

    • PubMed
    • Google Scholar
31. 1. Rangel A
    2. Hare T

    (2010) Neural computations associated with goal-directed choice

    Current Opinion in Neurobiology 20:262–270.

    https://doi.org/10.1016/j.conb.2010.03.001

    • PubMed
    • Google Scholar
32. 1. Ratcliff R
    2. Smith PL
    3. Brown SD
    4. McKoon G

    (2016) Diffusion decision model: current issues and history

    Trends in Cognitive Sciences 20:260–281.

    https://doi.org/10.1016/j.tics.2016.01.007

    • PubMed
    • Google Scholar
33. 1. Schreiner DC
    2. Renteria R
    3. Gremel CM

    (2020) Fractionating the all-or-nothing definition of goal-directed and habitual decision-making

    Journal of Neuroscience Research 98:998–1006.

    https://doi.org/10.1002/jnr.24545

    • PubMed
    • Google Scholar
34. 1. Shapiro MS
    2. Siller S
    3. Kacelnik A

    (2008) 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)

    Journal of Experimental Psychology: Animal Behavior Processes 34:75–93.

    https://doi.org/10.1037/0097-7403.34.1.75

    • Google Scholar
35. 1. Shull RL
    2. Mellon RC
    3. Sharp JA

    (1990) Delay and number of food reinforcers: effects on choice and latencies

    Journal of the Experimental Analysis of Behavior 53:235–246.

    https://doi.org/10.1901/jeab.1990.53-235

    • PubMed
    • Google Scholar
36. 1. Vandaele Y
    2. Cantin L
    3. Serre F
    4. Vouillac-Mendoza C
    5. Ahmed SH

    (2016) Choosing under the influence: a Drug-Specific mechanism by which the setting controls drug choices in rats

    Neuropsychopharmacology 41:646–657.

    https://doi.org/10.1038/npp.2015.195

    • PubMed
    • Google Scholar
37. 1. Vandaele Y
    2. Pribut HJ
    3. Janak PH

    (2017) Lever insertion as a salient stimulus promoting insensitivity to outcome devaluation

    Frontiers in Integrative Neuroscience 11:23.

    https://doi.org/10.3389/fnint.2017.00023

    • PubMed
    • Google Scholar
38. 1. Vandaele Y
    2. Vouillac-Mendoza C
    3. Ahmed SH

    (2019a) Inflexible habitual decision-making during choice between cocaine and a nondrug alternative

    Translational Psychiatry 9:109.

    https://doi.org/10.1038/s41398-019-0445-2

    • PubMed
    • Google Scholar
39. 1. Vandaele Y
    2. Mahajan NR
    3. Ottenheimer DJ
    4. Richard JM
    5. Mysore SP
    6. Janak PH

    (2019b) Distinct recruitment of dorsomedial and dorsolateral striatum erodes with extended training

    eLife 8:e49536.

    https://doi.org/10.7554/eLife.49536

    • PubMed
    • Google Scholar
40. 1. Vandaele Y
    2. Guillem K
    3. Ahmed SH

    (2020) Habitual preference for the nondrug reward in a drug choice setting

    Frontiers in Behavioral Neuroscience 14:78.

    https://doi.org/10.3389/fnbeh.2020.00078

    • PubMed
    • Google Scholar
41. 1. Vasconcelos M
    2. Monteiro T
    3. Aw J
    4. Kacelnik A

    (2010) Choice in multi-alternative environments: a trial-by-trial implementation of the sequential choice model

    Behavioural Processes 84:435–439.

    https://doi.org/10.1016/j.beproc.2009.11.010

    • PubMed
    • Google Scholar
42. 1. Vasconcelos M
    2. Monteiro T
    3. Kacelnik A

    (2013) Context-dependent preferences in starlings: linking ecology, foraging and choice

    PLOS ONE 8:e64934.

    https://doi.org/10.1371/journal.pone.0064934

    • PubMed
    • Google Scholar
43. 1. Vasconcelos M
    2. Monteiro T
    3. Kacelnik A

    (2015) Irrational choice and the value of information

    Scientific Reports 5:13874.

    https://doi.org/10.1038/srep13874

    • PubMed
    • Google Scholar
44. 1. Venniro M
    2. Zhang M
    3. Caprioli D
    4. Hoots JK
    5. Golden SA
    6. Heins C
    7. Morales M
    8. Epstein DH
    9. Shaham Y

    (2018) Volitional social interaction prevents drug addiction in rat models

    Nature Neuroscience 21:1520–1529.

    https://doi.org/10.1038/s41593-018-0246-6

    • PubMed
    • Google Scholar
45. 1. Wickelgren WA

    (1977) Speed-accuracy tradeoff and information processing dynamics

    Acta Psychologica 41:67–85.

    https://doi.org/10.1016/0001-6918(77)90012-9

    • Google Scholar

Author details

1. Youna Vandaele

   Lausanne University Hospital, Department of Psychiatry, Prilly, Switzerland

   Contribution

   Conceptualization, Formal analysis, Visualization, Writing - original draft, Writing - review and editing

   For correspondence

   youna.vandaele@chuv.ch

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0002-8389-8850

2. Magalie Lenoir

   1. Université de Bordeaux, Institut des Maladies Neurodégénératives, Bordeaux, France
   2. CNRS, Institut des Maladies Neurodégénératives, Bordeaux, France

   Contribution

   Investigation

   Competing interests

   No competing interests declared

3. Caroline Vouillac-Mendoza

   1. Université de Bordeaux, Institut des Maladies Neurodégénératives, Bordeaux, France
   2. CNRS, Institut des Maladies Neurodégénératives, Bordeaux, France

   Contribution

   Investigation

   Competing interests

   No competing interests declared

4. Karine Guillem

   1. Université de Bordeaux, Institut des Maladies Neurodégénératives, Bordeaux, France
   2. CNRS, Institut des Maladies Neurodégénératives, Bordeaux, France

   Contribution

   Investigation

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0002-9623-3698

5. Serge H Ahmed

   1. Université de Bordeaux, Institut des Maladies Neurodégénératives, Bordeaux, France
   2. CNRS, Institut des Maladies Neurodégénératives, Bordeaux, France

   Contribution

   Conceptualization, Resources, Writing - review and editing

   For correspondence

   serge.ahmed@u-bordeaux.fr

   Competing interests

   No competing interests declared

• 1,509

  views

• 142

  downloads

• 4

  citations

Views, downloads and citations are aggregated across all versions of this paper published by eLife.