Regulation of the exploration-exploitation trade-off captures long-term changes in rat behaviour DOI Open Access
François Cinotti, Étienne Coutureau, Mehdi Khamassi

et al.

Authorea (Authorea), Journal Year: 2023, Volume and Issue: unknown

Published: Oct. 3, 2023

In uncertain environments in which resources fluctuate continuously, animals must permanently decide whether to exploit what they currently believe be their best option, or instead explore potential alternatives case better opportunities are fact available. While such a trade-off has been extensively studied pretrained facing non-stationary decision-making tasks, it is yet unknown how progressively tune while learning the task structure during pretraining. Here, we compared ability of different computational models account for long-term changes behaviour 24 rats learned choose rewarded lever three-armed bandit across days We found that day-by-day evolution rat performance and win-shift tendency revealed progressive stabilization way regulated exploration-exploitation trade-off. successfully captured these behavioural adaptations using meta-learning model controlled by animal’s average reward rate.

Language: Английский

Stochastic decisions support optimal foraging of volatile environments, and are disrupted by anxiety DOI Creative Commons
Alex Lloyd, Ryan McKay, Nicholas Furl

et al.

Cognitive Affective & Behavioral Neuroscience, Journal Year: 2025, Volume and Issue: unknown

Published: Jan. 9, 2025

Abstract Adolescence is a developmental period of relative volatility, where the individual experiences significant changes to their physical and social environment. The ability adapt volatility one’s surroundings an important cognitive ability, particularly while foraging, near-ubiquitous behaviour across animal kingdom. As adolescents experience more in surroundings, we predicted that this age group would be adept than adults at using exploration adjust volatility. We employed foraging task with well-validated computational model characterise mechanisms volatile environments, preregistering hypothesis (aged 16–17; N = 91) exhibit optimal adaptation learning rate environmental compared 24+; 90). However, surprisingly, both exhibited suboptimal adjustment In contrast rate, it was instead participants’ stochasticity (i.e., decision variability) better resembled made by RL agent. Although heightened environment led participants often trial different responses facilitated discovery environment, also found anxiety impaired adaptive ability. finding environments contradicts expectations responsible for successful motivates future work on deleterious role plays when manage periods transition.

Language: Английский

Citations

1

Reviewing explore/exploit decision-making as a transdiagnostic target for psychosis, depression, and anxiety DOI Creative Commons
Alex Lloyd, Jonathan P. Roiser, Sarah Skeen

et al.

Cognitive Affective & Behavioral Neuroscience, Journal Year: 2024, Volume and Issue: 24(5), P. 793 - 815

Published: April 23, 2024

In many everyday decisions, individuals choose between trialling something novel or they know well. Deciding when to try a new option stick with an that is already known you, as the "explore/exploit" dilemma, important feature of cognition characterises range decision-making contexts encountered by humans. Recent evidence has suggested preferences in explore/exploit biases are associated psychopathology, although this typically been examined within individual disorders. The current review whether represents promising transdiagnostic target for psychosis, depression, and anxiety. A systematic search academic databases was conducted, yielding total 29 studies. Studies examining psychosis were mostly consistent showing explored more compared without psychosis. literature on anxiety depression heterogenous; some studies found exploration, whereas other demonstrated reduced exploration depression. However, subset employed case-control methods, there both also increased exploration. Due heterogeneity across literature, we suggest insufficient conclude alongside our advisory groups lived experience advisors, context candidate merits further investigation using well-powered, longitudinal designs. Such work should examine choices amenable intervention.

Language: Английский

Citations

8

One factor to bind them all: visual foraging organization to predict patch leaving behavior with ROC curves DOI Creative Commons
Marcos Bella-Fernández,

Manuel Suero Suñé,

Alicia Ferrer-Mendieta

et al.

Cognitive Research Principles and Implications, Journal Year: 2025, Volume and Issue: 10(1)

Published: April 5, 2025

Abstract Predicting quitting rules is critical in visual search: Did I search enough for a cancer nodule breast X-ray or threat baggage airport scanner? This study examines the predictive power of organization indexes like best-r, mean ITD, PAO, intersection rates as optimal criteria to leave foraging (looking several targets among distractors). In sample 29 adults, we compared static and dynamic foraging. Also, reanalyze data from diverse tasks lifespan already published replicate results. Using ROC curves, all results consistently show that measures outperform classic intake commonly used animal models predict human behavior. Importantly, combination traditional within unitary factor best predictor. Our findings open new research line studying decisions based on organization.

Language: Английский

Citations

0

Understanding explore-exploit dynamics in child development: current insights and future directions DOI Creative Commons
Seokyung Kim, Stephanie M. Carlson

Frontiers in Developmental Psychology, Journal Year: 2024, Volume and Issue: 2

Published: Sept. 23, 2024

Examining children's decisions to explore or exploit the environment provides a window into their developing metacognition and reflection capacities. Reinforcement learning, characterized by balance between exploring new options (exploration) utilizing known ones (exploitation), is central this discussion. Children initially exhibit broad intensive exploration, which gradually shifts toward exploitation as they grow. We review major theories empirical findings, highlighting two main exploration strategies: random directed. The former involves stochastic choices without considering information rewards, while latter driven reducing uncertainty for gain. Behavioral tasks such n-armed bandit, horizon, patch foraging are used study these strategies. Findings on bandit horizon showed mixed results whether decreases over time. Directed consistently with age, but its emergence depends task difficulty. In patch-foraging tasks, adults tend overexploit (staying too long in one patch) children overexplore (leaving early), whereas adolescents display most optimal balance. paper also addresses open questions regarding mechanisms supporting early application of strategies real-life contexts like persistence. Future research should further investigate relation cognitive control, executive function metacognition, explore-exploit strategies, examine practical implications adaptive learning decision-making children.

Language: Английский

Citations

3

The connecting brain in context: How adolescent plasticity supports experiential learning and development DOI Creative Commons
Amanda E. Baker, Adriana Gálvan, Andrew J. Fuligni

et al.

Developmental Cognitive Neuroscience, Journal Year: 2024, Volume and Issue: 71, P. 101486 - 101486

Published: Nov. 28, 2024

Language: Английский

Citations

3

The explore/exploit trade‐off: An ecologically valid and translational framework that can advance mechanistic understanding of eating disorders DOI
Kelsey E. Hagan,

Ivieosa Aimufua,

Ann F. Haynos

et al.

International Journal of Eating Disorders, Journal Year: 2024, Volume and Issue: 57(5), P. 1102 - 1108

Published: Feb. 22, 2024

Abstract The explore/exploit trade‐off is a decision‐making process that conserved across species and balances exploring unfamiliar choices of unknown value with choosing familiar options known to maximize reward. This framework rooted in behavioral ecology has traditionally been used study maladaptive versus adaptive non‐human animal foraging behavior. Researchers have begun recognize the potential utility understanding human psychopathology through trade‐off. In this article, we propose holds promise for advancing our mechanistic processes confer vulnerability maintain eating pathology due its neurodevelopmental bases, conservation species, ability be mathematically modeled. We present model how suboptimal can promote disordered recommendations future research applying pathology. Taken together, provides translational expanding etiologic maintenance models pathology, given developmental changes coincide time emergence evidence biased psychopathology. Additionally, disorders may improve knowledge their underlying pathophysiology, informing targeted clinical interventions such as neuromodulation pharmacotherapy. Public Significance Statement cross‐species whereby organisms choose between option reward or sampling options. hypothesize imbalanced preliminary data. significant advance neurocognitive mechanisms which could ultimately guide revisions inform novel interventions.

Language: Английский

Citations

1

Research on re-searching: interrupted foraging is not disrupted foraging DOI Creative Commons
Injae Hong, Jeremy M. Wolfe

Cognitive Research Principles and Implications, Journal Year: 2024, Volume and Issue: 9(1)

Published: May 15, 2024

Abstract In classic visual search, observers typically search for the presence of a target in scene or display. foraging tasks, there may be multiple targets same display (or “patch”). Observers and collect these items one patch until they decide to leave that move next one. This is highly rule-governed behavior. The current study investigated whether rules are disrupted when interrupted various manners. Experiment 1, was briefly then resumed patch. Experiments 2 3, each either ended voluntarily compulsorily after fixed amount time. cases, only all patches were visited. Overall, remained largely intact, though shows can overridden by demand characteristics task. results show participants tended perform approximately consistently despite interruptions. suggest behavior relatively simple environment resilient not easily interruption.

Language: Английский

Citations

1

Regulation of reinforcement learning parameters captures long‐term changes in rat behaviour DOI Creative Commons
François Cinotti, Étienne Coutureau, Mehdi Khamassi

et al.

European Journal of Neuroscience, Journal Year: 2024, Volume and Issue: 60(4), P. 4469 - 4490

Published: June 24, 2024

In uncertain environments in which resources fluctuate continuously, animals must permanently decide whether to stabilise learning and exploit what they currently believe be their best option, or instead explore potential alternatives learn fast from new observations. While such a trade-off has been extensively studied pretrained facing non-stationary decision-making tasks, it is yet unknown how progressively tune while the task structure during pretraining. Here, we compared ability of different computational models account for long-term changes behaviour 24 rats learned choose rewarded lever three-armed bandit across days We found that day-by-day evolution rat performance win-shift tendency revealed progressive stabilisation way regulated reinforcement parameters. successfully captured these behavioural adaptations using meta-learning model either rate inverse temperature was controlled by average reward rate.

Language: Английский

Citations

1

Applications of flexible ureteroscope lithotripsy with negative pressure suction and percutaneous nephrolithotomy in the treatment of 2–3 cm diameter renal calculi DOI

Juan Qi,

Dawei Ni,

Jiamin Shen

et al.

Clinical and Experimental Nephrology, Journal Year: 2024, Volume and Issue: unknown

Published: Nov. 14, 2024

Language: Английский

Citations

1

Developmental differences in children and adults’ enforcement of explore versus exploit search strategies in the United States and Turkey DOI
Hilal H. Şen, Sarah L. Kiefer, Ece Aksu

et al.

Developmental Science, Journal Year: 2024, Volume and Issue: 27(5)

Published: April 25, 2024

Abstract Across development, as children acquire a deeper understanding of their environment, they explore less and take advantage, or “exploit,” what already know. Here, we test whether also enforce exploration‐oriented search behaviors onto others. Specifically, ask are more likely to encourage agent versus exploit this pattern varies across childhood (between 3 6 years). We differs between adults, generalizes two different sociocultural contexts—Turkey the United States—that differ on dimensions that might relate children's decisions about exploration (e.g., curiosity‐focused educational practices, attitudes toward uncertainty avoidance). Participants (N = 358) watched an for rewards were asked at various points should “stay” (exploit) in current location, “go” (explore) new location. At all experiment, enforced significantly often than adults. Early agent's search, US Turkey; later younger (from both contexts) continue enforcing compared older children. These findings highlight not only highly exploratory themselves, but others—underscoring central role plays driving early cognitive development diverse contexts. Research Highlights The study examined developmental cross‐cultural differences adults’ enforcement explore‐exploit strategies. Children Turkey who exploitation often; results generally consistent cultures with small differences. Mirroring changes own behavior; tendency decreased 3‐ 6‐years age. Findings underscore “exploration mindset” decision‐making—even when has no direct benefits child themselves.

Language: Английский

Citations

0