Error Correction Dynamics in Rats’ Waiting Behaviors
Journal
Timing & Time Perception
ISSN
2213-4468
Publisher
Walter de Gruyter GmbH
Date Issued
2025-12-05
Author(s)
Pedro Espinosa-Villafranca
Frida García-Rangel
Emmanuel Alcalá
Valeria Pérez-Treviño
Jonathan Buriticá
Pablo Saavedra
Type
text::journal::journal article
Abstract
<jats:title>Abstract</jats:title>
<jats:p>Agents organize their behavior based on temporal regularities. However, interval timing is far from perfectly calibrated. Each time an agent undertakes a timing attempt, it may deviate from an ideal temporal state, producing durations that fall either below or above it. To adjust behavior adaptively, agents are presumed to correct their timing by monitoring both current and past deviations from this ideal state. We tested this assumption using a differential-reinforcement-of-low-rates (DRL) task with rats, where rewards are given for waiting a minimum (target) time before a response. Our results show that subjects are relatively successful in sustaining waiting times above the target interval after two consecutive rewarded attempts. However, they tend to struggle to break out of a streak of premature responses following consecutive failures. Interestingly, after a win–fail sequence, subjects frequently correct the error, while a fail–win sequence tends to impede sustained waiting above the target time. Some of these behavioral dynamics predict individual differences in global indices, like mean waiting time (a proxy for inhibitory control), the spread of waiting times (a proxy for timing precision), and the number of rewards obtained. Nevertheless, these dynamics do not account for session-by-session performance fluctuations within subjects. Our findings that rats adjusted their timing performance based on previous trials demonstrated operationally defined error correction, providing implicit support for error monitoring. These trial-by-trial adjustments challenge theories that view timing attempts as randomly sampled from stationary distributions; instead, distributions of waiting times may emerge from continuous corrections of deviations from an ideal state. In addition, neither error correction nor adaptive adjustment was universally present in our data, suggesting that mechanisms beyond temporal error monitoring and reward maximization contribute to timing performance.</jats:p>
<jats:p>Agents organize their behavior based on temporal regularities. However, interval timing is far from perfectly calibrated. Each time an agent undertakes a timing attempt, it may deviate from an ideal temporal state, producing durations that fall either below or above it. To adjust behavior adaptively, agents are presumed to correct their timing by monitoring both current and past deviations from this ideal state. We tested this assumption using a differential-reinforcement-of-low-rates (DRL) task with rats, where rewards are given for waiting a minimum (target) time before a response. Our results show that subjects are relatively successful in sustaining waiting times above the target interval after two consecutive rewarded attempts. However, they tend to struggle to break out of a streak of premature responses following consecutive failures. Interestingly, after a win–fail sequence, subjects frequently correct the error, while a fail–win sequence tends to impede sustained waiting above the target time. Some of these behavioral dynamics predict individual differences in global indices, like mean waiting time (a proxy for inhibitory control), the spread of waiting times (a proxy for timing precision), and the number of rewards obtained. Nevertheless, these dynamics do not account for session-by-session performance fluctuations within subjects. Our findings that rats adjusted their timing performance based on previous trials demonstrated operationally defined error correction, providing implicit support for error monitoring. These trial-by-trial adjustments challenge theories that view timing attempts as randomly sampled from stationary distributions; instead, distributions of waiting times may emerge from continuous corrections of deviations from an ideal state. In addition, neither error correction nor adaptive adjustment was universally present in our data, suggesting that mechanisms beyond temporal error monitoring and reward maximization contribute to timing performance.</jats:p>