Gonzalo Pablo Urcelay

In this selective review, we summarise recent studies from our laboratory in which we have assessed the effect of different variables (such as contiguity and contingency) on the generalization of prior learning. Much of this work was inspired by Pavlov’s work in which it was first documented that trace conditioning can broaden generalization, a finding that we have replicated in human participants. We then describe studies in which we have assessed whether the affective valence of stimuli and outcomes can affect generalization, and we find that indeed human participants can generalize across the valence of stimuli and this can interact with different outcomes. Finally, we review a study in which we assessed whether changes in a response-produced safety signal can influence avoidance and conclude that it does. We also observed that this generalizes and this depends on the similarity between trained and tested safety signals (i.e., outcomes). Overall, we conclude that generalization is a much more complex phenomenon than previously thoughts, and that more research is needed to understand this phenomenon and make it amenable to current theorising.

This study examined whether saline acts as an aversive stimulus for planarians and how it influences learning and memory. Using saline as an unconditioned stimulus, we first tested conditioned place avoidance. Saline did not support conditioned place avoidance acquisition; instead, it disrupted habituation to initially preferred tactile and visual stimuli. Further experiments assessed the establishment of conditioned place preference using food as the reinforcer with visual cues. Following training, exposure to saline impaired the expression of conditioned place preference for both light and dark contexts, indicating that the disruptive effect of exposure to saline extends beyond habituation. These findings suggest that saline interferes with the expression of learned behaviors through state-dependent retrieval failure caused by physiological changes induced by saline exposure. (PsycInfo Database Record (c) 2026 APA, all rights reserved).

A canonical view in the neuroscience of learning and memory literature is that failures in memory expression reflect storage failures, and hence, amnesic manipulations following training or following memory reactivation can permanently erase memory traces. In this review, we analyze extant literatures from the learning and memory domains suggesting that most if not all of these memory deficits can be restored with the appropriate retrieval cues. We contend that all experience-dependent manipulations conducted immediately after training or following memory reactivation result in new learning, which interferes with the original learning and hence makes information highly dependent on retrieval cues for memory expression. Thus, although acquisition and storage mechanisms are surely important, memory retrieval is a critical component of memory performance, with numerous findings from behavioral and neurobiological studies all converging on this general stance. These conclusions invite a rethinking of the learning and memory literatures and provide new avenues for research. (PsycInfo Database Record (c) 2025 APA, all rights reserved).

Uncertainty refers to the state in which the probability of an outcome following an event is unknown. Substantial progress has been made in terms of understanding how uncertainty modulates learning and decision making, however the effect of outcome uncertainty on the generalization of prior knowledge to novel situations has not been clearly documented. In two predictive learning experiments, participants experienced cue-outcome trials in which the probability that an outcome would follow a Gabor grating (vertical or horizontal) was deterministic (100%) or probabilistic (50% or 25%). Participants learned that one Gabor orientation predicted the outcome whereas the other did not. Following training, participants provided expectancy ratings of the outcome to the trained and to novel gratings at different orientations (generalization test). Experiment 1 documented that participants displayed more generalization to novel gratings when training was probabilistic. Experiment 2 replicated these results and further showed broader generalization when participants were trained with deterministic outcomes and then switched to probabilistic in the last training block, a manipulation that increased experienced prediction error and hence surprise. These results suggest that outcome uncertainty broadens generalization gradients, and this is mediated by unsigned prediction errors.

Two online experiments investigated whether the weakening of cue-outcome temporal contiguity impacts generalization gradients. Participants experienced strong temporal contiguity (the offset of a cue coincides with the onset an outcome) or weak temporal contiguity (there is a temporal gap between the offset of a cue and the onset of an outcome) in a continuous predictive learning task. Initial training involved a double negative discrimination, in which there was one reinforced cue (aqua-color rectangle) and two non-reinforced cues (one greener and one bluer than the reinforced cue), and participants had to make a response whenever they predicted the outcome. After training, in the generalization test phase, outcome expectancy in the presence of several stimuli along the full green-blue dimension was tested. Outcome expectancy was assessed with discrete ratings in Experiment 1 ( n = 252) and a behavioral dependent measure (button presses, similar to the training phase) in Experiment 2 ( n = 250). Regardless of the type of measure, when participants experienced weak cue-outcome temporal contiguity they showed a broader generalization response, although differences were more prominent in the green side of the color dimension. Furthermore, no significant differences emerged between the two groups experiencing strong temporal contiguity but differing in cue duration (controlling stimulus onset asynchrony, SOA), suggesting that cue-outcome temporal contiguity plays a more critical role than cue duration in shaping generalization. Overall, this result advances our understanding of the role of temporal contiguity at a basic level, while opening the possibility to study it as a critical factor in more applied settings.

Habits are central to daily life and some compulsive disorders, but reliably measuringthem in the lab is challenging. For the first time, across two experiments, we directlycompared two key experimental paradigms for studying lab-developed habits inhumans, the classic Outcome Devaluation task and a newer Response Remappingtask, by systematically manipulating time constraints for responses. In the outcomedevaluation task, longer preparation time led to more habitual actions,counterintuitively, possibly due to response withholding biases that initially maskedhabitual tendencies. In contrast, the response remapping task showed the expectedpattern – habits dominated under more response time pressure (200–500 ms) andweakened when more time was available. Overall, the response remapping taskelicited stronger habit tendencies than the outcome devaluation task, suggesting thateach paradigm engages distinct cognitive processes –response inhibition versusautomatic action selection– one requiring people to inhibit a learned response, andthe other relying on fast, automatic actions. Our findings reveal that habit expressionis highly sensitive to task structure and underlying mechanisms. This directcomparison underscores the consideration of distinct cognitive and motivationalfactors driving habitual responses and highlights that habits reflect multipledissociable mechanisms rather than a uniform phenomenon.

The developmental trajectory of the generalisation of avoidance has received limited research attention, despite generalisation and avoidance being key features of anxiety disorders. This study used a gamified avoidance task to investigate generalisation in 96 children aged 5–11 years old. The task was an adaption of the classic “Space Invaders” videogame, in which participants have to shoot at spaceships to win points but also avoid being shot by a large spaceship (i.e., aversive outcome). The appearance of the large spaceship (the outcome; which resulted in loss of points) was signalled by coloured sensors at the top of the screen. One sensor predicted the outcome (CS+; an aqua coloured sensor), and a second did not (CS-; orange). Participants could avoid the loss of points by moving into the safe areas of the screen in anticipation of the large spaceship. Following avoidance training, we then presented six generalisation stimuli varying along the blue-green dimension of the CS+, to assess generalisation. Our findings revealed that age significantly influenced generalisation gradients, with younger children (5-8) exhibiting broader gradients compared to older children (9-11). Regression analyses indicated that age, but not anxiety, was a significant predictor of generalisation. These results underscore the impact of ontogenetic changes on the generalisation of avoidance behaviour and highlight the importance of including young children in research to better understand the mechanisms underlying generalization and factors leading to anxiety disorders.

Previous studies have documented that overshadowing weakens when a CS-US trace is inserted, and we have interpreted such effect in terms of differential generalization in trace versus delay procedures, but direct evidence in humans is lacking. Two online experiments investigated whether the weakening of cue-outcome temporal contiguity impacts generalization gradients. Participants experienced strong temporal contiguity (the offset of a cue coincides with the onset an outcome) or weak temporal contiguity (there is a temporal gap between the offset of a cue and the onset of an outcome) in a continuous predictive learning task. Initial training involved a double negative discrimination, in which there was one reinforced cue (aqua-color rectangle) and two non-reinforced cues (one greener and one bluer than the reinforced cue), and participants had to make a response whenever they predicted the outcome. After training, in the generalization test phase, outcome expectancy in presence of several stimuli along the full green-blue dimension was tested. Outcome expectancy was assessed with discrete ratings in Experiment 1 (n=252) and a behavioral dependent measure (button presses, similar to the training phase) in Experiment 2 (n=250). Regardless the type of measure, when participants experienced weak cue-outcome temporal contiguity showed a broader generalization response, although differences were more prominent in the green side of the color dimension. Furthermore, no significant differences emerged between two groups experiencing strong temporal contiguity but differing in cue duration (controlling stimulus onset asynchrony, SOA), suggesting that cue-outcome temporal contiguity plays a more critical role than cue duration in shaping generalization. Overall, this result advances our understanding of the role of temporal contiguity at a basic level, while opening the possibility to study it as a critical factor in more applied settings.

Previous studies involving birds and humans have identified spatial proximity as a source of overshadowing between landmarks in navigation. In Herrera et al. (2024), subjects were trained in an open environment to locate a hidden goal with reference to an array of four landmarks placed at varying distances from it. Critically, two of the four landmarks (i.e., target landmarks) were placed at distances that were common among groups, whereas the remaining two were either proximal to, or distal from, them. Landmarks near the goal overshadowed (i.e., competed with) learning about the further ones, and this effect disappeared in the groups trained with distal landmarks. However, neither of these studies included a control group providing a baseline performance to assess the extent of competition; were the data indicative of overshadowing or facilitation of learning? Thus, we assessed whether spatial proximity determines overshadowing or facilitation between landmarks, including a control group trained with the target landmarks only. We conducted three experiments with varying training lengths: six training trials in Experiments 1 and 16 in Experiments 2 and 3. We also extended the landmark-goal distance in Experiment 3. In all experiments, we observed overshadowing in groups trained with closer nontarget landmarks (relative to the target) but no overshadowing when the nontarget landmarks were distal from the target landmarks. Overall, these experiments reveal that landmark-goal distance determines overshadowing between landmarks, a finding consistent with domain-general theories of learning, such as a modification of Pearce's configural model. (PsycInfo Database Record (c) 2025 APA, all rights reserved).

Goal-directed actions are sensitive to the causal association between actions and outcomes, as well as the value of those outcomes. Such sensitivity diminishes when actions become habitual. Based on recent findings in animals, we tested if delaying outcomes relative to actions would weaken sensitivity to outcome revaluation and reduce action rates. In three experiments (N = 290), participants made fictitious investments in companies within contexts that provided either immediate or delayed feedback. After training, participants were informed of a change in markets which affected both companies (one improved and the other worsened). Across all experiments, action rates were lower in the delayed-feedback condition, and outcome revaluation was stronger in the immediate-feedback condition. In addition, self-reported action-outcome knowledge was weaker in the delayed-feedback condition. These findings suggest that delays in reinforcement weaken the action-outcome association critical for goal-directed control. We discuss the potential mechanisms underlying this phenomenon in light of a contemporary theory of goal-directed behavior.