Overcoming Pavlovian bias in semantic space

Mechanisms underlying food devaluation after response inhibition to food

Multiple studies reveal that a requirement to stop a response to appetitive food stimuli causes devaluation of these stimuli. However, the mechanism underlying food devaluation after stopping is still under debate. The immediate-affect theory suggests that an increase in negative affect after stopping a response is the driving force for food devaluation. A competing value-updating theory presumes that food devaluation after stopping occurs through the need to align behavior with goals. The current study assessed how food devaluation after response inhibition is influenced by negative emotional reactivity and behavior-goal alignment on a trial-by-trial basis. The study included 60 healthy participants who completed a Food-Stop-Signal-Emotion task. Participants categorized high vs. low-calorie food stimuli and stopped their response upon encountering a stop signal. Subsequently, participants made subjective negativity ratings of negative- or neutral-valenced emotional images, and rated their desire to eat the previously depicted food. In contrast to predictions made by the immediate-affect account, food devaluation after stopping was not mediated nor moderated via changes in negative emotional reactivity after stopping. In support of the value-updating account, food devaluation was modulated by behavior-goal alignment, indicated by larger food devaluation after successful vs. failed stopping. In agreement with this theory, the findings indicate that devaluation occurs more strongly when performance aligns with the task requirement. This study sheds light on the mechanism that likely underlies food devaluation after stopping. Implications regarding applied use of food-inhibition trainings are discussed.

Emotional cues reduce Pavlovian interference in feedback-based go and nogo learning

It is easier to execute a response in the promise of a reward and withhold a response in the promise of a punishment than vice versa, due to a conflict between cue-related Pavlovian and outcome-related instrumental action tendencies in the reverse conditions. This robust learning asymmetry in go and nogo learning is referred to as the Pavlovian bias. Interestingly, it is similar to motivational tendencies reported for affective facial expressions, i.e., facilitation of approach to a smile and withdrawal from a frown. The present study investigated whether and how learning from emotional faces instead of abstract stimuli modulates the Pavlovian bias in reinforcement learning. To this end, 137 healthy adult participants performed an orthogonalized Go/Nogo task that fully decoupled action (go/nogo) and outcome valence (win points/avoid losing points). Three groups of participants were tested with either emotional facial cues whose affective valence was either congruent (CON) or incongruent (INC) to the required instrumental response, or with neutral facial cues (NEU). Relative to NEU, the Pavlovian bias was reduced in both CON and INC, though still present under all learning conditions. Importantly, only for CON, the reduction of the Pavlovian bias effect was adaptive by improving learning performance in one of the conflict conditions. In contrast, the reduction of the Pavlovian bias in INC was completely driven by decreased learning performance in non-conflict conditions. These results suggest a potential role of arousal/salience in Pavlovian-instrumental regulation and cue-action congruency in the adaptability of goal-directed behavior. Implications for clinical application are discussed.

The neurocognitive role of working memory load when Pavlovian motivational control affects instrumental learning

Research suggests that a fast, capacity-limited working memory (WM) system and a slow, incremental reinforcement learning (RL) system jointly contribute to instrumental learning. Thus, situations that strain WM resources alter instrumental learning: under WM loads, learning becomes slow and incremental, the reliance on computationally efficient learning increases, and action selection becomes more random. It is also suggested that Pavlovian learning influences people's behavior during instrumental learning by providing hard-wired instinctive responses including approach to reward predictors and avoidance of punishment predictors. However, it remains unknown how constraints on WM resources affect instrumental learning under Pavlovian influence. Thus, we conducted a functional magnetic resonance imaging (fMRI) study (N = 49) in which participants completed an instrumental learning task with Pavlovian-instrumental conflict (the orthogonalized go/no-go task) both with and without extra WM load. Behavioral and computational modeling analyses revealed that WM load reduced the learning rate and increased random choice, without affecting Pavlovian bias. Model-based fMRI analysis revealed that WM load strengthened RPE signaling in the striatum. Moreover, under WM load, the striatum showed weakened connectivity with the ventromedial and dorsolateral prefrontal cortex when computing reward expectations. These results suggest that the limitation of cognitive resources by WM load promotes slow and incremental learning through the weakened cooperation between WM and RL; such limitation also makes action selection more random, but it does not directly affect the balance between instrumental and Pavlovian systems.

Motivational learning biases are differentially modulated by genetic determinants of striatal and prefrontal dopamine function

Dopaminergic neurotransmission plays a pivotal role in appetitively motivated behavior in mammals, including humans. Notably, action and valence are not independent in motivated tasks, and it is particularly difficult for humans to learn the inhibition of an action to obtain a reward. We have previously observed that the carriers of the DRD2/ANKK1 TaqIA A1 allele, that has been associated with reduced striatal dopamine D2 receptor expression, showed a diminished learning performance when required to learn response inhibition to obtain rewards, a finding that was replicated in two independent cohorts. With our present study, we followed two aims: first, we aimed to replicate our finding on the DRD2/ANKK1 TaqIA polymorphism in a third independent cohort (N = 99) and to investigate the nature of the genetic effects more closely using trial-by-trial behavioral analysis and computational modeling in the combined dataset (N = 281). Second, we aimed to assess a potentially modulatory role of prefrontal dopamine availability, using the widely studied COMT Val108/158Met polymorphism as a proxy. We first report a replication of the above mentioned finding. Interestingly, after combining all three cohorts, exploratory analyses regarding the COMT Val108/158Met polymorphism suggest that homozygotes for the Met allele, which has been linked to higher prefrontal dopaminergic tone, show a lower learning bias. Our results corroborate the importance of genetic variability of the dopaminergic system in individual learning differences of action-valence interaction and, furthermore, suggest that motivational learning biases are differentially modulated by genetic determinants of striatal and prefrontal dopamine function.