Probing the role of perception in fear generalization

Know Your Movements: Poorer Proprioceptive Accuracy is Associated With Overprotective Avoidance Behavior

Pain-related avoidance of movements that are actually safe (ie, overprotective behavior) plays a key role in chronic pain disability. Avoidance is reinforced through operant learning: after learning that a certain movement elicits pain, movements that prevent pain are more likely to be performed. Proprioceptive accuracy importantly contributes to motor learning and memory. Interestingly, reduced accuracy has been documented in various chronic pain conditions, prompting the question whether this relates to avoidance becoming excessive. Using robotic arm-reaching movements, we tested the hypothesis that poor proprioceptive accuracy is associated with excessive pain-related avoidance in pain-free participants. Participants first performed a task to assess proprioceptive accuracy, followed by an operant avoidance training during which a pain stimulus was presented when they performed one movement trajectory, but not when they performed another trajectory. During a test phase, movements were no longer restricted to 2 trajectories, but participants were instructed to avoid pain. Unbeknownst to the participants, the pain stimulus was never presented during this phase. Results supported our hypothesis. Furthermore, exploratory analyses indicated a reduction in proprioceptive accuracy after avoidance learning, which was associated with excessive avoidance and higher trait fear of pain. PERSPECTIVE: This study is the first to show that poorer proprioceptive accuracy is associated with excessive pain-related avoidance. This finding is especially relevant for chronic pain conditions, as reduced accuracy has been documented in these populations, and points toward the need for research on training accuracy to tackle excessive avoidance.

The association between mnemonic discrimination ability and differential fear learning

BACKGROUND AND OBJECTIVES

It is important to be able to learn which stimuli in our surroundings predict aversive outcomes. To maintain emotional well-being, it is similarly important to be able to learn which stimuli predict safety. The ability to discriminate between stimuli that predict danger and safety has been suggested to not only have an emotional component, but also a cognitive one. One such candidate mechanism is mnemonic discrimination (MD), the ability to differentiate between two memories that are similar but not identical. In the present study, we wanted to examine if MD performance helps to explain inter-individual differences in the ability to acquire a differentiated fear response during fear conditioning.

METHODS

Participants performed a task assessing MD ability, and then underwent a fear conditioning procedure. Fear responses were measured using skin conductance responses (SCRs).

RESULTS

Results revealed no support for MD ability being associated with to which degree a differentiated fear response was acquired, or with the time needed to acquire such a response.

LIMITATIONS

Our only outcome measurement was SCRs. Future studies need to include fear ratings, expectancy ratings and neural responses. Future studies also need to examine this using a stimulus material where the conditioned stimulus and the safety stimulus are more difficult to distinguish from each other.

CONCLUSIONS

If MD ability has a role in inhibiting overgeneralization of fear learning, this does not seem to be driven by MD already during the initial learning.

Cortisol before extinction generalization alters its neural correlates during retrieval

While generalization of fear seems to be naturally acquired as frequently observed in fear-related disorders, extinction learning appears to be stimulus-specific. Thus, treatments aiming to generalize extinction learning comprise the chance to overcome stimulus-specificity and consequently reduce relapse. One suggested candidate is the timing-dependent administration of the stress hormone cortisol. In the present pre-registered, three-day fear conditioning study, we aimed to create a generalized extinction memory trace in 60 healthy men and women using multiple sizes of one conditioned stimulus (CS+G; generalized) during extinction training, whereas the other CS (CS+N; non-generalized) and the CS- were solely presented in their original sizes. Extinction training took place either after pharmacological administration of 20 mg cortisol or placebo. Following successful fear acquisition on day one, generalization effects during extinction training and retrieval were investigated in the comparison of CS+G and CS+N. Insula and dorsal anterior cingulate cortex (dACC) activation for CS+G as compared to CS+N extending to the second half of extinction training indicated prolonged fear processing during extinction training for the CS+G on day two. During retrieval on day three, an activation of the anterior hippocampus occurred for CS+N minus CS+G in the cortisol but not in the placebo group. Additionally, a more posterior hippocampal activation (compared to the other hippocampal activation) was observed for the contrast CS+G minus CS+N. In accordance with our hypotheses, amygdala and dACC responding during reinstatement test was reduced for the CS+G as compared to CS+N. However, cortisol did not modulate amygdala responding, but abolished the CS+G/CS+N differentiation in the dACC relative to placebo. Generalization and cortisol effects were not mirrored in skin conductance responses. In conclusion, extinction generalization processes appear to rely on prolonged fear processing still present in the second half of extinction training that in turn leads to reduced fear-related processing after reinstatement. Cortisol administration prior to extinction training, however, selectively reduced fear-related activation for standard extinction but did not further reduce fear-related activation for extinction generalization.

Stimulus-Based Extinction Generalization: Neural Correlates and Modulation by Cortisol

BACKGROUND

While healthy individuals and patients with anxiety disorders easily generalize fear responses, extinction learning is more stimulus specific. Treatments aiming to generalize extinction learning are urgently needed, since they comprise the potential to overcome stimulus specificity and reduce relapses, particularly in the face of stressful events.

METHODS

In the current 3-day functional magnetic resonance imaging fear conditioning paradigm, we aimed to create a generalized extinction memory trace in 60 healthy men and women by presenting multiple sizes of 1 conditioned stimulus during extinction training (CS+G; generalized), whereas the other conditioned stimulus was solely presented in its original size (CS+N; nongeneralized). Recall was tested on the third day after pharmacological administration of either the stress hormone cortisol or placebo.

RESULTS

After successful fear acquisition, prolonged activation of the amygdala and insula and deactivation of the ventromedial prefrontal cortex for CS+G compared with CS+N during extinction learning indicated sustained fear to the generalization stimuli. In line with our hypotheses, reduced amygdala activation was observed after extinction generalization on the third day in the contrast CS+G minus CS+N, possibly reflecting an attenuated return of fear. Cortisol administration before recall, however, blocked this effect.

CONCLUSIONS

Taken together, the findings show that extinction generalization was associated with decreased activation of the fear network during recall after prolonged activation of the fear network during extinction learning. However, the generalization of the extinction memory did not counteract the detrimental effects of stress hormones on recall. Thus, stimulus-based extinction generalization may not be sufficient to reduce relapses after stressful experiences.

The impact of prior and ongoing threat on the false alarm threshold for facial discrimination

BACKGROUND AND OBJECTIVES

Perceptual adaptations that facilitate rapid responses to threats can also lead to false alarms, or the failure to discriminate safe stimuli from signals of threat. We examined the impact of varying degrees of threat on false alarms in the perceptual discrimination of faces along the dimension of emotion (Experiment 1) or identity (Experiment 2).

METHODS

Participants first trained to discriminate between a target and nontarget face. Next, we tested their ability to identify the target in randomized presentations of the target, the nontarget, and nine novel stimuli morphed in 10% increments of similarity from the target to the nontarget. The task was completed under one of three randomized conditions: 1) Ongoing-Threat paired the target with an aversive outcome in both phases; 2) Prior-Threat paired the target with an aversive outcome in the training phase only; and 3) No-Threat paired the target with a neutral outcome in the training phase only.

RESULTS

In Experiment 1 (N = 90), Ongoing-Threat lowered the false alarm threshold for facial discrimination based on anger intensity compared to Prior-Threat and No-Threat. In Experiment 2 (N = 90), Ongoing-Threat and Prior-Threat each lowered the false alarm threshold for identity-based discrimination compared to No-Threat.

LIMITATIONS

The experiment did not measure generalization of threat responses.

CONCLUSION

Associating a facial expression or identity with threat leads to faster but less accurate discrimination of faces with similar features, particularly under conditions of ongoing threat. These experiments provide an avenue for examining the parameters that impact false alarms, which play a key role in anxiety disorders.

Acute stress leaves fear generalization in healthy individuals intact

Because threatening situations often occur in a similar manner, the generalization of fear to similar situations is adaptive and can avoid harm to the organism. However, the overgeneralization of fear to harmless stimuli is maladaptive and assumed to contribute to anxiety disorders. Thus, elucidating factors that may modulate fear (over)generalization is important. Based on the known effects of acute stress on learning, which are at least partly due to noradrenergic arousal, we investigated whether stress may promote fear overgeneralization and whether we could counteract this effect by reducing noradrenergic arousal. In a placebo-controlled, double-blind, between-subjects design, 120 healthy participants underwent a fear-conditioning procedure on Day 1. Approximately 24 hours later, participants received orally either a placebo or the beta-adrenergic receptor antagonist propranolol and were exposed to a stress or control manipulation before they completed a test of fear generalization. Skin conductance responses as well as explicit rating data showed a successful acquisition of conditioned fear on Day 1 and a pronounced fear generalization 24 hours later. Although physiological data confirmed the successful stress manipulation and reduction of noradrenergic arousal, the extent of fear generalization remained unaffected by stress and propranolol. The absence of a stress effect on fear generalization was confirmed by a second study and a Bayesian analysis across both data sets. Our findings suggest that acute stress leaves fear generalization processes intact, at least in a sample of healthy, young individuals.

Differences in perceptual memory determine generalization patterns

Although memory of past experiences is crucial for the ability to transfer knowledge to new situations, surprisingly little research has directly investigated the relationship between memory and generalization. The present study sought to investigate how the perceptual memory of a trained stimulus influences generalization to similar stimuli. Forty participants underwent a fear conditioning procedure on Day 1, and separate memory recall and generalization tests on Day 2. We focused on two aspects of perceptual memory: namely memory bias (i.e., over- or underestimation of stimulus magnitude) and uncertainty. We found that memory bias predicted the pattern of generalized self-reported (expectancy ratings) and psychophysiological responses (fear-potentiated startle responses). Memory uncertainty was measured in two ways: self-reported uncertainty ratings and variability in stimulus recall. We found that higher levels of self-reported memory uncertainty corresponded with a broader generalization gradient on US expectancy, while greater variability in memory recall was associated with a broader generalization gradient on fear-potentiated startle responses. Taken together, our findings suggest that memory is an important determinant of generalized behavior and illustrate the need to account for these interindividual differences in perceptual memory when examining the generalization of learned responses.

Perceptual variability: Implications for learning and generalization

The generalization of learned behavior has been extensively investigated, but accounting for variance in generalized responding remains a challenge. Based on recent advances, we demonstrate that the inclusion of perceptual measures in generalization research may lead to a better understanding of both intra- and interindividual differences in generalization. We explore various ways through which perceptual variability can influence generalized responding. We investigate its impact on the ability to discriminate between stimuli and how similarity between stimuli may be variable, rather than fixed, because of it. Subsequently, we argue that perceptual variations can yield different learning experiences and that interindividual differences in generalized responding may be understood from this perspective. Finally, we point to the role of memory and decision-making within this context. Throughout this paper, we argue that accounting for perception in current generalization protocols will improve the precision of obtained generalization gradients and the ability to infer latent mechanisms. This can inspire future attempts to use generalization gradients as a (clinical) predictor or to relate them to individual traits and neural correlates and, ultimately, may lead to new theoretical and clinical insights.

Perceptual errors are related to shifts in generalization of conditioned responding

Studies of perceptual generalization have recently demonstrated a close relationship between stimulus perception and conditioned responding, suggesting that incorrect stimulus perception might account for certain characteristics of generalization gradients. In this study, we investigated whether common phenomena, such as the area and peak shift in conditioned responding, relate to perceptual errors. After a differential conditioning procedure, in which one circle was paired with the presentation of an aversive picture whereas a different-sized circle was not, we combined a generalization test with a three-alternative forced-choice perceptual categorization task where participants had to indicate on every trial whether the presented circle was one of the two circles from the conditioning phase or a different one, after which US-expectancy ratings were collected. The typical peak and area shift were observed when conditioned responses were plotted on a physical dimension. However, when stimulus perception was incorporated generalization gradients diverged from the typical gradient. Both the area and peak shift largely disappeared when accounting for perceptual errors. These findings demonstrate the need to incorporate perceptual mechanisms in associative models.