Reducing fear overgeneralization in children using a novel perceptual discrimination task

Highlights

• A novel perceptual discrimination training task was assessed in a child sample.

• The training group exhibited improved basic perceptual discrimination abilities.

• Less overgeneralization of fear was observed in the training group.

Abstract

Fear generalization, while adaptive, can be detrimental when occurring in excess. To this end a perceptual discrimination training task was created with a goal of decreasing fear overgeneralization. The current study tested the effectiveness of the training task among typically-developing children.

Participants (n = 73) were randomly assigned into a training, placebo or no task group. Following a differential fear-conditioning task, participants in the first two groups underwent the discrimination training or placebo task. An assessment task was then administered. Finally, all participants completed a fear generalization test, consisting of 11 morphs ranging in perceptual similarity from the threat cue to the safety cue. Physiological and self-report measures were collected.

Fear-conditioning was achieved in both physiological and self-report measures. Further, in the assessment task, the training group showed better perceptual discrimination than did the placebo group. Last, the training group exhibited less overgeneralization of affective stimuli as indicated by a physiological measure than did the two control conditions.

Findings suggest that the perceptual discrimination training task effectively moderated fear overgeneralization in children. This adds to previous evidence of the task's effectiveness among adults.

Introduction

Generalization is an adaptive mechanism whereby prior knowledge of a specific stimulus transfers onto a novel stimulus. However, excessive generalization may be maladaptive. Such is the case with fear overgeneralization, the process through which fearful characteristics are attributed to a novel, usually neutral, stimulus based on a previous association with a similar but dangerous stimulus. This often occurs when the two stimuli are perceived to share perceptual characteristics (see Goldstone, Medin, & Schyns, 1997). As fear overgeneralization of innocuous stimuli can lead to avoidance and subsequent functional impairment, recent attempts have been made to decrease fear overgeneralization in adults through perceptual discrimination training tasks (Ginat-Frolich, Klein, Katz, & Shechner, 2017; Lommen et al., 2017). However, no such attempts have been made with youth. Therefore, the aim of the current study is to examine the impact of a perceptual discrimination training task on fear generalization in a sample of typically-developing children.

The past decade has seen a surge in research exploring the mechanism of fear generalization in human adults. This is typically modeled by creating a continuum from a threat to a safety cue using either gradients of perceptual similarity (e.g., color, shape; for a review, see Dymond, Dunsmoor, Vervliet, Roche, & Hermans, 2015) or categorical belonging (e.g., mammals, vehicles; for a review, see Dunsmoor & Murphy, 2015). Participants’ responses are measured either by self-report or physiological indices and plotted to form a curve representing the generalization gradient. Observed differences in the generalization gradient are then used to quantify fear overgeneralization, which is indexed by heightened levels of fear responding to stimuli on this continuum.

Fear overgeneralization can occur only after an initial fear has been established in connection with a specific stimulus. Differential fear conditioning tasks are often used to model this association in the laboratory. During fear acquisition, a neutral stimulus (conditional stimulus, CS) is repeatedly paired with an aversive stimulus (unconditional stimulus, US). Through this pairing, a new association is formed between the CS and US, resulting in the CS eliciting a fearful response in the absence of the US (threat cue; CS+). A second neutral stimulus is repeatedly presented alone, thereby signifying a lack of danger (safety cue; CS-). In contrast to studies with animal models where only physiological and behavioral responses are assessed, in human research, subjective ratings of fear are also collected. The use of multiple measures during both differential fear conditioning and generalization tasks is imperative, because physiological and self-reported indices assess different aspects of the fear response. Ultimately, the two indices, while different, are complementary.

Compared to the abundant differential fear conditioning studies conducted with adults or animals across development, research with human children is scarce, largely because of the complex methodological and ethical considerations related to fear learning research in developmental populations (see Shechner, Hong, Britton, Pine, & Fox, 2014). Further, the majority of developmental studies have focused on fear acquisition and fear extinction. Other fear-related mechanisms, such as fear generalization, have received less attention. Specifically, to the best of our knowledge, only two studies have examined fear generalization in typically-developing children (Glenn et al., 2012; Schiele et al., 2016), and one has done so with a clinical sample (El-Bar, Laufer, Yoran-Hegesh, & Paz, 2017). Additional studies have looked at generalization of extinction learning through threat/safety discrimination in both typically-developing children (Michalska, Shechner, Britton, Pine, & Fox, 2016) and clinical populations (Britton, 2013). Importantly, all of the above studies assessed perceptual fear generalization. Although it is difficult to directly compare across these studies, and further research is needed, it can be reasonably argued that fear overgeneralization occurs in children.

As the overgeneralization of fear is clearly maladaptive, recent attempts have been made to reduce it. To this end, a previous study with healthy adults assessed a novel perceptual discrimination training task aiming to increase basic perceptual discrimination and, as a result, improve affective learning (Ginat-Frolich et al., 2017). Following a differential conditioning task, participants were randomized to complete either the discrimination training or a placebo task. Thereafter, participants completed a discrimination assessment task, followed by a fear generalization test, consisting of 11 stimuli ranging in perceptual similarity from the CS + to the CS-. Results showed that the training group displayed better perceptual discrimination in the assessment task than did the placebo group. More importantly, this improvement transferred onto the affective stimuli in the fear generalization test, with participants in the training group showing less fear overgeneralization than did the placebo group.

Expecting similar effects in typically-developing children is reasonable for three reasons: first, the acquisition of fear through differential learning tasks has been found effective in studies with children (for review, see Shechner et al., 2014); second, children are able to distinguish between stimuli based on perceptual dimensions such as size and shape (see Goldstone, Medin, & Shyns, 1997); and third, children exhibit fear overgeneralization. The current study therefore sought to examine the effects of a perceptual discrimination training task on fear generalization, previously used with healthy adults, in a sample of healthy children. Given the limited number of studies assessing fear generalization in children, a third experimental condition was added. This “no task” group, who completed neither the perceptual discrimination training or placebo task nor the assessment task, served as an additional control by providing a picture of children's fear generalization patterns in the absence of an intervention.

The study has three primary hypotheses. First, children will achieve successful differential conditioning as measured by both self-reported fear and physiological indices. Second, following perceptual discrimination training, children will be better at discriminating between abstract shapes than will a placebo group. Third, improvement in perceptual discrimination will translate into enhanced discrimination of affective stimuli, ranging in perceptual similarity from a safety cue to a threat cue, resulting in decreased fear generalization in the training group.