Julia Wendt

Don’t fear ‘fear conditioning’: Methodological considerations for the design and analysis of studies on human fear acquisition, extinction, and return of fear

HighlightsOriginates from discussions on replicability and researchers degrees of freedom.Aims at stimulating discussions on methods applied in fear conditioning research.Addresses critical issues on terminology, design, methods, analysis.Serves as comprehensive compendium and critical evaluation of read‐out measures.Highlights methodological considerations when studying individual differences. ABSTRACT The so‐called ‘replicability crisis’ has sparked methodological discussions in many areas of science in general, and in psychology in particular. This has led to recent endeavours to promote the transparency, rigour, and ultimately, replicability of research. Originating from this zeitgeist, the challenge to discuss critical issues on terminology, design, methods, and analysis considerations in fear conditioning research is taken up by this work, which involved representatives from fourteen of the major human fear conditioning laboratories in Europe. This compendium is intended to provide a basis for the development of a common procedural and terminology framework for the field of human fear conditioning. Whenever possible, we give general recommendations. When this is not feasible, we provide evidence‐based guidance for methodological decisions on study design, outcome measures, and analyses. Importantly, this work is also intended to raise awareness and initiate discussions on crucial questions with respect to data collection, processing, statistical analyses, the impact of subtle procedural changes, and data reporting specifically tailored to the research on fear conditioning.

Brain activation during anticipation of interoceptive threat

The current study investigated the neural networks activated during the anticipation of potentially threatening body symptoms evoked by a guided hyperventilation task in a group of participants reporting either high or low fear of unexplained somatic sensations. 15 subjects reporting high and 14 subjects reporting low fear of somatic symptoms first learned that one of two cues predicted the occurrence of a hyperventilation task reliably producing body symptoms in all participants that were rated as more intense and unpleasant in the high fear group. During anticipation of unpleasant symptoms, high fear participants reported more intense body symptoms and showed potentiation of the startle reflex. After this learning session, participants were taken into the fMRI where the same cues either predicted the occurrence of hyperventilation or normoventilation, although the task was never performed in the scanner. During anticipation of hyperventilation all participants showed an increased activation of anterior insula/orbitofrontal cortex and rostral parts of the dorsal anterior cingulate cortex/dorsomedial prefrontal cortex (dACC/dmPFC). Brain activation of high compared to low fear participants differed in two ways. First, high fear participants showed an overall stronger activation of this network during threat and safe conditions indexing stronger anxious apprehension during the entire context. Second, while low fear participants no longer responded with stronger activation to the threat cue after experiencing that the hyperventilation challenge did not follow this cue, high fear participants continued to show stronger activation of the network to this cue. Activation of the rostral dACC/dmPFC was significantly correlated with reported fear of somatic symptoms. These data demonstrate that anticipation of interoceptive threat activates the same network that has been found to be active during anticipation of exteroceptive threat cues. Thus, the current paradigm might provide an innovative method to study anxious apprehension and treatment effects in patients with panic disorder.

Prefrontal function associated with impaired emotion recognition in patients with multiple sclerosis

Multiple sclerosis (MS) is associated with the occurrence of white matter plaques in the central nervous system. These are frequently located in areas interconnecting areas associated with the processing of emotions. Although recent behavioral studies indicated social and affective disturbances in many of these patients, functional studies investigating specific emotional recognition in MS are lacking. We used functional magnetic resonance imaging (fMRI) and lesion mapping in MS-patients to investigate correlates between these measures and emotional facial recognition. Eleven patients whose affective ability was impaired were compared with eleven unimpaired MS-patients and eleven healthy controls (HCs) using a facial expression matching task. Decreased recognition performance was limited to the detection of unpleasant facial expressions (sad, fearful, angry). In evaluating the functional activation maps for the unpleasant facial expressions, we found decreased insular and ventrolateral prefrontal cortex (VLPFC) activation in the impaired group versus the unimpaired groups. We found a close relationship between the inability of solving the task and decreased activation of the left VLPFC and the left anterior insula. In addition, we found a correlation between decreased performance accuracy and the presence of lesions in the left temporal white matter. These data suggest that emotion recognition deficits in MS-patients might be due to the interruption of processing emotionally relevant information, which leads to decreased activation of the VLPFC and the insula.

Brain activation during spatial updating and attentive tracking of moving targets

Keeping aware of the locations of objects while one is moving requires the updating of spatial representations. As long as the objects are visible, attentional tracking is sufficient, but knowing where objects out of view went in relation to ones own body involves an updating of spatial working memory. Here, multiple object tracking was employed to study spatial updating and its neural correlates. In a dynamic 3D-scene, targets moved among visually indistinguishable distractors. The targets and distractors either stayed visible during continuous viewpoint changes or they turned invisible. The parametric variation of tracking load revealed load-dependent activations of the intraparietal sulcus, the superior parietal lobule, and the lateral occipital cortex in response to the attentive tracking task. Viewpoint changes with invisible objects that demanded retention and updating produced load-dependent activation only in the precuneus in line with its presumed involvement in updating spatial working memory.

The functional connectivity between amygdala and extrastriate visual cortex activity during emotional picture processing depends on stimulus novelty

Enhanced perceptual processing of emotional stimuli may be accomplished via amygdala-back-projections into the inferior temporal cortex. In the current study, we investigated the influence of stimulus novelty on the covariation between these brain regions during emotional picture processing. Participants viewed repeatedly presented and novel emotional and neutral pictures during fMRI-scanning. The amygdala showed stronger activation to emotional arousing stimuli that decreased rapidly when the same pictures were presented repeatedly. Emotional modulation of the amygdala was reinstated when novel emotional and neutral pictures were presented. Inferior temporal cortex (ITC) showed increased activation during processing of emotional stimuli irrespective of picture repetition. ITC and amygdala activity were highly correlated only during processing of novel emotional pictures. Therefore, we concluded that enhanced perceptual processing of emotional stimuli is triggered by an initial significance detection and corresponding feedback information by the amygdala but is maintained by other mechanisms.

The brain's relevance detection network operates independently of stimulus modality

Brain regions associated with the processing of emotional stimuli are often also associated with the processing of social stimuli. Therefore, this network consisting of the amygdala, the anterior insula, the superior temporal sulcus (STS), and the orbitofrontal cortex (OFC) may rather be involved in more general relevance detection which should be independent of the sensory modality of the stimuli. In the current study, we used functional MRI to measure brain activations while participants either viewed pictures that varied in their emotional and social content or listened to sounds that varied along the same dimensions. The amygdala, the anterior insula, the STS, and the OFC showed increased activation during processing of emotional as well as social stimuli independent of the sensory modality in which the stimuli were presented. Moreover, social emotional stimuli elicited more pronounced activity in this network than stimuli with solely emotional or social content. These results indicate that the proposed network involved in relevance detection works independently of the source of relevance (emotional or social information mediated by the stimulus) and modality.

Fear-potentiated startle processing in humans: Parallel fMRI and orbicularis EMG assessment during cue conditioning and extinction

Studying neural networks and behavioral indices such as potentiated startle responses during fear conditioning has a long tradition in both animal and human research. However, most of the studies in humans do not link startle potentiation and neural activity during fear acquisition and extinction. Therefore, we examined startle blink responses measured with electromyography (EMG) and brain activity measured with functional MRI simultaneously during differential conditioning. Furthermore, we combined these behavioral fear indices with brain network activity by analyzing the brain activity evoked by the startle probe stimulus presented during conditioned visual threat and safety cues as well as in the absence of visual stimulation. In line with previous research, we found a fear-induced potentiation of the startle blink responses when elicited during a conditioned threat stimulus and a rapid decline of amygdala activity after an initial differentiation of threat and safety cues in early acquisition trials. Increased activation during processing of threat cues was also found in the anterior insula, the anterior cingulate cortex (ACC), and the periaqueductal gray (PAG). More importantly, our results depict an increase of brain activity to probes presented during threatening in comparison to safety cues indicating an involvement of the anterior insula, the ACC, the thalamus, and the PAG in fear-potentiated startle processing during early extinction trials. Our study underlines that parallel assessment of fear-potentiated startle in fMRI paradigms can provide a helpful method to investigate common and distinct processing pathways in humans and animals and, thus, contributes to translational research.

Resting heart rate variability is associated with inhibition of conditioned fear.

Startle blink as well as skin conductance responses (SCR) are widely used indices of learning processes associated with fear conditioning and extinction. During safety learning, the amygdala is under top-down inhibitory control by the prefrontal cortex (PFC). The capacity of the PFC to exert inhibitory control over subcortical brain structures may be indexed by resting state vagally mediated heart rate variability (HRV). The present study investigated the association of resting HRV with startle blink and SCR during conditioned fear inhibition and extinction. Participants first learned to discriminate a threat cue (A) signaling an aversive unconditioned stimulus from a safety signal (B), which were each presented together with a third stimulus X (AX+/BX-). Then, both the threat and safety signal were presented together (AB) to test whether the presence of the learned safety signal inhibits the fear response to the danger signal. Finally, AX was presented without reinforcement (AX-) to investigate fear extinction. Higher HRV was associated with pronounced fear inhibition and fear extinction. Resting HRV levels were associated with fear extinction as indexed by startle blink potentiation but not SCR, which presumably reflect more cognitive aspects of learning. Resting HRV may reflect the capacity of the prefrontal cortex to inhibit subcortical fear responses in the presence of safety or when former threat cues are presented in the absence of threat.

Genetic influences on the acquisition and inhibition of fear

As a variant of the Pavlovian fear conditioning paradigm the conditional discrimination design allows for a detailed investigation of fear acquisition and fear inhibition. Measuring fear-potentiated startle responses, we investigated the influence of two genetic polymorphisms (5-HTTLPR and COMT Val(158)Met) on fear acquisition and fear inhibition which are considered to be critical mechanisms for the etiology and maintenance of anxiety disorders. 5-HTTLPR s-allele carriers showed a more stable potentiation of the startle response during fear acquisition. Homozygous COMT Met-allele carriers, which had demonstrated delayed extinction in previous investigations, show deficient fear inhibition in presence of a learned safety signal. Thus, our results provide further evidence that 5-HTTLPR and COMT Val(158)Met genotypes influence the vulnerability for the development of anxiety disorders via different mechanisms.

Mechanisms of change: Effects of repetitive exposure to feared stimuli on the brain's fear network

Repetitive exposure to feared stimuli is considered as the essential element in therapy with phobic patients. However, the mechanisms mediating symptom reduction and their underlying neurobiological processes are poorly understood. Therefore, we presented the same fear-relevant and neutral stimuli repeatedly to individuals with high and low fear of animals during fMRI scanning. High-, but not low-fearful individuals showed an initial fear-stimulus-related potentiation of amygdala and insula activity. Potentiation of the amygdala in the high-fearful group habituated quickly, but insula activity was still potentiated during later repetition trials. Both groups showed an initial potentiation of the dorsomedial prefrontal cortex (dmPFC) that continuously decreased in low-, but not in high-fearful participants. Thus, within-session habituation may occur on an automatic processing level (amygdala), but does not cause lasting neural changes on a higher order cortical level (dmPFC).