Search

Allogeneic cell therapy as a means to break immunotolerance to solid tumors is increasingly used for cancer treatment. To investigate cellular alloimmune responses in a human tumor model, primary cultures were established from renal cell carcinoma (RCC) tissues of 56 patients. In three patients with stable RCC line and human leukocyte antigen (HLA)-identical sibling donor available, allogeneic and autologous RCC reactivities were compared using mixed lymphocyte/tumor cell cultures (MLTC). Responding lymphocytes were exclusively CD8+ T cells, whereas CD4+ T cells or natural killer cells were never observed. Sibling MLTC populations showed higher proliferative and cytolytic antitumor responses compared with their autologous counterparts. The allo-MLTC responders originated from the CD8+ CD62L(high)+ peripheral blood subpopulation containing naive precursor and central memory T cells. Limiting dilution cloning failed to establish CTL clones from autologous MLTCs or tumor-infiltrating lymphocytes. In contrast, a broad panel of RCC-reactive CTL clones was expanded from each allogeneic MLTC. These sibling CTL clones either recognized exclusively the original RCC tumor line or cross-reacted with nonmalignant kidney cells of patient origin. A minority of CTL clones also recognized patient-derived hematopoietic cells or other allogeneic tumor targets. The MHC-restricting alleles for RCC-reactive sibling CTL clones included HLA-A2, HLA-A3, HLA-A11, HLA-A24, and HLA-B7. In one sibling donor-RCC pair, strongly proliferative CD3+CD16+CD57+ CTL clones with non-HLA-restricted antitumor reactivity were established. Our results show superior tumor-reactive CD8 responses of matched allogeneic compared with autologous T cells. These data encourage the generation of antitumor T-cell products from HLA-identical siblings and their potential use in adoptive immunotherapy of metastatic RCC patients. (Cancer Res 2006; 66(23): 11447-54)

Supplementary Figure Legends 1-2 from Superior Antitumor In vitro Responses of Allogeneic Matched Sibling Compared with Autologous Patient CD8+ T Cells

Supplementary Figure 2 from Superior Antitumor In vitro Responses of Allogeneic Matched Sibling Compared with Autologous Patient CD8+ T Cells

Avoidance behavior is a core symptom of anxiety disorders that may hinder adaptation. Anxiety disorders are heterogeneous and previous research suggests to decompose anxiety into two dimensions: anxious apprehension and anxious arousal. How these two dimensions are associated with avoidance of and exposure to threatening stimuli, as well as their accompanying neural processes, is barely understood. We examined threat processing using event-related potentials (N1, LPP) from 134 individuals considering the influence of anxiety dimensions. During a two-phase picture-viewing task the participants watched neutral and threatening pictures, which they were instructed to either avoid or attend to during repeated presentations. Results showed that threatening compared to neutral pictures were associated with increased attention allocation (N1) and in-depth processing (LPP), modulated by task-instructions (lower during avoidance). Further, increased anxious apprehension was associated with heightened automatic attention (increased N1), followed by reduced LPP amplitudes for threatening pictures suggesting reduced in-depth processing. During re-exposure, threatening pictures were associated with increased in-depth processing, with no difference between previously avoided and maintained pictures. Together, these results illustrate that avoidance and high anxious apprehension seem to lead to similar neural changes in the processing of aversive images that may conflict with long-term adaptation.

We wanted to investigate what influence acute stress has on later fear generalization and specifically if noradrenergic arousal is necessary for the possible stress effects to evolve.

be characterized by early and fast orientation towards the stimulus, followed by a subsequent phase of attentional disengagement or avoidance (Mogg, Bradley, Miles, & Dixon, 2004; Mogg, Mathews, & Weinman, 1987). Patients suffering from anxiety disorders even show facilitated rapid attention towards threatening stimuli (Bar-Haim, Lamy, Pergamin, Bakermans-Kranenburg, & van, 2007; Bradley, Mogg, Falla, & Hamilton, 1998; Broadbent & Broadbent, 1988; Cisler & Koster, 2010; Mogg & Bradley, 2018), followed by decreased later processing (Mueller et al., 2009; Muhlberger et al., 2009; Weinberg & Hajcak, 2011a). Rapidly attending to threatening stimuli followed by a decreased in-depth processing, may hinder updating information about threatening stimuli, e.g., learning that threat is no longer existent (Salters-Pedneault, Tull, & Roemer, 2004). Specifically, it is suggested that avoiding the confrontation with threatening stimuli play a key role in the development and maintenance of anxiety disorders (Foa & Kozak, 1986). To target avoidance behavior, exposure therapy has proven to be very effective (Foa & Kozak, 1985; Hofmann & Smits, 2008). Yet, many patients do not respond to exposure therapy (Taylor, Abramowitz, & McKay, 2012), which highlights the need for investigating possible factors that explain this variation. It has been shown that preventing in-depth processing during threat experience inhibits habituation to those stimuli and can lead to an overshoot in reaction during subsequent re-exposure (Paul, Kathmann, & Riesel, 2016). Following up on this result it has been shown that this effect is especially increased in high worry participants (Kausche et al., in prep). It should be noted that the focus in clinical research, shifts from categorical characterizations of anxiety disorders to dimensional approaches (Sharp, Miller, & Heller, 2015), i.e., along mainly two latent dimensions: anxious arousal (i.e. hypervigilance and hyperarousal) and anxious apprehension (i.e. repetitive thoughts and worry). Following this dimensional approach, the Research Domain Criteria Initiative (RDoC) was established aiming at defining disorders according to psychophysiological and biological targets (Cuthbert & Kozak, 2013; Insel et al., 2010). Following this approach, anxiety disorder categories can be differentiated as either high on both dimensions (e.g. social anxiety disorder, SAD), high on anxious arousal and low on apprehension (e.g. panic disorder or specific phobia; Lang, McTeague, & Bradley, 2016), or low on anxious arousal and high on anxious apprehension (e.g. obsessive-compulsive disorder (OCD) and generalized anxiety disorder (GAD); Cox, Clara, Hills, & Sareen, 2010; Krueger, 1999). Thus, investigating attentional processes of threatening stimuli, focusing on habituation processes, has the potential to help us to explain why some patients do respond to exposure therapy and others do not (Taylor et al., 2012). Neural correlates of attentional processing of stimuli can be measured by use of event-related potentials (ERPs) using electroencephalography (EEG). Up to date, research found several ERPs, which are sensitive to the processing of threat stimuli. Early and rapid orienting towards threatening stimuli can be indexed by visual-evoked potentials (VEPs; Olofsson, Nodin, Sequeira, & Polich, 2008). The P1, for example, peaking between 100 and 130 ms after picture onset is larger for threatening compared to neutral (Holmes, Nielsen, & Green, 2008; Mueller et al., 2009; Muhlberger et al., 2009) and even positive pictures (Carretie, Hinojosa, Martin-Loeches, Mercado, & Tapia, 2004; Smith, Cacioppo, Larsen, & Chartrand, 2003). Furthermore, the P1 is differently be modulated by anxiety, as studies showed that high-trait individuals (Holmes et al., 2008), SAD patients (Mueller et al., 2009), as well as phobic patients (Michalowski et al., 2009) show an even enhanced P1, whereas a study with OCD patients did not reveal a difference between patients and healthy controls (HCs; Liu, Zakrzewski, Mathews, & Keil, 2021). In addition, studies suggest that the early peaking N1 (around 140-200 ms), which is thought to reflect a sensory gating mechanism of attention (Luck, Woodman, & Vogel, 2000; Vogel & Luck, 2000), shows a potentiated response in high anxious individuals To investigate more elaborated stimulus processing i.e., at a later stage of the neural processing cascade, the late positive potential (LPP) can be used (Cuthbert, Schupp, Bradley, Birbaumer, & Lang, 2000; Hajcak & Olvet, 2008; Paul, Kathmann, et al., 2016; Weinberg & Hajcak, 2011a, 2011b). The LPP occurs approximately 300 ms after stimulus onset and continues for the whole duration of stimulus presentation (Foti, Hajcak, & Dien, 2009; Schupp et al., 2000). It is sensitive for the emotional content of stimuli (Codispoti, Ferrari, & Bradley, 2006), modulated by trait anxiety (Holmes et al., 2008; Muhlberger et al., 2009), and can be influenced by emotion regulation strategies (Hajcak & Nieuwenhuis, 2006; Paul, Kathmann, et al., 2016). Across varies anxiety disorder, it was found that the LPP is enhanced for disorder specific pictures in OCD patients (Paul, Simon, Endrass, & Kathmann, 2016), phobic patients (Leutgeb, Schafer, & Schienle, 2009; Michalowski et al., 2009), and SAD patients (Moser, Huppert, Duval, & Simons, 2008; Muhlberger et al., 2009). In addition, GAD patients have been characterized by increased LPPs to threatening pictures (MacNamara & Hajcak, 2010; but see Weinberg & Hajcak, 2011a). This renders the LPP a candidate ERP to investigate information processing and potential habituation processes. In addition to ERPs elicited by threatening stimuli, the stimulus preceding negativity (SPN) rather represents anticipatory attention and peaks shortly before the onset of a stimulus (Bocker, Baas, Kenemans, & Verbaten, 2004; Bocker, Baas, Kenemans, & Verbaten, 2001). Studies showed an increased SPN for individuals showing fear of negative evaluation (Van der Molen et al., 2013), and a blunted SPN for high worry individuals (Grant, Judah, White, & Mills, 2015; Tanovic, Pruessner, & Joormann, 2018). However, in our unpublished data we found an increased SPN for high worry participants (Kausche et al., in prep). Because of these conflicting findings, we will not have a specific hypothesis regarding the SPN but rather investigate it in an exploratory manner.

Adequate fear learning is an important requirement to protect an individual from threatening or dangerous stimuli or situations. Key processes are the acquisition of fear, fear generalization, and fear extinction learning (Lonsdorf et al., 2017). Fear acquisition is characterized by associative, often pavlovian, learning processes (Pavlov, 1927). Central to this process is the formation of a fear association of a neutral stimulus (NS) with a threatening outcome (unconditioned stimulus, US). Through iterative pairing with the US, the NS becomes a conditioned stimulus (CS; Lissek et al., 2005; Ohman, Erikkson, & Olofsson, 1975). Fear generalization describes the process of applying acquired fear associations and responses to similar stimuli, without having to learn a new fear association (Dymond, Dunsmoor, Vervliet, Roche, & Hermans, 2015; Lissek et al., 2008). In contrast, fear extinction describes an inhibitory learning process, which involves learning that a previous with fear associated CS does not signal danger or threat anymore (Bouton, 1993; Craske, Treanor, Conway, Zbozinek, & Vervliet, 2014; Hermans, Craske, Mineka, & Lovibond, 2006). Some people however, show persistent or inflexible responses to threatening stimuli, leading to maladaptive behavior, which is thought to be a strong contributor to the development and maintenance of anxiety and stress-related disorders (Duits et al., 2015; Lissek et al., 2005). The strong symptom overlap and the concurrent heterogeneity in the response to e.g. treatment options between anxiety disorders, has led to a more transdiagnostic research approach over the last decade (Insel et al., 2010). As such, results suggest to characterize anxiety disorders along two latent dimensions: anxious arousal and anxious apprehension (Clark & Watson, 1991; Cox, Clara, Hills, & Sareen, 2010; Krueger, 1999; Sharp, Miller, & Heller, 2015). Anxious arousal is thought to represent a rather physiological component, associated with hypervigilance and hyperarousal, whereas anxious apprehension reflects a more cognitive component associated with negative and repetitive thoughts and worry. Following this approach, anxiety disorder categories can be differentiated as either high on both components (e.g. social anxiety disorder, SAD), high on anxious arousal and low on apprehension (e.g. panic disorder or specific phobia; Lang, McTeague, & Bradley, 2016), or vice versa (e.g. obsessive-compulsive disorder (OCD) and generalized anxiety disorder (GAD); Cox et al., 2010; Krueger, 1999). Meta-analytic evidence on fear conditioning processes in anxiety disorders reveals that anxiety patients compared to healthy controls (HC) showed increased fear responses to conditioned safety cues during fear acquisition, and reduced fear extinction (Duits et al., 2015). Because there were – aside from post-traumatic stress disorder (PTSD) – not enough studies of other anxiety disorders included, this meta-analysis could not differentiate this effect between different anxiety disorders. However, subsequent studies suggest that anxiety disorders that vary along anxiety dimensions show different outcomes, e.g. in fear acquisition. As such, OCD as well as SAD patients show differential fear conditioning, i.e. an increased response to the CS+ compared to the CS- (Ahrens et al., 2016; Geller et al., 2017). However, OCDs compared to HCs show a stronger fear response to the CS+ (Geller et al., 2017), whereas SAD patients compared to HCs show larger ratings to both the CS+ and the CS- (Ahrens et al., 2016). Such an investigation might help to explain why not all patients respond to exposure therapy (Taylor, Abramowitz, & McKay, 2012), although being an effective treatment for most anxiety patients (Foa & Kozak, 1985; Hofmann & Smits, 2008). To this end, we plan to examine different fear and safety learning processes in anxiety disorders that might differ along the anxiety dimensions. We will focus on explicit rating data, fear potentiated startle (FPS), and electrocortical fear responses. The startle reflex can be defined as highly preserved, defensive reaction (Landis & Hunt, 1939) and is increased in the presence of threat (Davis & Astrachan, 1978). Electrocortical fear responses can be measured by event-related potentials (ERPs). Fear-relevant compared to fear-irrelevant stimuli attract more attention (Ohman, Flykt, & Esteves, 2001) and it has been suggested that anxiety disorders even show a hypervigilance towards those fear-relevant stimuli (Broadbent & Broadbent, 1988; Cisler & Koster, 2010; Grupe & Nitschke, 2013; Mogg & Bradley, 2018). Specifically, it has been shown that emotional compared to neutral stimuli are associated with increased automatic processing, as indicated by a larger P1 (Holmes, Nielsen, & Green, 2008; Muhlberger et al., 2009). Interestingly, the P1 can be modulated by anxiety, as studies showed that high-trait individuals (Holmes et al., 2008), SAD patients (Mueller et al., 2009), as well as phobic patients (Michalowski et al., 2009) show an enhanced P1 to aversive pictures, whereas a study with OCD patients did not reveal a difference between patients and healthy controls (HCs; Liu, Zakrzewski, Mathews, & Keil, 2021). In addition, to the P1, the N1 represents an interesting ERP, indicating fast automatic processing (Luck, Woodman, & Vogel, 2000; Vogel & Luck, 2000), which seems to be potentiated in high anxious individuals (Holmes et al., 2008; Mueller et al., 2009; Yang et al., 2020). At later stages, emotional stimuli are associated with more in-depth processing, indicated by the late-positive potential (LPP; Cuthbert, Schupp, Bradley, Birbaumer, & Lang, 2000; Hajcak, Weinberg, MacNamara, & Foti, 2012). Across varies anxiety disorders, there often was an increased LPP for stimuli associated with threat for anxious participants (Leutgeb, Schafer, & Schienle, 2009; MacNamara & Hajcak, 2010; Michalowski et al., 2009; Moser, Huppert, Duval, & Simons, 2008; Muhlberger et al., 2009; Paul, Simon, Endrass, & Kathmann, 2016; but see Weinberg & Hajcak, 2011a).

Anxiety disorders are highly prevalent (Baxter et al., 2013), debilitating, and account for an immense burden of disease within mental disorders as measured by disability-adjusted life years (Baxter et al., 2014; Murray et al., 2012). We thus need to understand the neural underpinnings of different phenotypical expressions of anxiety and—on this basis—establish tailored interventions. In other words, research is required to bridge the gap between basic science and clinical practice to enable the application of precision medicine principles in psychotherapy. Following this perspective, the Research Domain Criteria initiative (RDoC) promotes research that integrates fundamental biobehavioral findings to understand the mechanisms that underlie mental disorders (Cuthbert & Insel, 2013). This includes research on genetic, molecular, cell, neural circuitry, and physiological levels as well as behavioral, self-report and paradigm associations. If steady findings across these units of analysis emerge, the findings could be used to develop etiology-based diagnostics and promote mechanism-based interventions (Cuthbert & Insel, 2013). Prior studies often limited their focus on single units of analysis (e.g., self-reports) or compared patients with a specific disorder category (e.g., generalized anxiety disorder; GAD) to a healthy control sample, resulting in two major limitations: First, it is unclear, whether a certain mechanism is altered across multiple units of analysis. For instance, a certain effect might be present in experimental data but not in self-report data. These blind spots can be reduced by intersecting the examination of specific mechanisms thought to underlie psychopathology across units of analysis. Second, if specific disorder categories are compared to healthy control samples to understand etiologic factors, the researcher must presuppose coherent and homogenous disorder categories (cf. Krueger & Markon, 2006). Consequently, a non-comorbid sample is required to fit into a narrow diagnostic scheme, i.e., the study object (disorder category) serves the method (diagnostic) and not vice versa, reducing the generalizability of findings. An exception of this is Lang et al. (2016), who performed a re-analysis of physiological data in patients across the anxiety disorder spectrum. Lang et al. phenotyped anxiety disorders on a spectrum from circumscribed fear (e.g., specific phobia) along a generalization gradient (across e.g., circumscribed and generalized social anxiety disorder; SAD) to anxious misery (e.g., obsessive-compulsive disorder; OCD and GAD). With increasing comorbidity, functional impairment, and chronicity, an anxiety disorder phenotype moves towards the pole anxious misery. A composite index of defensive preparedness (i.e., startle), sympathetic (i.e., SCR) and more parasympathetic (i.e., heart rate) activation revealed that this dimension was traced in physiological reactivity: While circumscribed fear coincided with physiological hyperreactivity, anxious misery concurred with physiological hyporeactivity. However, the data originated from studies using a disorder–idiographic aversive imaginary paradigm (McTeague & Lang, 2012) and the uncertainty of those threat narratives was not directly manipulated. Neurobiology-informed models suggest that uncertainty plays a central role in anxiety disorders: Patients with anxiety disorders display increased threat expectancies (Grupe & Nitschke, 2013) and show an increased intolerance of uncertainty (IU; Boelen & Reijntjes, 2009; Carleton et al., 2013; Gentes & Ruscio, 2011; Holaway et al., 2006; Mahoney & McEvoy, 2012; McEvoy et al., 2019; Tolin et al., 2003). To experimentally manipulate uncertainty, Schmitz & Grillon (2012) developed an instructed threat of shock paradigm, the NPU-threat test. Here, the degree of certainty varies directly, which distinguishes responses to predictable from responses to unpredictable threat. Specifically, in the predictable (P) condition, the subjects receive a highly predictable aversive stimulus whose occurrence is signaled by a cue. In contrast, subjects receive unsignaled aversive stimuli in the unpredictable (U) condition, whereas in the neutral (N) condition, subjects never receive an aversive stimulus. The potentiation of physiological responses (typically startle reflex) from condition N to U is labelled anxiety potentiation, the potentiation from N to P is labelled fear potentiation (Schmitz & Grillon, 2012). A wealth of studies directly manipulated temporal uncertainty via the NPU-threat test in clinical samples (e.g., Gorka et al., 2017; Grillon et al., 2017; Klahn et al., 2017; Lieberman et al., 2017; Stevens et al., 2018). One study employing magnetoencephalography (Klahn et al., 2017) found that patients with panic disorder displayed stronger ventromedial prefrontal cortex activation in threat and safety conditions compared to patients with specific phobia and healthy controls. In line with the uncertainty and anticipation model of anxiety (Grupe & Nitschke, 2013) this suggests deficient neural discrimination between threat and safety in individuals with more generalized anxiety (i.e., panic disorder) compared to circumscribed fear (i.e., specific phobia). Grillon et al. (2017) compared patients with—partially overlapping—diagnoses of GAD and SAD. A subset of these patients had a history of panic attacks. While SAD increased fear potentiated startle, a history of panic attacks rendered patients hypersensitive to unpredictable threat. In effect, the presence of panic attacks might indicate greater symptom severity and dimensionally position the affected closer to the pole anxious misery. This indicates an increasing sensitivity to unpredictable threat along the dimension of circumscribed fear to anxious misery (Lang et al., 2016). While some studies support the association of panic disorder increasing responses to unpredictable threat (Grillon et al., 2008; Shankman et al., 2013), other studies found contradictory response patterns (Gorka et al., 2014; Stevens et al., 2018). Specifically, Gorka et al. (2017) found—antithetical to Grillon et al. (2017)—that individuals with more circumscribed fears (i.e., SAD and specific phobia) showed elevated anxiety potentiated startle compared with individuals with anxious misery (i.e., GAD). These inconsistent findings from group comparisons of anxiety disorders call for more granular research that includes dimensional constructs thought to underlie anxiety psychopathology. Essentially, broad and heterogeneous disorder categories might overshadow a possible hypersensitivity to unpredictable threat in more severe and generalized anxiety psychopathology or in specific subtypes of patients characterized by a certain symptom or trait. In fact, the association between panic disorder and hypersensitivity to unpredictable threat has been shown to rely on IU (Gorka et al., 2014; Nelson et al., 2016). Yet, studies simultaneously assessing the impact of categorical and dimensional psychopathology on threat anticipation constitute a gap in the clinical research. As such, a dimensional and commonly shared decomposition of anxiety proposes two anxiety dimensions, which are considered psychologically and neurally distinguishable (Cox et al., 2010; Heller et al., 1997; Lang et al., 2016; Sharp et al., 2015): First, anxious apprehension grasps the tendency for verbal, future-oriented rumination (which can be conceptualized as a behavioural consequence of high IU; Einstein, 2014). Second, anxious arousal captures the propensity to experience symptoms of physiological hyperarousal when confronted with mild stressors (Clark & Watson, 1991; Sharp et al., 2015; Watson et al., 1995). While the evidence for anxious apprehension modulating responses to temporal uncertain threat is scarce and inconsistent (Nelson & Shankman, 2011; Rutherford et al., 2020), self-report anxious arousal has not yet been linked to threat anticipation. Usually, questionnaires assess the delineated constructs dimensionally, resembling the RDoC zeitgeist (Cuthbert & Insel, 2013). However, the diagnostic systems ICD and DSM used in clinical practice adopt a categorical approach for the description of psychopathology (American Psychiatric Association, 2013; World Health Organization, 2020). Thus, in order to span inferences from fundamental science to the clinical practice, the present project takes into account (1) clinical spectra of dimensionally assessed symptomatology and aims to compare (2) categorical and dimensional approaches describing anxiety psychopathology. Moreover, while panic disorder and GAD predominate in previous clinical research using the NPU-threat test (e.g., Gorka et al., 2013; Grillon et al., 2017), further anxiety disorders have not been focused on. As such, OCD has—to the best of our knowledge—not yet been examined in this context which constitutes a second gap in the literature. To address these gaps in the literature, the present project intends to examine the influence of anxiety disorder categories ranging from circumscribed fear to anxious misery on threat anticipation and incorporate dimensional constructs that underlie phenotypic expressions of anxiety symptomatology. Importantly, the symptom severity within a disorder category might vary within disorder categories and will be examined exploratorily. Previous studies used the startle reflex to assess defensive responding (Blumenthal et al., 2005). Auditory startle probes trigger the startle reflex. The event-related potential (ERP) component N100 locked to startle probes emanates from early sensory processing these startle probes, the P300 reflects attentional allocation to startle probes at a later stage of neural processing. In a more aversive context, a general neural response pattern involves increased startle (Schmitz & Grillon, 2012), increased N100 (for unpredictable threat; Nelson, Hajcak, et al., 2015) and decreased P300 (Ferry & Nelson, 2020; MacNamara & Barley, 2018; Nelson et al., 2015). In this, the P300 represents the captivation of attentional resources to the threat cue instead of the startle probe.

Fear responses are typically not limited to the actual threatening stimulus but generalize to other stimuli resembling the threatening stimulus. Although this fear generalization is generally adaptive, fear overgeneralization is maladaptive and assumed to contribute to anxiety disorders. Despite the clinical relevance of fear (over)generalization, how the extent of fear generalization is modulated remains not well understood. Based on the known effects of stress on learning and memory, we tested here the impact of major stress mediators, glucocorticoids and noradrenergic arousal, on fear generalization. In a laboratory-based, placebo-controlled, double-blind, between-subject design, 125 healthy participants first underwent a fear conditioning procedure. About 24 h later, participants received orally either a placebo, hydrocortisone, the α2-adrenoceptor antagonist yohimbine, leading to increased noradrenergic stimulation, or both drugs before a test of fear generalization. Skin conductance responses as well as explicit rating data revealed that yohimbine intake led to enhanced fear memory expression, i.e. an enhanced responding to the CS+ but not to stimuli resembling the CS+. Moreover, neither enhanced safety learning nor a mere enhancement of perceptual discrimination ability could explain this result. In contrast to yohimbine, hydrocortisone had no significant effect on fear memory. These findings suggest that noradrenergic arousal strengthens fear memory expression and have important implications for mental disorders in which the overgeneralization of conditioned fear is prominent.