Common and separable behavioral and neural mechanisms underlie the generalization of fear and disgust

This is a self-archived version of an original article. This version may differ from the original in pagination and typographic details. Author(s): Title: Year: Version: Copyright: Rights: Rights url: Please cite the original version: CC BY-NC-ND 4.0 https://creativecommons.org/licenses/by-nc-nd/4.0/ Common and separable behavioral and neural mechanisms underlie the generalization of fear and disgust © 2022 Elsevier Inc. All rights reserved. Accepted version (Final draft) Wang, Jinxia; Sun, Xiaoying; Becker, Benjamin; Lei, Yi Wang, J., Sun, X., Becker, B., & Lei, Y. (2022). Common and separable behavioral and neural mechanisms underlie the generalization of fear and disgust. Progress in NeuroPsychopharmacology and Biological Psychiatry, 116, Article 110519. https://doi.org/10.1016/j.pnpbp.2022.110519 2022 1 1 Common and separable behavioral and neural mechanisms underlie the 2 generalization of fear and disgust 3 4 5 Jinxia Wang1,2, Xiaoying Sun3, Benjamin Becker4*, Yi Lei1* 6 7 1Institute for Brain and Psychological Sciences, Sichuan Normal University, Chengdu 610066, China 8 2 Faculty of Education and Psychology, University of Jyvaskyla, Finland 9 3Ningxia College of Construction, Ningxia 750021, China 10 4Clinical Hospital of Chengdu Brain Science Institute, MOE Key Laboratory for Neuroinformation, 11 University of Electronic Science and Technology of China, Chengdu, China 12 13 14 Corresponding author: Yi Lei; Benjamin Becker 15 Email: [email protected]; [email protected] 16 17 18 19 20 21 22 23 24 25 26 27 2 Abstract 28 Generalization represents the transfer of a conditioned responses to stimuli that 29 resemble the conditioned stimulus (CS). Previous studies on generalization of defensive 30 avoidance responses have primarily focused on fear and have neglected disgust 31 generalization, which represents a key pathological mechanism in some anxiety 32 disorders. In the present study we examined common and distinct mechanisms of fear 33 and disgust generalization by means of a fear or disgust multi-CS conditioning and 34 generalization paradigm with concomitant event-related potential (ERPs) acquisition in 35 n = 62 subjects. We demonstrate that compared to fear, disgust-relevant generalized 36 stimuli (GS) elicited larger expectancy ratings and longer reaction times (RTs) 37 reflecting stronger ratings of ‘risk’. On the electrophysiological level, increased P2 38 amplitudes were found in response to conditioned CS+ versus CS− across both 39 domains, possibly reflecting higher motivational and attentional salience of aversive 40 conditioned stimuli per se. Contingent negative variation (CNV) amplitude was 41 significantly larger for disgust-CS+ than disgust-CS−, showing stronger preparation of 42 the disgust US. Additionally, we found that the contingent negative variation (CNV) 43 fear generalization gradient, and CNV amplitude were increased with similarity to CS+. 44 In contrast the CNV to disgust-GS did not differ and did not reflect disgust 45 generalization. Together this may indicate that the CNV represents a highly fear46 specific index for generalization learning. This study provides the first neurobiological 47 evidence for common and distinct generalization learning in fear versus disgust 48 suggesting that dysregulations in separable defensive avoidance mechanisms may 49 underly different anxiety disorder subtypes. 50 Keywords: Multi-conditioned stimulus conditioning; fear; disgust; event-related 51 potentials; defensive responses 52 3 1. Introduction 53 Appropriate fear generalization represents an evolutionarily adaptive defensive 54 mechanism allowing organisms to respond immediately to and avoid future potential 55 dangers (Arnaudova et al., 2017). However, fine-grained balance between 56 generalization and discrimination is vital for the organism to distinguish between safety 57 threat signals in order to facilitate adaptive behavior in an ever changing environment 58 (Sangha et al., 2020). The vast majority of previous studies on the underlying defensive 59 and learning mechanisms have employed classical Pavlovian fear conditioning 60 paradigms, during which repeated pairing with an aversive stimulus (Unconditioned 61 Stimulus: US), renders an initially neutral stimulus (Conditioned Stimulus: CS+) or 62 similar stimuli (Generalized Stimulus: GS) that resemble the original CS, as a trigger 63 for the fear response (Conditioned Response: CR) (Yau & McNally, 2018). 64 The Pavlovian fear conditioning paradigm has been widely employed to examine 65 mechanistic dysregulations in anxiety-related disorders, characterized by over66 generalization, impaired extinction, and excessive avoidance (Duits et al., 2015; Pittig 67 et al., 2018). Specifically, while no discrimination difference (CS+ minus CS−) was 68 observed doing conditioning, anxiety patients exhibited stronger expression of fear to 69 the CS− (safety signal; predicting the absence of an aversive US), which may reflect an 70 over generalization to a safety cue or deficient fear inhibition to the safety signal (Lissek 71 et al., 2013; Jovanovic et al., 2010). Anxious individuals have shown decreased 72 ventromedial prefrontal cortex engagement during both, conditioning and extinction 73 recall indicating dysregulated safety and fear learning (Marin et al., 2017). Furthermore, 74 patients with generalized anxiety disorders exhibit a shallower generalization gradient 75 suggesting that an overgeneralization of fear to safe stimuli may contribute to the 76 development and maintenance of pathological anxiety (Lissek et al., 2014). 77 4 While a large body of research has investigated the important role of fear in 78 anxiety disorders, accumulating evidence suggests that disgust-related mechanisms 79 may also contribute to psychopathological dysregulations (e.g., Armstrong & Olatunji, 80 2017; Cisler et al., 2009; Ludvik et al., 2015). Both, fear and disgust represent adaptive 81 defensive-avoidance mechanisms which have evolved to avoid potential threats in 82 terms of predators or contaminations, respectively (Woody & Teachman, 2000). 83 Individuals with high disgust proneness are more susceptible to developing 84 dysregulated avoidance responses in terms of contamination-associated obsessive85 compulsive disorder (OCD), blood-injection-injury phobia, and small animal phobias 86 (e.g., Bhikram et al., 2017; Cougle et al., 2016; Hirai et al., 2018; Olatunji et al., 2017). 87 Woody et al. (2005) moreover demonstrated that disgust plays an important role in 88 avoidance symptoms in spider phobias such that individuals with high fear experienced 89 both, stronger anxiety and disgust as compared to individuals with low fear. Although 90 OCD and post-traumatic stress disorders (PTSD) were removed from anxiety disorder 91 category in the DSM-5, both conditions are closely linked to exaggerated fear and 92 disgust reactivity (McGuire et al., 2016). In parallel to studies examining dysregulations 93 in fear learning, Pavlovian disgust conditioning models have been successfully applied 94 to determine disgust-associated pathomechanisms in contamination-based OCD (Stein 95 et al., 2001) as well as PTSD (Badour et al., 2013). Furthermore, accumulating 96 evidence suggests a direct association between symptoms of contamination-based OCD 97 and disgust sensitivity (Olatunji et al., 2010), and – in contrast to fear – acquired disgust 98 responses are highly resistant to extinction as indexed by subjective experience as well 99 as behavioral indices (Mason, & Richardson, 2010). 100 Further evidence for distinct yet also interacting mechanisms underlying fear and 101 disgust learning comes is provided by developmental studies reporting that children 102 5 experienced increased disgust after vicarious fear learning by presenting novel animals 103 (CSs) with adult faces expressing fear (USs) as well as increased fear experience after 104 vicarious disgust learning (Askew et al., 2014). Klucken et al. (2012) investigated the 105 neural basis of fearand disgust-conditioning and demonstrated that both aversive 106 learning mechanisms involved common neural circuits encompassing the occipital 107 cortex, the nucleus accumbens, the orbitofrontal cortex, and the dorsal anterior 108 cingulate cortex, with higher disgust sensitivity being associated with increased insula 109 activation. However, common and distinct generalization gradients and underlying 110 differentiable electrophysiological responses during fearand disgust-generalization 111 have not been systematically examined. 112 An increasing number of recent studies examined the temporal dynamics of fear 113 conditioning by means of electrophysiological approaches such as event-related 114 potentials and demonstrated that early attention components, including P2 and P3, 115 showed enhanced amplitude in response to CS+ compared to CS− (Junghöfer et al., 116 2015; Junghöfer et al., 2017; Sperl et al, 2021). Further, studies focusing on the late 117 positive potential (LPP) component suggest a sustained attention to CS+ probably 118 representing the newly acquired fear (Pavlov & Kotchoubey, 2019; Ventura-Bort et al., 119 2016). Krusemark and Li (2011), employed visual fear and disgusting stimuli of natural 120 objects in a visual search paradigm with concomitant event-related potential (P1) 121 acquisition, contrasting the effects of the two defensive responses on early neural 122 indices of sensory perception and attention. The results showed that, compared to 123 neutral stimuli, fear images elicited a larger P1 (96 ms) amplitude whereas disgust 124 images evoked an attenuated P1 amplitude, demonstrating an opposite pattern of early 125 sensory discrimination. Despite these initial findings on differential early perceptual 126 discrimination the common and separable ERP-responses underlying generalization 127 6 may further allow to determine the process-specific contribution of neurobiological 128 separable fear and disgust mechanisms to segregate separable psychopathological 129 markers for fear and disgust-related anxiety disorders. 130 Against this background the present study aimed to determine common and 131 distinct behavioral and neural signatures of fear and disgust generalization during 132 associative learning. Particularly, to ensure a sufficient signal-to-noise ratio, we 133 adopted the MultiCS conditioning learning paradigm. In MultiCS conditioning, many 134 similar stimuli were paired with the aversive US (CSs-US association), whereas the 135 same number of similar CS was presented alone (CSs-no US association) (Rehbein et 136 al., 2018). MultiCS conditioning paradigms use a series of similar and complex stimuli 137 to comprise an affective category, making the associative learning process more 138 complex and avoiding rapid extinction in the generalization test (Steinberg et al., 2012; 139 Steinberg et al., 2013). To avoid carry over effects, habituation and expectations in the 140 experimental design participants were divided into two groups in this study, with one 141 group completing the fear generalization paradigm, and the other group completing the 142 disgust learning paradigm. Thus, we analyzed the acquisition and the generalization 143 phase of these two aversive conditioning processes. We hypothesized that the US 144 expectancy of CS+/GS+ would be significantly higher than that of CS−/GS−. Based on 145 the different evolutionary functions of disgust (avoidance of contamination from a 146 class of stimuli) and fear (anticipation of physical attack e.g. in a highly specific 147 context) (e.g. Curtis, de Barra., & Aunger, 2011) we expected enhanced generalization 148 for the conditioned disgusting-CS+ as compared to the fearful-CS+. On the ERP 149 activation level, we hypothesized that (1) fear-conditioned and disgust-conditioned 150 CS+ and GS+ would evoke an early attentional bias reflected by P2; (2) LPP amplitude 151 would be modulated by both stimuli types reflecting that both stimuli types capture 152 7 strong sustained attention possible suggesting threat monitoring; and (3) differential 153 electrophysiological modulation of disgust-relevant CS/GS versus the fear-related 154 CS/GS, in particular larger LPP amplitudes response to conditioned disgust-CS+ than 155 to fear-CS+ given that previous studies reported a larger attentional bias for disgusting 156 than fear stimuli (Charash & McKay, 2002; Carretié et al., 2011) and stronger 157 interference by disgusting stimuli (Cisler et al., 2009; van Hooff et al., 2013). 158 159 2. Materials and Methods 160 2.1 Participants 161 A priori sample size calculation (G*Power) indicated that 52 participants in total 162 would be sufficient to achieve a medium effect size of 0.20, an alpha level of 0.05, and 163 a 1-beta level of 0.80 (Erfelder, Faul, & Buchner, 1996; Faul et al., 2007; Hendrikx et 164 al., 2021). We recruited 62 healthy college participants (27 women; Mage = 20.87; SDage 165 = 2.51) who were randomly assigned to either fearor disgust-associative learning 166 (n = 31 per group; Agefear = 21.07 ± 2.92; Agedisgust = 20.68 ± 2.09). Five participants 167 (three in fear group: Nfear = 28 and two in disgust group: Ndisgust = 29) were excluded 168 from the final data analysis because they rated the US expectancy of the CS− larger 169 than that of the CS+. All participants had normal or corrected vision and had no history 170 of psychiatric or neurological diseases (according to self-report). All subjects had a BDI 171 score < 13 and STAI < 50 which is in the normal range and thus indirectly confirm the 172 absence of mood, anxiety disoders. Participants provided written informed consent and 173 received monetary compensation. The research was approved by the Medicine Ethics 174 Committee of Shenzhen University and the experimental protocol was established, 175 according to the ethical guidelines of the Helsinki Declaration. 176 2.2 Stimuli 177 8 2.2.1 CS and GS 178 The generalized stimuli used in this study were a modified version of those used 179 in a previous study to maximize signal to noise ratio (Lissek et al., 2008). The 180 conditioned and generalized stimuli were a series of shapes including a circle, triangle, 181 square, and parallelogram, and each shape was presented in a separate block. 182 Specifically, each shape was designed with 10 stimuli, continuously increasing in size 183 (5.08–14.22 cm in diameter, 20% increments) (Figure 1). The assignment of CS+ was 184 counter-balanced between blocks. For two of the four blocks, the smallest stimuli (5.08 185 cm) served as CS+ paired with the US (75% reinforcement), and the largest stimuli 186 served as CS− (14.22 cm). In the remaining two blocks, the largest stimuli were used 187 as CS+, and the smallest stimuli were used as CS−. The remaining stimuli in the middle 188 served as GS. 189 190 Figure 1. The conditioned stimuli (CS) and generalized stimuli (GS) used in the present 191 study, adapted from the procedures of Lissek et al. (2008). The database included four 192 different shapes, each with 10 stimuli continuously increasing in size. The smallest and 193 the largest stimuli served as CS+ and CS−, respectively, and the CS+ and CS− were 194 counterbalanced across participants. The remaining stimuli served as GS. 195 15 Figure 3. US expectancy ratings and mean response time were collected for each trial 322 in the fear acquisition (A) and generalization tasks (B). Error bars represent standard 323 mean errors. CS = conditioned stimulus; GS = generalized stimulus; US = 324 unconditioned stimulus 325 On the other hand, a main effect of Block was found, F (2.039, 112.118) = 33.043; p 326 < .001, η2 = 0.375. Additionally, the Block × Conditioned Stimulus Type interaction 327 was significant, F (2.505, 137.750) = 5.489, p = 0.003, η2 = 0.091. Simple effect analysis 328 showed that, for CS+, the RTs of Acq1 was longer than that of Acq2 ([DiffM 209.170, 329 p < .001; 95%CI (103.405; 314.935)]), Acq3 ([DiffM 181.251, p < .001; 95%CI 330 (72.953; 289.548)]) and Acq4 ([DiffM 165.572, p < .001; 95%CI (60.999; 270.145)]). 331 For CS−, the RTs of Acq1 was larger than that of Acq2 ([DiffM 168.373, p = .001; 332 95%CI (57.986; 278.761)]), Acq3 ([DiffM 228.738, p < .001; 95%CI (110.017; 333 347.459)]) and Acq4 ([DiffM 312.096, p < .001; 95%CI (178.267; 445.924)]); Further, 334 the RTs of Acq2 ([DiffM 143.722, p < .001; 95%CI (43.904; 243.541)]) and Acq3 335 ([DiffM 83.358, p = .004; 95%CI (19.993; 146.722)]) were longer than that of Acq4 336 (Figure 4). 337 338 Figure 4. The time course of response time during the fear acquisition (means ± SEMs). 339 3.3 Generalization phase 340 3.3.1 Subjective expectancy ratings 341 16 The US ratings in both groups in the generalization phase exhibited a significant 342 main effect of Conditioned Type (F(2.279,125.359) = 230.779; p < .001，ηp2= .808). 343 Bonferroni corrected post-hoc analysis revealed that US ratings significantly differed 344 across generalized stimuli (p < .001) except CS+ with GS1 and CS− with GS2 (p 345 ＞ .05) , exhibiting a gradient of generalization. Furthermore, the US ratings during 346 generalization were characterized by a main effect of Emotion Type (F (1,55) = 9.699; p 347 = .003，ηp2= .150) and their interaction (F (2.279,125.359) = 5.808; p = .003，ηp2= .096). 348 Simple effect analysis showed that the five types of disgust-related GS (GS1 349 [DiffM .649, p = .010; 95%CI (.165; 1.134)]), GS2 ([DiffM 1.084, p = .003; 95%CI 350 (0.381; 1.787)]), GS3 ([DiffM 1.311, p = .002; 95%CI (.505; 2.116)]), GS4 ([DiffM 351 1.085, p = .003; 95%CI (.388; 1.782)]) and CS− ([DiffM .840, p = .007; 95%CI (0.243; 352 1.437)]) were larger than those of fear-related GS (Figure 3 #3). 353 3.3.2 Reaction times 354 The main effects for Conditioned Type (F(2.326,127.911) = 31.704; p < .001，ηp2 = 355 .366) reached significance. Apart from GS1 with GS2, and GS4 with CS−, RTs showed 356 an overall downward trend for those followed by Bonferroni corrected post-hoc 357 analysis. The main effect of Emotion Type, F(1,55) = 3.042; p = .087，ηp2 = .052, and 358 the Conditioned Type by Emotion Type interaction, F (2.326,127.911) = 2.005; p = .078， 359 ηp2 = .035, were both non-significant (Figure 3 #4). 360 3.4 ERPs 361 3.4.1 Conditioning phase 362 3.4.1.1 P2 363 P2 was characterized by a marginal significant main effect of Conditioned Type 364 (F(1,55) = 3.635; p = .062，ηp2=.062). Bonferroni corrected post-hoc analysis indicated 365 17 that CS+ evoked an enhanced P2 amplitude during the threat learning process compared 366 with CS−. However, the Emotion Type (F(1,55)= .030; p = .864，ηp2 = .001) and 367 Conditioned Type × Emotion Type (F(1,58)= 2.090; p = .154，ηp2 = .037) were not 368 significant (Figure 5). 369 370 Figure 5. P2 and CNV responses during fear and disgust acquisition. (A) Stimulus371 logged ERPs at FCz channels for CS+fear, CS−fear, CS+disgust, and CS−disgust 372 conditions. (B) The averaged ERP (Fc1, Fcz, Fc2, F1, Fz, F2) of the grand average 373 amplitude of P2 and CNV under different emotional conditions. (C) The scalp 374 18 topography of the grand average amplitude of P2 and CNV under different emotional 375 conditions. 376 CNV = contingent negative variation; CS = conditioned stimulus; ERP = event-related 377 potential 378 3.4.1.2 CNV 379 We found a significant main effect of Conditioned Type (F(1,55)= 7.630; p = .008 380 ，ηp2= .122) and a marginal significant interaction effect (F(1,55)= 3.788; p = .057，ηp2 381 = .064). Simple effect analysis showed that CNV amplitude was greater in response to 382 Disgust-CS+ compared with Disgust-CS−, whereas CNV amplitudes did not differ in 383 the late time window between the Fear-CS+ and the Fear-CS− conditions. Similarly, 384 the CNV ERP results revealed no significant main effect of Emotion Type with CNV 385 values (F(1,55)= 1.932; p = .170，ηp2= .034) (Figure 5). 386 3.4.2 Generalization phase 387 3.4.2.1 CNV 388 CNV analysis yielded a significant main effect of Conditioned Type (F(3,165)= 389 3.459, p = .018, ηp2= .059) and a significant interaction effect (F(3,165)= 3.573; p 390 = .015，ηp2= .061). However, we did not find a significant main effect of Emotion 391 Type (F(1,55)= .032, p = .859, ηp2= .001). The simple effect analysis revealed that the 392 CNV amplitudes of GS1, GS2, GS3 and GS4 were not significantly different under the 393 disgust condition (p > .05), but the CNV amplitude of GS1 [DiffM -7.351, p = .003; 394 95%CI (-12.718; -1.985)] and GS2 [DiffM -4.418, p = .032; 95%CI (-8.592; -.245)] 395 were significantly higher than that of GS4. The difference between GS1 and GS3 was 396 marginally significant in the fear condition. Overall, the CNV amplitude showed a 397 generalization gradient (Figure 6). 398 399 19 400 Figure 6. The CNV results during fear generalization. (A) Stimulus-logged ERPs at 401 FCz channels for GS1, GS2, GS3, and GS4 under fear and disgust conditions. (B) The 402 averaged ERP (Fz, Cz, FCz) of the grand average amplitude of CNV under different 403 emotional conditions. (C) The scalp topography of the grand average amplitude of CNV 404 under different emotional conditions. 405 CNV = contingent negative variation; CS = conditioned stimulus; ERP = event-related 406 potential; GS = generalized stimulus 407 408 4. Discussion 409 The present study aimed at determining common and differential behavioral and 410 neural responses during disgust and fear generalization by means of capitalizing on a 411 20 multi-CS conditioning and generalization paradigm with concomitant ERP acquisition. 412 On the behavioral level we found greater US expectancy ratings for CS+ than for CS− 413 in both emotional domains, indicating successful acquisition of CS+-US contingencies 414 and an effective experimental manipulation (Koban et al., 2018; Wong & Lovibond, 415 2017). Individuals reported elevated US expectancy ratings for disgust-CS− as 416 compared to fear-CS−, possibly reflecting that fear induces a stronger discriminative 417 conditioning with respect to the safety signal (CS− , Takemoto & Song, 2019), or 418 alternatively that the fear-related CS− might show a stronger inhibition relative to 419 disgust-relevant CS−. 420 In the generalization phase, the US expectancy ratings showed a gradual decline 421 as a function of decreasing CS+ similarities across both emotion types. Ratings of US 422 expectancy provide an index of ’subjective CS discrimination’ and drive the 423 conditioned response and associated generalization gradients (Lonsdorf et al., 2017; 424 Harvie et al., 2017). Expectancy ratings for disgust generalization stimuli were however 425 generally higher than for the fear generalization stimuli reflecting a stronger ratings of 426 ‘risk’ for disgust than for fear. Fear may occur in response to immediate threats, 427 perceived as a risk of injury or death, whereas disgust is an emotional response to 428 stimuli considered distasteful or contaminative (Curtis, 2011). Although both represent 429 defensive avoidance reactions characterized by aversive negative arousal and 430 withdrawal, previous studies suggested that it was harder to remember contaminating 431 vs. threatening stimuli since disgust is associated with avoidance and suppressed 432 sensory exposure (Susskind et al., 2008). Thus, one possible explanation for this 433 stronger rating of ‘risk’ in disgust in turn lead to a relatively poor accuracy of the CS 434 memory representation. Similar stimuli were wrongly categorized to the original one, 435 leading to a border generalization gradient (Zenses et al., 2021). The stronger US 436 21 expectancy ratings in disgust could reflect an evolutionary adaptive mechanism given 437 the often less explicit indices of pathogen contamination as compared to a direct, e.g. 438 attack-related, threat. Together the findings underscore differential behavioral and 439 neural signatures of fear and disgust generalization which may contribute differentially 440 to psychiatric conditions with dysregulations in aversive avoidance mechanisms, e.g. 441 anxiety or obsessive-compulsive disorders (Armstrong & Olatunji, 2017). 442 On the behavioral level, we observed that RTs gradually decreased over the 443 learning course during acquisition, which supports the view that RTs may reflect the 444 level of confidence (Lissek et al., 2008) with a higher confidence in the estimation of 445 risks leading to decreasing RTs. Further, we found that decreasing RTs with decreasing 446 similarity with the CS+ in generalization which may be explained in terms of the 447 reinforcement rate, because the CS+-US association was 75% whereas the CS−were 448 always presented alone (Lei et al., 2019). Several associative learning studies employed 449 RTs to assess the associative strength between specific events and outcomes (Craddock 450 et al., 2012). Comparing the stimuli that resemble to CS+, stimuli similar to CS− 451 required less time to make decisions. A short RT to the outcome indicates a strong 452 associative strength, whereas a longer RT may suggest a comparably weaker 453 associative strength between the event and its outcome. 454 Regarding the ERP results in the acquisition phase, we observed increased P2 455 amplitude for CS+ relative to CS− irrespective of emotion type. The early modulatory 456 effect on the P2 demonstrates an electrophysiological index of directed selective 457 attention (Ugland et al., 2013). A similar P2 modulation effect was found in the study 458 by Kluge et al. (2011) employing electric shock as US during a fear acquisition 459 paradigm. Previous studies suggested that increased early P2, in response to 460 emotionally aversive stimuli, may reflect automatic attention capture and threat-related 461 22 attention biases (Lei et al., 2019; Willner et al., 2020). The enhanced P2 amplitudes for 462 conditioned salient stimuli may index motivated attention (Zheng et al., 2019). 463 Together the findings indicate that conditioned fear and disgust engage comparable 464 early attentional resource engagement and salience processes. From a biological 465 perspective, both fear and disgust require rapid defensive avoidance responses in the 466 face of threatening stimuli, and thus early threat detection and deployment of attentional 467 resources towards both classes of stimuli represents a critical initial step of the 468 defensive avoidance response (Buck et al., 2018). 469 Associative learning describes the acquisition of stimulus-outcome contingencies 470 and conditioned threat CS+ predicts the occurrence of the US. The CS+ could elicit an 471 anticipation of US occurrence due to this predictive relationship (Pittig, et al., 2018). 472 The CNV components are hypothesized to index a processes of cognitive appraisal and 473 contingency evaluation (eg., Proulx & Picton, 1984; Regan & Howard, 1991). The 474 current analyses showed that parietooccipital CNV amplitudes were significantly larger 475 in response to conditioned disgust-CS+ than to disgust-CS−, which might reflect the 476 cognitive processes of anticipation and preparation of defensive responses to a potential 477 disgust triggering stimulus (US). 478 The current analyses showed that parietooccipital CNV amplitudes were 479 significantly larger in response to conditioned disgust-CS+ than to disgust-CS−, yet 480 interestingly the CNV amplitudes did not significantly differ for the fear-associated 481 CS+ than CS− stimuli. Previous aversive conditioning studies using ERPs found 482 increased CNV amplitude in response to CS+ in response to stimuli which may induce 483 subjective feelings of fear as well as disgust (e.g. small animal pictures Regan & 484 Howard, 1995), suggesting that biological salient threat stimuli can induce a 485 modulation of motivated attention or sustained attention bias. The findings resonate 486 23 with previous lesion and brain imaging studies suggesting common yet also separable 487 neural responses to fear and disgust-inducing stimuli (e.g. Stark et al., 2003, 2007). 488 Although some features of the defensive avoidance reaction in response to disgust and 489 fear are similar other features such as the specific facial expression or the subjective 490 experience differ. The behavioral responses may specifically differ in terms of the 491 evolutionary function in terms of danger avoidance. Moreover, disgust may manifest in 492 OCD with contamination fears thus suggesting differential underlying biological 493 processes (Knowles et al., 2018). Differentiating temporal dynamics of ERPs that 494 respond to the fear and disgust may thus represent an important neurobiological 495 differentiation between the defensive avoidance reactions and psychiatric conditions 496 characterized by fear versus disgust dysregulations. 497 Nelson at al. (2014) examined the electrodermal activity of fear generalization by 498 using ERPs. The results revealed that LPP was more enhanced for CS+ relative to CS−, 499 whereas it did not differ among GS, indicating that this component is not sensitive to 500 fear generalization. Our results exhibited an overall CNV fear generalization gradient, 501 furthermore, the GS showed an attenuated CNV effect with decreasing similarity to 502 CS+. One possible explanation for this CNV gradient pattern was that the late-latency 503 periods may index the fear generalization for CS+. These findings may suggest that 504 CNV in particular may reflect anticipation of the GS-US association. As for the CNV 505 in disgust generalization, the CNV amplitude did not differ among the GS (GS1, GS2, 506 GS3, GS4) stimuli. This might suggest that the subtle differences between disgust 507 generalized stimuli could not be detected by CNV. Considering the absence of adequate 508 evidence, caution should be exercised when considering these interpretations, and 509 further research is warranted. 510 Findings of the present study need to be considered in the context of limitations. 511 24 First, we applied a between-subject design to avoid cross-stimulus conditioning or 512 extinction, and the participants were randomly assigned to fear or disgust learning 513 groups. Thus, individual variability between groups might contribute to the findings. If 514 one kind of CS are conditioned to fear and another kind of CS are conditioned to 515 disgust, this limitation may be overcome in future research. Second, eye movement 516 patterns can provide temporal accuracy measures of emotional stimuli processing. 517 Thus, examining how fear and disgust learning affect eye tracking differently could 518 reflect the perceptual and cognitive process in these two learned threats. Future research 519 should consider using eye-movement methodology in conjunction with ERPs to 520 investigate the fear versus disgust generalization pattern. Third, the pictures used in this 521 study to manipulate the type of US were rather weak and might have impacted the 522 results, especially for the CNV electrophysiological index. Unpleasant odors, for 523 example Civette, which smells like feces, could be used to serve as disgust-US, 524 however, might be difficult to match with the fear-associated stimulus. Further studies 525 can use more disgustor fear-evoking US instead of images to lead to a stronger CS526 US association. Finally, the present work may provide implications for clinical research 527 on fearand disgust-associated disorders such as contamination OCD. Pathology 528 models suggest that both, exaggerated contamination fear and heightened disgust 529 proneness play a role in the development and maintenance of this condition (Eyal, Dar, 530 & Liberman, 2021). The present results indicate that the underlying defensive 531 avoidance mechanisms are – at least in healthy individuals – separable. Despite the 532 limited generalization of the present findings to clinical populations (although see the 533 importance for proof-of-concept studies for clinical OCD (Abramowitz et al., 2021) – 534 future studies may examine common and separable contributions of dysregulations in 535 these domains as potential pathoand vulnerability-mechanism for contamination OCD. 536 31 Mason, E. C., & Richardson, R. (2010). Looking beyond fear: the extinction of other 708 emotions implicated in anxiety disorders. Journal of anxiety disorders, 24(1), 63– 709 70. doi:10.1016/j.janxdis.2009.08.007 710 McGuire, J. F., Orr, S. P., Essoe, J. K., McCracken, J. T., Storch, E. A., & Piacentini, 711 J. (2016). Extinction learning in childhood anxiety disorders, obsessive 712 compulsive disorder and post-traumatic stress disorder: implications for 713 treatment. Expert review of neurotherapeutics, 16(10), 1155–1174. doi: 714 10.1080/14737175.2016.1199276. 715 Moeck, E. K., Matson, L. A., & Takarangi, M. (2021). Mechanisms underlying memory 716 enhancement for disgust over fear. Cognition & emotion, 1–7. Advance online 717 publication. https://doi.org/10.1080/02699931.2021.1936460 718 Nelson, B. D., Weinberg, A., Pawluk, J., Gawlowska, M., & Proudfit, G. H. (2015). An 719 event-related potential investigation of fear generalization and intolerance of 720 uncertainty. Behavior Therapy, 46, 661-670. doi:10.1016/j.beth.2014.09.010 721 Olatunji, B. O., Armstrong, T., & Elwood, L. (2017). Is disgust proneness associated 722 with anxiety and related disorders? A qualitative review and meta-analysis of 723 group comparison and correlational studies. Perspectives on Psychological 724 Science, 12, 613–648. doi:10.1177/1745691616688879 725 Olatunji, B. O., Forsyth, J. P., & Cherian, A. (2007). Evaluative differential 726 conditioning of disgust: a sticky form of relational learning that is resistant to 727 extinction. Journal of Anxiety Disorders, 21, 820–834. 728 doi:10.1016/j.janxdis.2006.11.004 729 Olatunji, B. O., Moretz, M. W., Wolitzky-Taylor, K. B., McKay, D., McGrath, P. B., 730 & Ciesielski, B. G. (2010). Disgust vulnerability and symptoms of contamination731 based OCD: descriptive tests of incremental specificity. Behavior therapy, 41(4), 732 475–490. https://doi.org/10.1016/j.beth.2009.11.005 733 Olatunji, B. O., Wolitzky-Taylor, K. B., Willems, J., Lohr, J. M., & Armstrong, T. 734 (2009). Differential habituation of fear and disgust during repeated exposure to 735 threat-relevant stimuli in contamination-based OCD: an analogue study. Journal 736 of Anxiety Disorders, 23, 118–123. doi:10.1016/j.janxdis.2008.04.006 737 Pavlov, Y.G., & Kotchoubey, B. (2019). Classical conditioning in oddball paradigm: A 738 comparison between aversive and name conditioning. Psychophysiology 56, 739 e13370. doi:10.1111/psyp.13370 740 Peyrot, C., Brouillard, A., Morand-Beaulieu, S., & Marin, M. F. (2020). A review on 741 32 how stress modulates fear conditioning: Let's not forget the role of sex and sex 742 hormones. Behaviour research and therapy, 129, 103615. 743 https://doi.org/10.1016/j.brat.2020.103615 744 Pittig, A., Treanor, M., LeBeau, R. T., & Craske, M. G. (2018). The role of associative 745 fear and avoidance learning in anxiety disorders: gaps and directions for future 746 research. Neuroscience & Biobehavioral Reviews, 88, 117–140. 747 doi:10.1016/j.neubiorev.2018.03.015 748 Proulx. G. B., & Picton, T. W. (1984). The effects of anxiety and expectancy on the 749 CNV. Annals of the New York Academy of .Sciences, 425, 617–622. 750 doi:10.1111/j.1749-6632.1984.tb23586.x 751 Regan. M., & Howard, R. (1991). Controllability, predictability, and event-related 752 potentials to fear-relevant and fear-irrelevant stimuli. Journal of 753 Psychophysiology, 5, 43–57. 754 Regan, M., & Howard, R. (1995). Fear conditioning, preparedness, and the contingent 755 negative variation. Psychophysiology, 32, 208–214. doi:10.1111/j.1469756 8986.1995.tb02950.x 757 Rehbein, M. A., Pastor, M. C., Moltó, J., Poy, R., López-Penadés, R., & Junghöfer, M. 758 (2018). Identity and expression processing during classical conditioning with 759 faces. Psychophysiology, 55, e13203. doi:10.1111/psyp.13203 760 Sangha, S., Diehl, M. M., Bergstrom, H. C., & Drew, M. R. (2020). Know safety, no 761 fear. Neuroscience & Biobehavioral Reviews, 108, 218–230. 762 doi:10.1016/j.neubiorev.2019.11.006 763 Sperl, M. F. J., Wroblewski, A., Mueller, M., Straube, B., & Mueller, E. M. (2021). 764 Learning dynamics of electrophysiological brain signals during human fear 765 conditioning. NeuroImage, 226, 117569. doi:10.1016/j.neuroimage.2020.117569 766 Stark, R., Schienle, A., Walter, B., Kirsch, P., Sammer, G., Ott, U., . . . Vaitl, D. (2003). 767 Hemodynamic responses to fear and disgust-inducing pictures: an fMRI study. 768 International Journal of Psychophysiology, 50(3), 225-234. doi:10.1016/S0167769 8760(03)00169-7 770 Stark, R., Zimmermann, M., Kagerer, S., Schienle, A., Walter, B., Weygandt, M., & 771 Vaitl, D. (2007). Hemodynamic brain correlates of disgust and fear ratings. 772 NeuroImage, 37(2), 663-673. doi:10.1016/j.neuroimage.2007.05.005 773 Stein, D. J., Liu, Y., Shapira, N. A., & Goodman, W. K. (2001). The psychobiology of 774 obsessive–compulsive disorder: how important is the role of disgust? Current 775 33 Psychiatry Reports, 3, 281–287. doi:10.1007/s11920-001-0020-3 776 Steinberg, C., Bröckelmann, A.K., Rehbein, M., Dobel, C., & Junghöfer, M. (2013). 777 Rapid and highly resolving associative affective learning: convergent electroand 778 magnetoencephalographic evidence from vision and audition. Biological 779 Psychology, 92, 526–540. doi:10.1016/j.biopsycho.2012.02.009 780 Steinberg, C., Dobel, C., Schupp, H.T., Kissler, J., Elling, L., Pantev, C., & Junghöfer, 781 M. (2012). Rapid and highly resolving: affective evaluation of olfactorily 782 conditioned faces. Journal of Cognitive Neuroscience, 24, 17–27. 783 doi:10.1162/jocn_a_00067 784 Susskind, J. M., Lee, D. H., Cusi, A., Feiman, R., Grabski, W., & Anderson, A. K. 785 (2008). Expressing fear enhances sensory acquisition. Nature neuroscience, 11(7), 786 843–850. https://doi.org/10.1038/nn.2138 787 Takemoto, M., & Song, W.J. (2019). Cue-dependent safety and fear learning in a 788 discriminative auditory fear conditioning paradigm in the mouse. Learning & 789 Memory, 26, 284–290. doi:10.1101/lm.049577.119 790 Tuominen, L., Boeke, E., DeCross, S., Wolthusen, R. P., Nasr, S., Milad, M., Vangel, 791 M., Tootell, R., & Holt, D. (2019). The relationship of perceptual discrimination 792 to neural mechanisms of fear generalization. NeuroImage, 188, 445 –455. 793 doi:10.1016/j.neuroimage.2018.12.034 794 Ugland, C.C., Dyson, B.J., & Field, A.P. (2013). An ERP study of the interaction 795 between verbal information and conditioning pathways to fear. Biological 796 Psychology, 92, 69–81. doi:10.1016/j.biopsycho.2012.02.003 797 van Hooff, J. C., Devue, C., Vieweg, P. E., & Theeuwes, J. (2013). Disgustand not 798 fear-evoking images hold our attention. Acta psychologica, 143(1), 1–6. 799 https://doi.org/10.1016/j.actpsy.2013.02.001 800 Ventura-Bort, C., Löw, A., Wendt, J., Dolcos, F., Hamm, A.O., & Weymar, M. (2016). 801 When neutral turns significant: brain dynamics of rapidly formed associations 802 between neutral stimuli and emotional contexts. European Journal of 803 Neuroscience, 44, 2176–2183.s. doi:10.1111/ejn.13319 804 Vriends, N., Michael, T., Blechert, J., Meyer, A.H., Margraf, J., & Wilhelm, F.H. 805 (2011). The influence of state anxiety on the acquisition and extinction of fear. 806 Journal of Behavior Therapy and Experimental Psychiatry, 42, 46–53. 807 doi:10.1016/j.jbtep.2010.09.001 808 34 Wang, J., Sun, X., Lu, J., Dou, H., & Lei, Y. (2021). Generalization gradients for fear 809 and disgust in human associative learning. Scientific reports, 11(1), 14210. 810 https://doi.org/10.1038/s41598-021-93544-7 811 Waters, A.M., Peters, R.M., Forrest, K.E., & Zimmer-Gembeck, M. (2014). Fear 812 acquisition and extinction in offspring of mothers with anxiety and depressive 813 disorders. Developmental Cognitive Neuroscience, 7, 30–42. 814 doi:10.1016/j.dcn.2013.10.007 815 Willner, C.J., Jetha, M.K., Segalowitz, S.J., & Gatzke-Kopp, L.M. (2020). 816 Neurophysiological evidence for distinct biases in emotional face processing 817 associated with internalizing and externalizing symptoms in children. Biological 818 Psychology, 150, 107829. doi:10.1016/j.biopsycho.2019.107829 819 Wong, A.H.K., & Lovibond, P.F. (2017). Rule-based generalisation in single-cue and 820 differential fear conditioning in humans. Biological Psychology, 129, 111–120. 821 doi:10.1016/j.biopsycho.2017.08.056 822 Woody, S.R., McLean, C., & Klassen, T. (2005). Disgust as a motivator of avoidance 823 of spiders. Journal of Anxiety Disorders, 19, 461–475. 824 doi:10.1016/j.janxdis.2004.04.002 825 Woody, S.R., & Teachman, B.A. (2000). Intersection of disgust and fear: normative 826 and pathological views. Clinical Psychology Science and Practice, 7, 291–311. 827 doi:10.1093/clipsy/7.3.291 828 Yau, J.O.Y., & McNally, G.P. (2018). Brain mechanisms controlling Pavlovian fear 829 conditioning. Journal of Experimental Psychology: Animal Learning and 830 Cognition, 44, 341–357. doi:10.1037/xan0000181 831 Zenses, A. K., Lee, J. C., Plaisance, V., & Zaman, J. (2021). Differences in perceptual 832 memory determine generalization patterns. Behaviour research and therapy, 136, 833 103777. https://doi.org/10.1016/j.brat.2020.103777 834 Zheng, P., Lyu, Z., & Jackson, T. (2019). Effects of trait fear of pain on event-related 835 potentials during word cue presentations that signal potential pain. European 836 Journal of Neuroscience, 50, 3365–3379. doi:10.1111/ejn.14495 837 838