Does the assessment of different combinations of within-phase subjective measures influence electrodermal responding and between-phase subjective ratings during fear conditioning and extinction experiments?

Highlights

• Assessing no within-phase ratings was associated with return of electrodermal responding.

• Within-phase US expectancy ratings resulted in differential conditioning and sustained extinction into the test phase than the other conditions.

• Assessing within-phase CS evaluations interfered with extinction of SCRs and exhibited smaller CS- SCRs in all phases than other conditions.

• Assessing two within-phase ratings increased physiological arousal to both CSs during acquisition and extinction compared to all other conditions.

• Assessing two within-phase ratings resulted in differential SCRs returning at the test phase.

Abstract

The type and number of dependent measures assessed have varied between fear conditioning and extinction experiments and it remains unclear whether methodological differences influence results. Measuring skin conductance responses (SCRs) and between-phase subjective ratings between four conditions; Con condition had no within-phase ratings; US Exp condition included within-phase US expectancy ratings only; CS Eval condition included within-phase CS evaluations only and All Meas condition included both. All Meas condition exhibited larger SCRs compared to other conditions during acquisition, extinction, and return of differential responding at test. Differential SCRs did not extinguish in CS Eval and Con conditions, and CS Eval condition exhibited smaller CS- SCRs than other conditions throughout phases. US Exp condition revealed differential conditioning, successful extinction, and no return of differential SCRs. Between-phase ratings were not affected by within-phase ratings. Researchers should consider assessing different combinations of within-phase subjective ratings depending on the aims of their research.

Introduction

During fear conditioning, a participant learns to fear a neutral conditioned stimulus (CS+) that is paired with an aversive unconditioned stimulus (US). In differential fear conditioning, the participant is also exposed to another conditioned stimulus (CS-) that is not associated with the aversive US. Throughout the experiment, various responses are assessed to index fear conditioning, extinction, and extinction retention. These typically include one or more subjective ratings (e.g., CS evaluation; US expectancy; subjective fear/anxiety) that are assessed either within-phase (i.e., on every trial or intermittently throughout each phase of the experiment) and/or after the completion of each phase (i.e., between-phase ratings). US expectancy ratings provide an indication of the extent to which the participant anticipates the delivery of the US upon presentation of a CS (Lovibond & Shanks, 2002), and are suggested to be more valid and reliable when assessed within-phase than between-phase (Dawson & Reardon, 1973; Lovibond & Shanks, 2002). CS valence ratings provide an indication of evaluative learning and the extent to which the participant’s liking of the conditioned stimulus changes with new learning experiences and are often assessed within- and between-phase (De Houwer, Thomas, & Baeyens, 2001). Participants subjective fear/anxiety ratings are also commonly assessed, either within and/or between-phases (Britton et al., 2013; Lau et al., 2008). Objective measures of electrodermal responding (e.g., skin conductance responses; SCR) and defensive responding (e.g., fear potentiated startle; FPS) are also assessed on a trial-by-trial basis during each phase (see Lipp, 2006). SCR is a measure of sweat gland activity associated with the sympathetic division of the autonomic nervous system and provides a measure of physiological arousal to salient stimulus events (Blumenthal et al., 2005; Boucsein et al., 2012; Dawson, Schell, & Courtney, 2011).

If differential conditioning has occurred, the participant will display significantly larger US expectancy ratings on CS + than CS- trials, more negative evaluations of the CS + than the CS-, and larger SCRs to the CS + than the CS- (Duits et al., 2015). During the extinction phase, the participant is repeatedly presented with both CS + and CS- without the US, and if extinction learning has occurred, the participant displays a reduction in US expectancies, negative CS evaluations, and SCRs on CS + trials, and is often indexed as no significant difference between the CSs by the end of extinction (Duits et al., 2015). The test phase is procedurally similar to extinction and provides an index of the extent to which extinction learning is retained (i.e., no significant differences between the CSs) or if fear has returned (i.e., responses to the CS + are larger than to the CS-) (Vervliet, Baeyens, Van den Bergh, & Hermans, 2013).

A recent systematic review of design parameters for differential fear conditioning and extinction experiments by Ryan, Zimmer-Gembeck, Neumann, and Waters (2019) revealed that prior studies have differed considerably in the extent to which self-report and physiological measures of fear conditioning and extinction studies have been assessed within-phase and/or between-phases. Moreover, variation in the type and method of assessing dependent variables in prior studies has made it difficult to determine whether these differences actually affect learning (cf. Lonsdorf et al., 2017). Subjective responses assessed within-phase are likely to increase engagement and processing of CS-US associations more so than passive participation when subjective measures are assessed only before and after each phase of the experiment (Collins & Shanks, 2002). Furthermore, it is unclear whether assessing CS evaluations vs US expectancies or both measures differentially affect learning.

It is important to directly compare the effects of within-phase measures on other within- and between-phase learning indices for several reasons. First, the observation of differing outcomes as a function of the measures assessed across studies could have implications for establishing methodological consistency, as some key findings of differences in extinction or return of fear may, in part, be influenced by differences in the assessment of the dependent measures (Bach & Melinscak, 2020; Vervliet et al., 2013). Indeed, Lonsdorf et al. (2017) suggested that the inclusion of within-phase measures does not only assess but can also influence learning. Second, implications of such differences could apply to exposure therapy, in that asking patients to provide their threat expectancies and evaluations of feared stimuli during exposure trials might either enhance, or interfere with, exposure therapy outcomes (Fullana et al., 2020; Waters, LeBeau, & Craske, 2017). Third, if differences in outcomes of extinction training are influenced by assessing within-phase vs no within-phase measures (e.g., US expectancy, CS evaluations or both), this might raise important theoretical questions about the role of cognitive processes during fear acquisition and extinction. For instance, requiring within-phase evaluations of US expectancy and CS valence might engage more top-down cognitive resources and thus influence learning more than passively viewing the CSs without active within-phase ratings (e.g., Hofmann, 2008; Lovibond, 2004; Waters & Craske, 2016). In addition to theoretical implications, such findings would have further implications for the extent to which novel treatment strategies should enhance or minimise cognitive strategies during extinction/exposure therapy to improve outcomes (Lipp, Waters, Luck, Ryan, & Craske, 2020).

One study compared intermittently assessed subjective CS fear ratings and no ratings on SCR and FPS during acquisition, extinction and reinstatement phases (i.e., asking participants three times throughout nine conditioning trials “How much stress, fear, or anxiety did you experience the last time you saw symbol X?” with the X referring to one of the CSs at a time) (Sjouwerman, Niehaus, Kuhn, & Lonsdorf, 2016). This study included three conditions in which SCRs were assessed in all three conditions, one additionally included both FPS and ratings, one included FPS but not ratings, and the other group included ratings but no FPS. The results relating to the impact of intermittently presented fear ratings of the CSs showed that SCRs and FPS during acquisition differed between the CS + and CS- regardless of whether fear ratings were measured. In other words, there were no significant differences in discrimination learning between groups that made fear ratings and groups that did not. Although caution should be taken in using the lack of a statistically significant difference as evidence for the lack of an effect, the findings nevertheless support the conclusion that obtaining fear ratings will not interfere with the acquisition of SCRs and FPS. However, in the last half of trials in the extinction phase, the group that did not provide fear ratings exhibited significant differences in SCRs to the CS + compared to the CS- whereas the groups providing CS fear ratings did not. For FPS, differential responding in both groups extinguished by the end of the extinction phase, with no significant differences between the CSs. Thus, groups providing within-phase subjective CS fear ratings extinguished SCRs with no statistically significant difference between CSs compared to the group that did not provide within-phase CS fear ratings.

A study by Warren et al. (2014) compared FPS when assessing the absence or presence of within-phase trial-by-trial US expectancy during every trial of acquisition, extinction, re-extinction and reinstatement.¹ This study was conducted over three days, using two CS + and one CS- (geometric shapes), the US was an airblast to the larynx, and the FPS startle was white noise. During fear conditioning on day 1, both groups showed significant conditioning effects, however the within-phase US expectancy group revealed an increase in FPS to the two CS + stimuli compared to the no US expectancy group. Although extinction occurred on day 2 in both conditions, with FPS to the CS + comparable to the CS- by the last block of extinction, the within-phase US expectancy group exhibited larger FPS overall with a significant main effect of group to the CSs compared to the no ratings group. Moreover, both conditions relapsed during the retest phase as indexed by a significant increase of FPS to the first block of the test phase compared to the last block of extinction. However, the within-phase US expectancy group again exhibited significantly higher levels of FPS overall than the no US expectancy group. Thus, providing within-phase US expectancy ratings appeared to elevate participants’ defensive physiological responding compared to providing no subjective measures. These findings are in contrast to Sjouwerman et al. (2016) who did not find that CS fear ratings had a differential effect on FPS. This may be due to differences in the types of ratings being assessed (i.e., fear of the CSs versus US expectancies) as well as the difference between within-phase ratings i.e., trial-by-trial (Warren et al., 2014) and intermittent (Sjouwerman et al., 2016). Warren et al. (2014) did not include electrodermal responding so the effects of within-phase US expectancy ratings on SCR could not be assessed.

These studies provide some indication that providing within-phase subjective CS fear ratings enhance extinction of electrodermal reactivity and US expectancies increase defensive physiological responding. However, no studies have compared the effect of assessing individual or combinations of within-phase subjective measures relative to no within-phase subjective measures on within-phase electrodermal responses and subjective measures assessed between-phases. Therefore, it is important to consider theoretical models of fear learning to formulate hypotheses about the effect of within-phase subjective ratings on within-phase electrodermal responding and between-phase subjective ratings.

In the review by Lonsdorf et al. (2017), it was suggested that within-phase assessment of US expectancy might enhance attention to the CS-US contingency and “may influence (i.e., boost) the learning process” rather than just measuring it (p 268). According to Boddez, Baeyens, Luyten, and Vansteenwegen (2013), expecting the danger of the US plays a causal role in determining physiological responses (i.e., greater autonomic CR, see Davey, 1992) and avoidance behaviours. Thus, enhancing attention to the CS-US contingency by providing within-phase US expectancy ratings might enhance the processing of the predictability of the US and thus, attenuate electrodermal responding to the CS + compared to conditions without within-phase ratings. Moreover, by enhancing attention to the changing CS-no US contingency by providing within-phase US expectancy ratings during extinction, differences in electrodermal responding between the CS + and CS- might be expected to extinguish (i.e., no longer significantly different) compared to conditions that do not require within-phase US expectancy ratings during extinction. No studies to date have assessed SCR with and without assessing US expectancy ratings.

Lonsdorf et al. (2017) suggested that affect ratings taken within-phase may differentially induce several potential learning-related processes, i.e., they may “generally interact with the development of contingency awareness, unintentionally trigger other memory processes (i.e., reconsolidation after reactivation through CS + presentation during the rating) or directly influence the strength of the developing CR due to the emotion regulating effect of affect labelling” (p. 269). The type of CS evaluation assessed has varied amongst studies, with some studies assessing how fearful or afraid participants are of the CS (Lau et al., 2008; Sjouwerman et al., 2016) or how much they evaluate the valence of the CS as being unpleasant-pleasant (Waters, LeBeau et al., 2017; Waters, Theresiana, Neumann, & Craske, 2017; Zbozinek, Hermans, Prenoveau, Liao, & Craske, 2015). It is unclear whether the different focus of these affective-related ratings differentially impact learning. De Houwer et al. (2001) reviewed the evaluative conditioning literature (i.e., the liking of a stimulus resulting from pairing with a negative or positive stimulus) in regard to Pavlovian conditioning, establishing that these ratings were resistant to extinction. Lovibond (2004) similarly discussed the possibility that during the rating of CS evaluations, some emotional responses might be activated due to a memory or representation of a stimulus, without the active anticipation of the actual occurrence of that stimulus (i.e., CS- no US contingency) and therefore CS evaluations might be resistant to extinction procedures. The variation that exists among studies in ‘what’ is being evaluated (e.g., fear or valence) might also give rise to differing results. However, in relation to rating the degree to which one evaluates the CS as pleasant-unpleasant (cf. De Houwer et al., 2001; Lovibond, 2004), then negative CS evaluations would be expected to be more resistant to extinction.

Whereas separately assessing US expectancy ratings and CS evaluation ratings within-session may be expected to have different effects on conditioning (i.e., US expectancies enhancing CS-US contingency learning) and extinction learning (i.e., CS evaluations slowing extinction; US expectancies enhancing extinction) compared to no assessment of within-phase ratings, the effect of assessing both measures relative to no measures or either measure alone within-session is unclear. Lovibond (2004) suggested that evaluative conditioning differs from expectancy learning in three ways 1) it occurs independently of conscious awareness of the CS-US relationship; 2) it displays affective reactions of a like/dislike nature; and 3) it can be more resistant to extinction (Lovibond, 2004). Furthermore, assessing CS evaluations within-phase has been proposed to influence CS-US contingency awareness, trigger other memory processes, and potentially influence the strength of the conditioned response (CR) itself (Hofmann, De Houwer, Perugini, Baeyens, & Crombez, 2010; Lonsdorf et al., 2017). Therefore, when providing both ratings within-phase compared to no within-phase ratings or either measure alone, attending to or recalling negative CS evaluations might increase electrodermal arousal to the CSs during acquisition, as expected for the CS evaluation alone condition relative to no subjective measures control condition. Similarly, providing both ratings within-phase or ratings of CS evaluations alone during extinction might be expected to slow extinction of electrodermal responses to the CS + compared to the CS- relative to conditions without CS evaluation ratings, due to the increase of threat salience of the learning context overall.

Although ratings taken within and between-phases have been proposed to engage different evaluation strategies; the former based on an accumulation of learning (within-phase) and the latter on memory recall (between-phase) (Lipp, 2006), no studies to date have directly examined whether providing within-phase subjective ratings differentially affect between-phase ratings. The present study addressed this gap by determining whether providing US expectancy, CS evaluations and both types of ratings differentially influenced acquisition, extinction and extinction retention of not only within-phase electrodermal responding but also between-phase subjective ratings.

This study compared four conditions in which all participants provided within-phase electrodermal responses (i.e., SCRs) and between-phase ratings of CS valence, CS arousal and subjective anxiety. The four conditions differed in the extent to which they assessed additional within-phase subjective ratings. The Control (Con) condition involved no other within-phase measures; the US Expectancy (US Exp) condition additionally assessed within-phase US expectancy ratings only; the CS Evaluation (CS Eval) condition additionally assessed within-phase CS valence ratings only; and the All Measures (All Meas) condition additionally assessed both within-phase US expectancy ratings and within-phase CS valence ratings.

In the context of differential conditioning to the CS + compared to the CS- during acquisition, the following predictions were made based on the assumption that within-phase US expectancies enhances attention to CS-US contingencies (Lonsdorf et al., 2017) and within-phase CS evaluations impacts fear memories (Lipp, 2006). First, it was hypothesised that providing within-phase US expectancy ratings alone (US Exp condition) would increase the predictability of the US on CS + trials and thus attenuate orienting (First Interval Responses) and anticipatory (Second Interval Responses) SCRs to the CS + as well as to the US (Third Interval Responses). Second, it was hypothesised that within-phase CS evaluations given alone (CS Eval condition) or with US expectancies (All Meas condition) would elevate electrodermal arousal and thus enhance these responses compared to the Con condition (in all intervals).

In extinction and test, relative to the Con condition in which extinction of differences in SCRs to the CS + and CS- was expected to return at test in FIRs, (a) the US Exp condition was expected to attain extinction of differential SCRs to the CSs and retain this at test in all intervals, and (b) the CS Eval and All Meas conditions were expected to exhibit larger first, second and third interval SCRs to the CS + compared to the CS- during extinction and test. Analyses were conducted to uncover if the combination of assessed measures between conditions had any effect on the factor of Block, for exploratory purposes. As there is no precedence to predict the between-group differences as a function of the different within-phase ratings conditions, these analyses were conducted for exploratory purposes. Supplementary analyses were conducted to examine (a) the effect of providing ratings of both within-phase measures of US expectancy and CS evaluations in the All Meas compared to US Exp and CS Eval conditions, and (b) the effect of providing within-phase US expectancies and CS evaluations alone ratings or in combination on between-phase CS evaluation ratings and subjective anxiety.