The Influence of Verbal
Instructions on Fear
Learning
Master Thesis
Name: Benedikt Aink
Student- Number: 10003893
Date: 01-06-2015
Supervisors: Tom Beckers & Angelos Miltiadis-
Krypotos
Department of Clinical Psychology
Universiteit van Amsterdam
2 Table of Contents ABSTRACT………... 3 INTRODUCTION………. 3 METHOD………... 7 RESULTS………... 14 GENERAL DISCUSSION………... 23 ACKNOWLEDGEMENTS………... 24 REFERENCES………... 25
3
Abstract
Both anxiety and fear are related to different forms of psychopathology. Previous research in the field of fear learning has found that fear can be acquired via different pathways. The goal of our study was to test whether fear responses differ in magnitude when fear is acquired via instructions or via experience. In addition to previous studies in the field, we measured all the components of responses to fear in order to gain a complete picture of these responses across the different indices of fear: subjective experiences, physiological responses and avoidance tendencies. We also used Bayesian statistics to analyse our results, a statistical approach that allowed us to gather both evidence for differences in the magnitude of fear responses between the different types of fear learning, as well as that supporting the absence of such differences. The results indicate that both pathways lead to comparable fear responses. A limitation of the present study was that no evidence was revealed for different responses with regards to both physiological measures and avoidance behaviour. Results of our study allow us to gain a better understanding on the acquisition of fear and the crucial role language plays in fear learning.
Introduction
Fear is one of the basic emotions serving an organism’s survival by informing it of a potential threat and triggering subsequent fear reactions. Fear can take both adaptive and maladaptive forms. It is adaptive when triggered from genuinely threatening stimuli (e.g. dangerous animals; harmful events) in advance of their occurrence. It is, however, maladaptive when fear is triggered in the absence of a real threat, (e.g. house spiders). In those cases, fear can ultimately lead to emotional distress and avoidance behaviour. It eventually deteriorates individuals’ quality of life as seen in different anxiety disorders and phobias (Levenson, 2003; American Psychiatric Association, 2013). With an estimated life-time prevalence of 19,6% in the Netherlands (De Graaf et al., 2010), anxiety disorders were found to be the most prevalent of all psychiatric disorders here. It is, therefore, crucial to understand how fears and anxiety disorders are acquired.
4
According to Mowrer (1993) fear is a learned response to those conditioned stimuli that prefigure injury, and mainly serves as a motivator for behaviour that brings about a state of relief or security (such as avoidance behaviour).
Traditionally, fear acquisition towards essentially innocuous cues was hypothesized to be the result of encountering a traumatic event (Rachman, 1991). A laboratory paradigm that mimics the acquisition of fear towards innocuous cues is Pavlovian conditioning that entails the pairing of an initially neutral stimulus (CS; e.g., the picture of a geometrical object) with an unconditioned stimulus (the US; e.g., a shock) so that the conditioned stimulus (CS) is sufficient to evoke conditioned responses (CR; e.g., fear) even in absence of the US (Pavlov, 1927).
Although in the early days of psychology it was assumed that fear responses towards essentially innocuous stimuli are learned via direct experience, it later became apparent that Pavlovian conditioning alone could not explain the acquisition of most anxiety disorders (Rachman, 1977). Due to the absence of memories of aversive direct experiences in various individuals with anxiety disorders (e.g. Olsson & Phelps, 2007), Rachman (1977) suggested a three-pathway theory wherein anxiety disorders can be acquired via two more pathways than just direct experience: indirect vicarious learning and transmission of instructions. Vicarious learning is the modification of behaviour resulting from exposure to modelling stimuli (Bandura, 1965). In fear learning by transmission of information/instructions, individuals learn to fear a neutral stimulus based on verbal information about the potential threatening properties of the stimuli. For example, a fear of flying can be acquired via instructions: previous studies have found that most people who have a phobia towards planes did not report a direct negative experience with planes (Nousi et al., 2008). Instructional learning relies on aversive memories or imageries of experiences, without any actual experiences as in Pavlovian conditioning (Olsson and Phelps, 2007)
Differences between individual fear responses after conditioning and instructional learning have already been studied before. Olsson and Phelps (2004) compared differences in skin conductance responses between the three pathways of fear learning. In their experiment participants in the Experience group were conditioned by receiving an electric shock that served as a US after they saw the CS+. In the Vicarious learning group, participants observed another person that pretended to receive an electric shock after seeing the CS+. Participants in the Instructed
5
learning group were merely told that they would receive electric shocks after seeing the CS+. Participants were then confronted with the CS+ and CS- (a stimulus that was never followed by a shock) and their Skin Conductance Responses (SCRs; a measure of physiological arousal that is heightened in fearful situations) were assessed. While similar levels of fear were acquired via the three pathways, awareness of the stimuli was necessary for changes in SCRs in the Instructed learning group while it was not for the other types of learning. Olsson and Phelps assumed that instructional learning must be a different process than direct conditioning with distinct qualities, possible due to the activation of a different side of the amygdala, a part of the brain mainly involved in emotional learning (for more information on brain processes involved in emotional learning, see Funayama et al., 2001 and Morris et al., 1998).
Raes et al. (2014) found that conditioning can strengthen fear responses that have been learned by a combination of instructions and experience. Participants were told that two different stimuli would be followed by a shock although only one of them was actually paired with a shock. SCRs and expectancies of participants to receive an electric shock (i.e., US-expectancies) were enhanced to fear-evoking stimuli when a combination of instructions and experiences was used rather than experience alone.
Lastly, Field and Storksen-Coulson (2007) found that beliefs acquired by verbal information strengthen avoidance tendencies if paired with negative direct experiences. Also, avoidance tendencies were as strong after receiving negative information as after an actual negative encounter with the CS. This result suggests that avoidance tendencies can be influenced by instructions as well as by experience.
Taken together, the results above indicate that language (i.e. instructional learning) can play an important role in the acquisition of fears: not only did the combination of negative direct experiences with instructions heighten fear responses; the instructions alone seemed sufficient to evoke different fear responses, such as physiological, subjective and behavioural ones.
A limitation of the previous studies is that the combination of all comportments of fear was never tested in a single experiment; it remains unclear whether a combination of the different fear responses differs in magnitude if learned via conditioning or instructions.
In general, there are three different fear response systems that can be activated by fear: the subjective experience, the autonomic nervous system (ANS), and
6
behaviour (Mauss & Robinson, 2009)1. Although it would be intuitively assumed that the different emotion response systems strongly cohere, experimental data does not support this assumption (see Mauss et al., 2005 for a brief summary of studies). Therefore, the measurement of all components of fear is crucial to gain a better picture of how the different learning pathways influence the acquisition of fear responses.
The goal of our study was to test whether fear responses do differ anyhow in magnitude when fear is learned, either via direct experience or via instructions. In extension to previous studies, we measured all components of fear as described above: subjective experiences, the autonomic nervous system and behaviour.
In our study, subjective experiences were measured by self-report of fear (Mauss, 2009).
The physiological fear responses of ANS and CNS were assessed by performing a facial electromyography (EMG) of the startle reflex to a loud noise (Mauss, 2009). The startle reflex is a defensive reflex in potentially dangerous situations (Yeoman’s et al. 2002) that is enhanced in aversive fear conditioning when confronted with fear-relevant stimuli (Weike et al., 2007; Cook & Turpin, 1997). SCRs can be used to assess physiological fear responses of the ANS. Increased skin conductance responses indicate increased physiological arousal (McCleary, 1950). It was found that SCRs are influenced both by positive as well as by negative valence US – unlike startle reflexes that are only potentiated to aversive US (Hamm & Vaitl, 1996), – therefore both measures were assessed.
Avoidance tendencies can be measured with a symbolic approach-avoidance reaction task (AAT) wherein the interaction between the valence of the presented stimuli (i.e., positive or negative) and the instructed response (i.e., approach or avoid) influences reaction times on a computer task (Chen, Bargh, 1999). The task is based on the finding that negative stimuli (such as the CS+) are avoided faster while positive or neutral stimuli (the CS-) are approached faster, rather than vice versa (Chen & Bargh, 1999).
Because anxiety disorders can be acquired via these two pathways, we might be able to make suggestions on how fear responses can differ depending on the role
1 Some models even suggest the central nervous system (CNS) is a fourth system involved in emotional responses. (Buck, 1999)
7
that language plays in fear and fear learning, with potential implications for the etiology and treatment of anxiety disorders.
In the present study we investigated whether fear responses differ or do not differ if learned via either instructions or direct conditioning. In extension to previous studies, Bayesian statistics were used to gather both evidence for the presence of group differences as well for the absence (Dienes, 2011), because the absence of differences cannot be tested in frequential statistics (Rouder et al., 2011).
Our first hypothesis was that fear learning takes place both via direct conditioning as well as via instructions. After participants had learned the CS-US contingencies either by instructions or experience, fear responses were expected to be higher to the CS+ than to the CS- in both groups. This would be indicated by higher responses to the CS+ compared to the CS- on all measures of the three components of fear and avoidance tendencies and would provide evidence for fear learning via both pathways.
During further exploration, we tested whether fear responses and avoidance tendencies to the CS+ do or do not differ between groups. If measures of fear responses are similar independent of the pathway of learning, the hypothesis that fear responses to the CS+ after instructional learning are the same as fear responses after experience could be supported, while any differences corroborate the hypothesis that fear responses are different if learned via different pathways.
Method
Participants
Fifty-four participants who reported no prior conditioning experience were recruited using the UvA Onderzoekspanel (https://www.lab.uva.nl/spt/). Participants that reported any (traumatic) experiences with electric shocks, presence of any kind of psychiatric disorder, or being under the influence of drugs were not tested.
The CS-US contingency-awareness was assessed to test whether our participants learned to differentiate between the CSs. Learning this differentiation is crucial to comparing fear responses to the CSs. Therefore, participants that gave incorrect answers on the Exit-questionnaire (e.g.: reporting that the CS- was followed
8
by the US) or reported incorrect US-expectancies during the trials of the test phase (e.g.: when the CS- was expected to be followed by a shock), were replaced.
After replacing 16 contingency- unaware participants, a total of 38 participants (20 in the Instructed group and 18 in the Experience group) were used for data analysis. (20 women, mean age = 23.0 years, age range: 18-30 years. During further exploration, the main results did not change whether the CS-US unaware participants were included in the analysis or not.
Materials and Apparatus
The technical descriptions of the used electric stimulus, the assessment of the startle reflex and of the SCR are based on Soeters & Kindt (2011). The experiment was approved by the ethical commission of the University of Amsterdam (EC number: 2014-CP-3887).
Stimuli.
Pictures of a square and a cylinder (see Figure 1) depicted from four different viewpoints served as CS+ or CS-. Assignment of the cube or cylinder to the CS+ or the CS- was counterbalanced across participants. Each CS was presented in the middle of a black background on a computer screen.
Electric shocks that served as US were applied for 2 ms to the wrist of the non-preferred hand via a pair of electrodes and a conduct gel between electrodes and skin. Level of shocks was determined in a work-up procedure wherein shock intensity was increased till participants considered it to be “uncomfortable, but not painful” (Soeters & Kindt, 2011).
A picture of a curtain was used to cover the pictures of the CSs (See Figure 1), so that participants had to guess which CS was hidden behind the curtain.
Figure 1. Example of the square (left) and the cylinder (middle) that either served as
9 Startle-reflex.
The participants’ eye blink-startle-reflex to a loud noise (40 ms; 104 dB) was assessed by presenting a noise binaurally through headphones after each CS during the conditioning trials (Headphone model MD-4600; Compact Disc Digital Audio, Monacor). The startle potentiation was accessed by two 7-mm Ag/AgCl electrodes that were filled with electrolyte gel and positioned 1 cm under the pupil and 1 cm below the lateral canthus. A ground reference was placed on the forehead. The eye blink EMG activity was measured with a front-end amplifier. Any noise was removed with a filter that was set at 50 Hz. Integration was done via a true-RMS converter. Responses were measured within 0 and 250 ms from the administration of the probe. For data analysis, the means of the square roots for all the trials in each phase were computed.
Skin conductance response.
Electrodermal activity was assessed by an input device with a sine-shaped excitation voltage (+0.5 V) of 50 Hz. Two Ag/AgCl electrodes of 20 × 16 mm were connected with the input device (Soeters & Kindt, 2011). The electrodes were connected to the first and third fingers of the non-preferred hand.
Determination of SCRs elicited by the CS was assessed by comparing the baseline (i.e., 1 sec before the CS was presented) to peak responses within 1–7 sec after the onset of the stimulus (Effting & Kindt, 2007). For data analysis, the absolute value of the logarithm of the SCRs was computed and the algebraic signs of the original values were transmitted to the new values (Milad et al., 2005).
Subjective experience.
The participants had to report if they expected the US during each conditioning trial on an 11-point Likert-scale by moving the cursor of a slider. The scale was made of different coloured LEDs that represented the cursor position and was labelled from “certainly no electric stimulus” (-5) to “a certain electric stimulus” and responses in between (“uncertain”) were scored as 0.
10 Avoidance tendencies.
The following description of the AAT is based on Krypotos et al. (2014). Two blocks of 4 practice trials and 16 test trials each were used for the AAT. Two of the 4 viewpoints of the stimuli were presented during the practice trials. During the 16 test trials, the four geometrical presentations of CS+ and CS- were tested in random order. Hereby, a white manikin figure (71 mm × 95 mm) appeared either on the bottom or top half of a black computer screen. The picture of the CS with the accompanying frame was presented after 1500 ms on the opposite side of the screen. Participants had to react as quickly and accurately as possible to the presentation of the CSs. After the first block, instructions were inverted. The manikin could be moved upwards by pressing the “Y”-key (marked with an upward-arrow-sticker) or downwards by pressing the “B” key (marked with a downward-arrow-sticker) on a standard computer keyboard. Reaction times between CS-presentation and the emitted response were used as dependent variables.
Self-report measures.
US- pleasantness ratings were assessed on 11-point Likert scales that ranged from -5 (‘did not like the shock at all’) to +5 (‘liked the shock very much’) by asking participants “How pleasant was the shock for you?”
Fear ratings participants were recorded on 11-point Likert scales that ranged from -5 (‘not fearful at all’) to +5 (‘very fearful’) by asking “How fearful is this stimulus for you?”
Participants were asked to report their gender, birth date, and hand- preference. Also, evaluative fear ratings of pleasantness, intensity, startlingness of the US, and valence of both CSs and US were assessed on 11-point Likert-scales that ranged from -5 (indicating low levels) to +5 (indicating high levels). Motivation to complete tasks and questionnaires were assessed on millimetre scales that ranged from 0 mm (indicating lowest motivation) to 100 mm (indicating highest motivation).
State anxiety and trait anxiety were assessed with the Dutch translation of the State-Trait Anxiety Inventory (STAI) wherein fear in different situations (state anxiety) and as a personality trait (trait anxiety) is assessed on a 4-point Likert scale (Spielberger et al., 1970; Van der Ploeg, 2000).
Tendencies to react fearful to anxiety were assessed with the Dutch version of the Anxiety Sensitivity Index (ASI; Peterson & Reiss, 1993).
11 Contingency awareness questions.
To assess whether participants were contingency aware, each participant was asked to point out which CS had been followed and which one had not been followed by a shock.
Design
Participants were randomly assigned to either the direct conditioning or the instructional learning group. Participants in the Experience group sometimes received a shock after an initially neutral stimulus presentation (CS+; e.g. a picture of a cube) and received no shock after the presentation of another neutral stimulus (CS-; e.g., a picture of a cylinder).
The experiment consisted of three phases for all participants: the habituation-, the acquisition- and the test-phase (see Figure 2). There were 2 trials in the habituation-phase, 6 in the acquisition-phase and 2 in the test-phase for each CS.
Figure 2. Procedure of conditioning trials in habituation, conditioning and test
phase.
Procedure
First, participants were asked to read and sign the informed consent brochure. Their state anxiety was assessed. After that, the SCR-, EMG- and US-electrodes were attached to the participants and headphones were placed. The individual level of the US was determined. Participants then received instructions about the task (see Figure 3 for a complete picture).
2 CS+ vs. 2 CS -Habituation 6 CS + vs. 6 CS -Acquisition 2 CS + vs. 2 CS -Test
12
The following measures were obtained across all conditioning phases: the expectancy ratings of the US had to be given within 7 seconds after each CS-presentation, just before the startle probe was administered. The startle probes were administrated 7 seconds after each CS- presentation. Inter-trial intervals (ITIs) varied quasi-randomly between 15, 20 or 25 seconds with a mean of 20 seconds and no more than two consecutive trials taking place.
Figure 3. Picture of the complete procedure of the experiment in chronological order.
Habituation phase
During the habituation phase, participants in both groups were exposed twice, without reinforcement, to two random selected viewpoints of the CS+ and the CS- that were revealed behind a picture of a curtain.
Acquisition phase
Next, participants in the Instructed group saw a picture of the curtain (see Figure 1) twelve times, 6 times for each CS. It was followed by a shock five times if the CS+ was hidden behind the curtain. The curtain was visible for ten seconds during each trial in the Instructed group and did not open during this phase. This procedure was used to test whether participants learn about the contingency of CSs that was hidden behind the curtain and the US only with the instructions that were given after the acquisition. In the Experience group, participants saw the picture of the curtain twelve times but, as opposed to the picture in the Instructed group, it did open up after two seconds and the CS behind the curtain was revealed. Participants were exposed to six presentations of the CS+ of which five were followed by the US.
STAI-S
Install SCR, EMG andUS- electrodes Acquisition
In stru cti on s o r B re ak Hab ituat ion
Test Disconnect SCR, EMG
and US- electrodes
AAT Fear - ratings Pleasant ness -ratings ASI STAI-T
Evaluative ratings of pleasantness, intensity, startlingness of the US and
13 Test phase
During the test phase, the participants in the Instructed group were instructed that they only received an electric shock during the previous phase when a specific CS (either the cube or the cylinder) was hidden behind the curtain, and that they did not receive an electric shock when the other CS was presented behind the curtain. Participants in the Experience group were merely instructed onscreen to take a one-minute break but did not receive any task-relevant instructions.
Next, as in the habituation phase, participants in both groups were exposed twice to both previously learned stimuli without receiving any shocks.
Finally, all electrodes and the headphones were disconnected.
AAT phase
Then, all participants were asked to complete the AAT. Participants received the instruction that they had to react as fast and as accurate as possible to the horizontal or vertical frame. During the first block, participants were instructed to either approach landscape-orientated pictures by moving the manikin towards the picture and to avoid portrait-orientated pictures by moving the manikin away from the picture, or vice versa (see Figure 4). Four practice trials were performed prior to the test block to ensure that participants had understood the instructions.
Then in the second block, the instructions were inverted. Four practice trials were performed prior to this test block as well.
The block order was counterbalanced across participants. The manikin disappeared 500 ms after it was moved for the first time. Incorrect responses were followed by a red cross. The inter-trial intervals were set at 2000 ms.
14
Figure 4. Example of sequences of avoidance trial (above) and approach trial
(below) as seen by participants during the AAT.
Self-reports phase
After the completion of the AAT, participants were asked, onscreen, to assess fear and pleasantness ratings for each of the CS+ - and the CS- - viewpoints. At the end of the experiment, participants were asked to fill in the ASI and the STAI-T. Also, evaluative fear ratings of pleasantness, intensity, startlingness of the US and valence (of both, CSs and US) were assessed.
Debriefing
After completion of the experiment participants were debriefed and rewarded.
Results
Data analysis
For all our analysis the alpha level was set to .05. To assess possible differences between groups (in levels of motivation, US-valences, gender, age, hand preference and the levels of the US), a one-way ANOVA was conducted for all participants.
Independent t-tests were conducted to assess group differences in the STAI-S, STAI-T and the ASI.
1. 2. 3.
15
For each type of fear response the means of all measures of fear in habituation, acquisition, and test phase were computed separately for each CS. Differences in all measures of fear responses were assessed with 2 (Group: direct conditioning vs. instructional learning) × 2 (Stimulus: CS+ vs. CS-) repeated-measures Analysis of Variances (ANOVAs). Differences in startle potentiation, SCRs and US-expectancies were analysed with different 2 (Group: direct conditioning vs. instructional learning) × 2 (Stimulus: CS+ vs. CS-) ANOVAs for each phase (Habituation, Acquisition and Test) with stimulus as within-subject factor and group as between-subject factor.
For the AAT, median reaction times were computed for each CS (CS+ vs. CS-) × Response (approach vs. avoidance). Then, reaction times were analysed in a 2 (Stimulus: CS+ vs. CS-) × 2 (Response: approach vs. avoidance) × 2 (Group: direct conditioning vs. instructional learning) ANOVA whereby stimuli and response-type served as within-subject factors and group as between-subject factor.
The Bayesian analysis was done using the BayesFactor (Morey et al., 2014) package for RStudio (2012) with scripts written by Tessa Blanken and Angelos- Miltiadis Krypotos. The main effects were assessed by comparing a model with one main effect to a model with both main effects. Interaction effects were analysed by comparing a model of both main effects and interactions to a model of only main effects. The Bayes factors for each main effects and interactions were computed. While larger Bayes factors represented evidence that the results were more likely to occur if the alternative hypothesis was right, small Bayes factors indicated evidence for the nul-hypothesis. Jeffrey’s (1961: 432) interpretations of Bayes factors (BF10)
were used to interpret evidence (see Table 1).
Table 1
Jeffrey’s Interpretations of Bayesian Factors (k) in Terms of Strength of Evidence for Alternative Hypothesis (H1) and Nul-Hypothesis (H0) (adapted from Jeffrey, 1961).
Bayes Factor (k) Strength of Evidence
> 100 Extreme evidence for H1
30 – 100 Very strong evidence for H1
10 – 30 Strong evidence for H1
16
1 – 3 Anecdotal evidence for H1
1 No evidence
1/3 – 1 1/3 – 1/10
Anecdotal evidence for H0 Moderate evidence for H0
1/10 – 1/30 Strong evidence for H0
1/30 – 1/100 Very strong evidence for H0
< 1/100 Extreme evidence for H0
Manipulation and group randomization check
The results of the manipulation and group randomization analysis are shown in Table 2. No differences between the groups were found in terms of hand-preference, US- pleasantness, US-intensity, US-startlingness ratings, or motivation to complete questionnaires and the computer task, or the level of the US.
Levels of state- anxiety, trait- anxiety and anxiety sensitivity were indifferent in both groups.
Table 2
Assessment of Group Differences by Means, Group Size (N) and Standard Deviations (Std. Dev.) for the Instructions (1) and Experience (2) Group and Degrees of Freedom (df), F-values, Significance Levels and Effect Sizes (ηp²) of the ANOVA. Values
marked with “*”represent t-test results.
Measure Group Mean Std.
Dev. F p ηp² Hand-preference Instructed 1.15 .37 Experience 1.17 .38 0.019 .892 .001 US-pleasantness Instructed 3.75 2.82 Experience 4.28 2.51 0.366 .549 .010 US-intensity Instructed 2.95 0.39 Experience 2.83 0.51 0.623 .435 .017 US-startlingness ratings Instructed 3.60 0.68 Experience 3.50 0.71
17 0.197 .660 .005 Questionnaire- motivation Instructed 8.56 1.13 Experience 9.00 1.03 1.596 .215 .042 Computer-task- motivation Instructed 8.76 0.96 Experience 8.76 1.46 0.000 .998 .000 Age Instructed 22.63 3.13 Experience 23.42 3.37 0.556 .461 .015 Gender Instructed 1.55 0.51 Experience 1.50 0.51 0.090 .766 .002 Level of electric stimulus (mA) Instructed 15.77 6.31 Experience 14.28 6.24 State anxiety Instructed
32.35 7.94 0.536 .469 .015 Trait anxiety Anxiety sensitivity Experience Instructed Experience Instructed Experience 32.11 35.75 36.22 9.45 11.56 1.74 8.66 7.95 3.52 6.46 0.096* -0.174* -1.266* .942* .863* .214* .000 .001 .043 US- expectancies
Figure 5 shows the mean expectancies for each CS, across phases, for each group. During the habituation-phase, participants in both groups reported that they expected the US almost equally after both CSs. No main effect of CS was revealed, F(1, 36) < 1; BF10 = 0.23. This result did not differ between groups, F(1, 36) = 1.046,
p = .313, ηp² = .028; BF10 = 0.49.
During fear acquisition, participants reported different expectations of the US to the CSs as indicated by a CS × Group interaction, F(1, 36) = 9.96, p = .003, ηp² =
.217; BF10 = 12.13. An exploratory analysis was conducted to assess these results per
group: while participants in the Instructed group reported similar expectancies of the US, F(1, 19) < 1, participants in the Experience group reported higher expectancies of the US in case of a CS+ -rather than of a CS-- trial, F(1, 17) = 12.84, p = .002, ηp² =
.430. In line with our hypothesis, it can be suggested that participants in the Instructed group could not predict when the US would occur because the CSs were hidden
18
behind the curtain, while participants in the Experience group had already learned about its relationship.
In the test phase, participants in both groups reported higher expectations of the US after they saw the CS+ than after CS-. A main effect of CS was revealed, F(1, 36) =6.591, p = .015, ηp² = .155; BF10 = 7.87. This result indicated that participants
learned to expect the US after CS+ but not after the CS-. Learning took place in both groups, no interaction effect between type of CS and group was found, F(1, 36) < 1; BF10 = 0.31. In an exploratory analysis it was found that US-expectancies of
participants in the Instructed group differed between the CSs, F(1, 19) = 8.997, p = .007, ηp² = .321. Noticeably participants in the Experience group did not report
different expectancies of the US to the CSs, F(1, 17) = 1.884, p = .188, ηp² = .100.
Figure 5. Mean US-expectancies per CS in the Instructed (left panel) and Experience
group (right panel) during the habituation-, acquisition- and test-phases. Error bars represent standard errors.
Startle- reflex potentiation
Figure 6 shows mean EMG-responses for each CS, in each phase, across groups. During the habituation phase participants did not show differences in startle potentiation to the CSs, as revealed by a main effect of the CS, F(1, 36) < 1; BF10 =
0.32. Also, no interaction effect between CS and group was found, F(1, 36) = 2.58, p = .117, ηp² = .067; BF10 = 0.86, indicating that the baseline startle reflexes were
comparable between groups because participants could did not differentiate between the CSs.
19
During the fear acquisition phase, participants still did not show differences in EMG- responses to the CSs, F(1, 36) < 1, BF10 = 0.29.
A trend was visible in the CS × group interaction, F(1, 36) = 3.862, p = .057, ηp² = .097; BF10 = 1.28. It can be suggested that participants (at least in the
Experience group) started to learn to differentiate between the CSs. This presumption was supported by an exploratory analysis that revealed that participants in the Experience group had different startles to the CSs, F(1, 17) = 6.851, p = .018, ηp² =
.287, while participants in the Instructed group did not show different startles to the CSs, F(1, 19) < 1.
During the test phase, participants did not show different EMGs to the CSs, as a main effect of the CS was revealed, F(1, 36) < 1; BF10 = 0.24. No CS × group
interaction was found either, F(1, 36) < 1,; BF10 = 0.36; participants had similar EMG
responses for both CSs in both groups. The result that startles did not differ to the CSs was not in line with our hypothesis, because higher EMG responses to the CS+ than to the CS- were expected.
Figure 6. EMG-measures per CS in the Instructed (left panel) and Experience group
(right panel) during the habituation-, acquisition- and test-phases. Error bars represent standard errors.
SCRs
One participant in the Instructed group was excluded from the SCR data analysis due to a broken electrode. Figure 7 shows the mean SCRs for each CS, in each phase, across the two groups.
20
During the habituation phase, participants had similar SCRs to CS+ and CS-, no main effect of the type of stimuli was revealed, F(1, 35) = 2.29, p = .140, ηp² =
.061; BF10 = 0.49. Also, no differences in SCRs between groups were found, that is no
CS × group interaction was revealed, F(1, 35) < 1; BF10 = 0.39. This indicates that no
baseline differences existed between the SCRs and the CSs, amongst groups, during habituation.
In the acquisition phase, participants had different SCRs to the CSs in both groups as indicated by a Group × CS interaction, F(1, 35) = 17.667, p < .001, ηp² =
.335; extreme evidence for differences was revealed in Bayesian analysis, BF10 =
104.24 In an exploratory analysis, the results were checked per group. While participants in the Instructed group had similar SCRs to the CSs, F(1, 19) = 1.160, p = .296, ηp² = .061, participants in the Experience group had higher SCRs to the CSs,
F(1, 17) = 19.899, p < .001, ηp² = .539. It can be suggested that knowing about the
CS-US contingency influenced SCRs when fear was learned via experience.
During test phase, participants had similar SCRs to both CSs, as a main effect of CS revealed, F(1, 35) < 1; BF10 = 0.33. Participants in both groups had similar
SCRs to the CSs, no Group × CS interaction was revealed, F(1, 35) < 1; BF10 = 0.34.
Unexpectedly, SCRs did not differ to the CSs.
Figure 7. Mean Skin Conductance Responses per CS in the instructed (left panel) and
experience (right panel) group during habituation, acquisition and test phase. Error bars represent standard errors.
21 Fear ratings
Figure 8 shows the mean fear ratings for each CS across groups. After the test phase participants gave higher fears ratings for the CS+ than for the CS- as revealed by the main effect of type of CS, F(1, 36) = 94.57, p < .001, ηp² = .72; BF10 =
1.749784e+12. Participants in both groups reported similar levels of fear to the two CSs. Although a trend for higher fear ratings to the CS+ in the Experience group was visible, no CS × group- interaction was revealed, F(1, 36) = 2.307, p = .138, ηp² =
.060. Contrarily, Bayesian data analysis revealed anecdotal evidence for differences between groups, BF10 = 1.80.
Figure 8. Mean scores for the fear ratings per CS in the Instructed (left panel) and
Experience group (right panel). Error bars represent standard errors.
CS- pleasantness ratings
Figure 9 shows the mean pleasantness ratings for each CS across the two groups. After the test phase, participants reported lower levels of pleasantness of the CS+ than of the CS-; a main effect of type of CS was revealed, F(1, 36) =22.627, p < .001, ηp² = .386; BF10 = 29982.18. Participants in both groups reported similar levels
of pleasantness toward the CSs, no interaction effect between type of CS and group was found, F(1, 36) = 3.033, p = .090, ηp² = .078; BF10 = 0.87.
22
Figure 9. Mean Pleasance ratings per CS in the Instructed (left panel) and
Experience group (right panel). Error bars represent standard errors.
Approach- and avoidance task
Figure 10 shows median reaction times per task, CS and group. CS+ and CS -participants responded equally fast during both tasks, indicated by no CS × Task- interaction, F(1, 36) < 1, p = .836, ηp² = .001; BF10 = 0.35. No interaction effect
between types of CS, type of task and group was revealed, F(1, 36) < 1; BF10 = 0.40.
Unexpectedly, it can be suggested that participants did not show avoidance behavior when they were confronted with the CS+.
Figure 10. Median reaction times of the AAT in milliseconds for approaching and
avoiding as seen in Instructed (left panel) and Experience group (right panel). Error bars represent standard errors.
23
General discussion
The results of the present study demonstrated that instructions can play a major role in the acquisition of fear. Language seemed to be as powerful as aversive experiences in instilling fear. In line with previous studies (Raes et al., 2014; Field and Storksen-Coulson, 2007), instructions about the CS-US contingency increased self-reported fear beliefs just as direct aversive experiences did. Expectancies of the US after the CS+ were higher than the CS- after participants had learned about the contingencies, either via instructions or via direct experiences. Contrary to previously conducted experiments (Olsson & Phelps, 2004; Raes et al., 2014), we failed to find any differences in physiological measures of fear responses (such as SCRs, startle potentiation to the stimuli and avoidance tendencies).
Still, instructions influenced the magnitude of fear responses as experience did; no differences in startle potentiation, SCRs and avoidance behaviour were revealed between the two groups. This result supports Rachman’s (1977) 3- pathway model and underlines the role that language can play in the acquisition of anxiety disorders.
The Bayesian approach enhanced confidence in our results and affirmed evidence for the nul-hypotheses of no differences. Bayesian results were mainly in line with frequential results. Interestingly, anecdotal evidence was found for differences between groups in fear ratings while no differences were revealed in frequential statistics. All other findings were in line with the results of the frequential analysis. The high variability of SCRs might have diminished the power to compare groups by this measure as in other experiments (Ahmed & Lovibond, 2015).
The lack of differences in avoidance tendencies was not in line with Chen & Bargh’s (1999) finding that reaction times are slower when the movement is congruent with the valence of the stimuli. Due to the failure to replicate this finding, as in another experiment (Rotteveel et al., 1999), the link between avoidance tendencies and fear learning in the AAT needs to be considered further.
Extinction during and after the test phase potentially decreased differences in physiological fear responses and avoidance tendencies to the stimuli. To rule out the possibility that the unreinforced trials during test phases triggered extinction, an exploratory analysis of expectancy ratings, EMGs and SCRs was conducted wherein only the results of the first trial of the test phase were analyzed. This analysis revealed that that the major results of all measures were robust when compared to the means of
24
the test phase. It can be suggested that extinction during test phase did not influence the results. Still, it remains unclear whether extinction after test phase and during the AAT influenced avoidance tendencies. For future research we suggest the minimization of distraction of participants by keeping the breaks between different tasks as short as possible. Additionally, on-screen instructions might reduce potential extinction, because contact with the researcher could potentially alter emotional processing.
A potential limitation of the experiment concerns the continues use of startle tones after each CS- presentation (which were more irritating than the electric shock according to verbal comments of some participants). Therefore the startle tone could have functioned as another US rather than a CS that somehow inhibited the effects of the electric shock. Other experiments showed that startle tones can be successfully used as US (Brunzell & Kim, 2001; Neumann & Waters, 2006) indicating that both CS+ and CS- could have been reinforced so that differential learning might have been inhibited, as seen in the lack of differences in physiological fear responses and avoidance tendencies to the CSs. We suggest the assessment of fear responses in future research at the end of each phase twice, rather than after each CS to prevent the conditioning effects of the tone.
In conclusion, our results suggest that fear can be acquired both via instruction as well as via experience. Although discriminative fear learning was not observed in all measures of fear responses, language can initiate the acquisition of fears. Therefore, it is necessary to focus on language as a potential factor in installing fears that could ultimately lead to anxiety disorders. By understanding the mechanisms that play a role in the acquisition of fears we can ultimately develop models that could help to prevent such anxiety disorders. Further studies are needed to assess languages positive functions, such as potential preventing and relearning of fear.
Acknowledgements
I would like to thank all the people who helped and supported me with writing this research: Angelos Miltiadis-Krypotos, Tessa Blanken & the participants in the experiment. This research was funded by VIDI grant (#452-09-001) from NWO appointed to Tom Beckers.
25
References
Ahmed, O., & Lovibond, P. F. (2015). The Impact of Instructions on Generalization of Conditioned Fear in Humans. Behavior Therapy.
doi:10.1016/j.beth.2014.12.007
American Psychiatric Association (2013. Diagnostic and Statistical Manual of Mental Disorders, 5th Edition. doi:10.1176/appi.books.9780890425596.893619 Bandura, A. (1965). Vicarious processes: A case of no-trial learning. Advances in
Experimental Social Psychology, 2, 1-55. doi:10.1016/s0065-601(08)60102-1
Beckers, T., Krypotos, A. M., Boddez, Y., Effting, M., & Kindt, M. (2013). What's wrong with fear conditioning?. Biological Psychology, 92, 90-96.
doi:10.1016/j.biopsycho.2011.12.015
Brunzell, D. H., & Kim, J. J. (2001). Fear conditioning to tone, but not to context, is attenuated by lesions of the insular cortex and posterior extension of the intralaminar complex in rats. Behavioral neuroscience, 115, 365-375. doi:10.1037/0735-7044.115.2.365
Buck, R. (1999). The biological affects: A typology. Psychological Review, 106, 301- 336. doi:10.1037/0033-295x.106.2.301
Chen, M., & Bargh, J. A. (1999). Consequences of automatic evaluation: Immediate behavioral predispositions to approach or avoid the stimulus. Personality and Social Psychology Bulletin, 25, 215-224.
doi:10.1177/0146167299025002007
Cook, E., & Turpin, G. (1997). Differentiating orienting, startle, and defense responses: The role of affect and its implications for psychopathology. Attention and orienting: Sensory and motivational processes, 137-164.
De Graaf, R., Ten Have, M., & Van Dorsselaer, S. (2010). De psychische gezondheid van de Nederlandse bevolking: NEMESIS II: Opzet en de eerste resultaten [The Mental Health of the Dutch Population: Netherlands Mental Health Epidemiological and Incidence study NEMESIS II: Design and First Results]. Utrecht, Netherlands: Trimbos Institute.
Ekeberg, Ø., Kjeldsen, S. E., Greenwood, D. T., & Enger, E. (1990). Correlations between psychological and physiological responses to acute flight phobia
26
stress. Scandinavian Journal of Clinical & Laboratory Investigation, 50, 671-677. doi:10.3109/00365519009089186
Effting, M., & Kindt, M. (2007). Contextual control of human fear associations in a renewal paradigm. Behaviour Research and Therapy, 45, 2002-2018. doi:10.1016/j.brat.2007.02.011
Faul, F., & Erdfelder, E. (1992). GPOWER: A priori, post-hoc, and compromise power analyses for MS-DOS [Computer program]. Bonn, FRG: Bonn University, Department of Psychology.
Field, A. P., & Storksen-Coulson, H. (2007). The interaction of pathways to fear in childhood anxiety: A preliminary study. Behaviour Research and Therapy, 45, 3051-3059. doi:10.1016/j.brat.2007.09.001
Funayama, E. S., Grillon, C., Davis, M., & Phelps, E. A. (2001). A double dissociation in the affective modulation of startle in humans: effects of unilateral temporal lobectomy. Journal of Cognitive Neuroscience, 13, 721-729. doi:10.1162/08989290152541395
Hamm, A. O., & Vaitl, D. (1996). Affective learning: Awareness and aversion. Psychophysiology, 33, 698-710. doi:10.1111/j.1469-8986.1996.tb02366.x Jeffreys, H. (1961). Theory of probability. Oxford, UK: Oxford University Press.
doi:10.1093/gji.6.4.555
Krypotos, A. M., Effting, M., Arnaudova, I., Kindt, M., & Beckers, T. (2014). Avoided by Association Acquisition, Extinction, and Renewal of Avoidance Tendencies Toward Conditioned Fear Stimuli. Clinical Psychological Science, 2, 336-343. doi:10.1177/2167702613503139
Lang, P. J., Bradeley, M. M., & Cuthbert, B. N. (1997). Motivated attention: Affect, activation, and action. In P. J. Lang, R. F. Simons, & M. T. Balaban (Eds.), Attention and orienting: Sensory and motivational processes (pp. 137–164). Mahwah, NJ: Erlbaum. doi:10.4324/9780203726457
Levenson, R. W. (2003). Blood, sweat, and fears. Annals of the New York Academy of Sciences, 1000, 348-366. doi:10.1196/annals.1280.016
Mauss, I. B., Levenson, R. W., McCarter, L., Wilhelm, F. H., & Gross, J. J. (2005). The tie that binds? Coherence among emotion experience, behavior, and physiology. Emotion, 5, 175- 190. doi:10.1037/1528-3542.5.2.175
Mauss, I. B., & Robinson, M. D. (2009). Measures of emotion: A review. Cognition and Emotion, 23, 209-237. doi:10.1080/02699930802204677
27
McCleary, R. A. (1950). The nature of the galvanic skin response. Psychological Bulletin, 47, 97-116. doi:10.1037/h0059810
Milad, M. R., Orr, S. P., Pitman, R. K., & Rauch, S. L. (2005). Context modulation of memory for fear extinction in humans. Psychophysiology, 42, 456-464.
doi:10.1111/j.1469-8986.2005.00302.x
Morey, R. D., Rouder, J. N., Jamil, T., & Morey, M. R. D. (2014). Package ‘BayesFactor’.
Morris, J. S., Öhman, A., & Dolan, R. J. (1998). Conscious and unconscious emotional learning in the human amygdala. Nature, 393, 467-470. doi:10.1038/30976
Mowrer, O. H. (1939). A stimulus-response analysis of anxiety and its role as a reinforcing agent. Psychological Review, 46, 553-565. doi:10.1037/h0054288 Neumann, D. L., & Waters, A. M. (2006). The use of an unpleasant sound as an
unconditional stimulus in a human aversive Pavlovian conditioning procedure. Biological psychology, 73, 175-185. doi:10.1016/j.biopsycho.2006.03.004 Nousi, A., Haringsma, R., van Gerwen, L. J., & Spinhoven, P. (2008). Different flying
histories in flying phobics: Association with psychopathology and treatment outcome. Aviation, Space, and Environmental Medicine, 79, 953-959. doi:10.3357/asem.2189.2008
Olsson, A., & Phelps, E. A. (2004). Learned fear of “unseen” faces after Pavlovian, observational, and instructed fear. Psychological Science, 15, 822-828. doi:10.1111/j.0956-7976.2004.00762.x
Olsson, A., & Phelps, E. A. (2007). Social learning of fear. Nature neuroscience, 10, 1095-1102. doi:10.1038/nn1968
Pavlov, I. P. (1927). Conditioned reflexes. An Investigation of the Physiological Activity of the Cerebral Cortex (pp. 337-351). London.
doi:10.1017/s0031819100028242
Peterson, R. A., & Reiss, S. (1993). Anxiety sensitivity index revised test manual. IDS Publ.. doi:10.1037/t00033-000
Pineles, S. L., Orr, M. R., & Orr, S. P. (2009). An alternative scoring method for skin conductance responding in a differential fear conditioning paradigm with a long‐duration conditioned stimulus. Psychophysiology, 46, 984-995. doi:10.1111/j.1469-8986.2009.00852.x
28
Rachman, S. (1977). The conditioning theory of fear acquisition: A critical examination. Behaviour Research and Therapy, 15, 375-387. doi:10.1016/0005-7967(77)90041-9
Rachman, S. (1991). Neo-conditioning and the classical theory of fear acquisition. Clinical Psychology Review, 11, 155–173.
doi:10.1016/0272-7358(91)90093-a
Raes, A. K., De Houwer, J., De Schryver, M., Brass, M., & Kalisch, R. (2014). Do CS-US pairings actually matter? A within-subject comparison of instructed fear conditioning with and without actual CS-US pairings. PloSOne, 9, e84888. doi:10.1371/journal.pone.0084888
Rotteveel, M., Gierholz, A., Koch, G., Van Aalst, C., Pinto, Y., Matzke, D., Steingroever, H.,Verhagen, J., Beek, T. F., Selker, R., Sasiadek, A. &
Wagenmakers, E. (2015). On the automatic link between affect and tendencies to approach and avoid: Chen and Bargh (1999) revisited.. Front. Psychol., 6, 335-363. doi: 10.3389/fpsyg.2015.00335
Rouder, J. N., Morey, R. D., Speckman, P. L., & Province, J. M. (2012). Default Bayes factors for ANOVA designs. Journal of Mathematical Psychology, 56, 356-374. doi:10.1016/j.jmp.2012.08.001
RStudio (2012). RStudio: Integrated Development Environment for R [Computer Software]. RStudio: Boston, MA.)
Soeter, M., & Kindt, M. (2011). Disrupting reconsolidation: pharmacological and behavioral manipulations. Learning & Memory, 18, 357-366.
doi:10.1101/lm.2148511
Spielberger, C. D., Gorsuch, R. L., & Lushene, R. E. (1970). Manual for the state-trait anxiety inventory.
Van der Ploeg, H.M. (2000). Handleiding bij de Zelf-Beoordelings Vragenlijst. Amsterdam: Harcourt Assessment B.V.
Weike, A. I., Schupp, H. T., & Hamm, A. O. (2007). Fear acquisition requires awareness in trace but not delay conditioning. Psychophysiology, 44, 170-180. doi:10.1111/j.1469-8986.2006.00469.x
Yeomans, J. S., Li, L., Scott, B. W., & Frankland, P. W. (2002). Tactile, acoustic and vestibular systems sum to elicit the startle reflex. Neuroscience &
