Examining the Effectiveness of Observed & Derived Pathways of Avoidance

Abstract

This study compared aversive learning through solely social observational learning

and derived avoidance learning using a recorded learning model video where they receiving a

mild electric shock when a selected circle (CS+) is shown on-screen. The stimulus will

predict the delivery of shock (CS+) whilst another contextually similar stimulus will present

no negative consequences (CS-). Participants are instructed of an avoidance behaviour

response (AV+).The inferred condition participants will be presented with RCP trials of non-

arbitrary and arbitrary training and testing of Same and Opposite relations, as shown in

Dymond et al., (2008) study, to test for transformation of functions and derived relational

responding in later trials i.e. avoidance acquisition and extinction.

The between- subject factor Group has two levels; observed and inferred. Two within-

subject factors (1) Conditioned Stimulus, which has two levels: CS+ and CS-. (2) Learned

Behaviour which has two levels: Avoidance (AV) and Non-Avoidance (NAV). Thus

combination of the three factors created six experimental conditions. The DerivedCS+ will be

compared against the inferred response rates of CS+ and CS-. There were three dependent

variables: Avoidance behaviour responses of the shock in each phase, expectancy of shock

taken from Likert scales at the end of each phase, and AAQ-II results.

Comparing Observed EXT CS+ and CS- and Inferred EXT CS+ and CS-; no

significant differences were found between the learning groups and conditioned stimuli,

showing the two learning groups were just as effective as each other in producing a fear-

evoked avoidance response. In a 2*2*2 multi-way mixed ANOVA, results showed

significant main effects for conditioned stimuli, learned behaviour and learning group.

Comparing group against conditioned stimuli and learned behaviour (Figure 8 and 9) it is

clear that Observed rated higher in expectancy of shock to both, this could be attributed to the

1

observed group understanding the CS-US contingency and the value of CS- more than the

Inferred group.

In conclusion, the study found that observed group may have been taught through

direct means, by watching the learning model in the avoidance acquisition phase. Future

studies would amend this. The inferred condition was successful as derived relational

responding was produced.

2

A common debate in the topic of fear learning is how fear is firstly acquired, be it

through direct or indirect means. Fear acquisitions to environmental stimuli are fundamental

to the fight or flight mechanism, thus essential for survival of human and non-humans alike.

Understanding fear conditioning will develop and strengthen the understanding of phobias

and anxiety disorders to help build clinical models (Barot, Chung, Kim & Bernstein, 2009;

Field, 2006).

The first model of classical fear conditioning was coined Pavlovian conditioning and

suggested fear acquisition of an object can only surface through first-hand experience with an

aversive stimulus (Olsson, Nearing & Phelps, 2007; Siddle & Bond, 1988). The individual,

be it human or non-human, acquires a conditioned fear response (CR) to a formerly neutral

stimulus (CS) through direct experience pairing with an aversive occurrence, the

unconditioned stimulus (US). This acquisition was thought to be maintained by a two-factor

theory i.e. the escape from the CS and instant decrease in fear eliciting factors. This model is

famously illustrated in the ‘Little Albert’ study (Watson & Rayner, 1920) where they

conditioned a fear response to a previously neutral CS i.e. white rabbit/rat/dog, by pairing

CS-US (loud noise). Results showed Little Albert generalised and elicited a fear response not

only to white rabbit etc., but also Santa Claus and white hair.

Research has shown that the amygdala is essential in the acquisition, storage and

expression of the conditioned fear (Davis & Whalen, 2001; Kim & Jung, 2006; Olsson et al.,

2007). Barot, Kyono, Clark & Bernstein’s (2008) study showed clear visualisations of

neurons within the amygdala receiving CS-US convergence information, during acquisition

phase of a learned association. Barot et al., (2009) study showed the interaction of the

amygdala and the hippocampus, where the hippocampus processes and transmits contextual

CS information to the amygdala i.e. hippocampus is responsible for contextual learning

whilst amygdala responsible for cue learning (Olsson & Phelps, 2007; Phillips & LeDoux,

3

1992). Damage to these areas is coupled with reports of CS-US contingency learning but

inhibition of autonomic response (LaBar, LeDoux, Spencer & Phelps, 1995). Delgado, Jou,

Ledoux & Phelps (2009) also illustrated the role of the striatum due to its reward-related

processing ability to rule emotions and avoid negative outcomes. Participants were instructed

upon seeing the aversive stimulus they could avoid shock by choosing the correct action

(AV+). This depicts a mechanism of active coping, where the striatum exerts control over the

emotional responses to influence the decision making (Delgado, Li, Schiller & Phelps, 2008;

LeDoux & Gorman, 2001; Schultz, 2006). These studies clearly illustrate the networks of

learning paradigms, storage and reinforcer types.

However, over the years research grew and the understanding that fear can only be

acquired through direct means was soon diminished. The Pavlovian Model was suggested to

be extended to consider indirect forms. Field (2006) suggested two pathways of avoidance;

(1) vicarious (observational) learning and (2) verbal information i.e. instructions. Another

issue with the Pavlovian Model was the two-factor theory. Seligman (1968) developed a

variation of the model and described the safety-signal theory, which considered an acquired

conditioned emotional response (CER) only to a signal that predicted the onset of shock, for

any other signal a conditioned response of pressing for food was shown. (Lovibond, Chen,

Mitchell & Weidemann, 2011). The safety-signal theory can be applied to this present study

as it is the intention that participants will produce an avoidance response to the CS-US

contingency and not for any other stimuli, unless derived.

Humans’ current socio-cultural environment can present indirect means of obtaining

fearful stimuli through verbal communications and social modulation i.e. expressions,

imitation etc. Mineka & Cook (1993) hypothesized that indirect fear learning is procedurally

the same as direct fear learning, and can be just as effective. This happens by the observation

of another’s reaction to the CS acting as a US. Others reaction elicits anxiety and therefore a

4

fear response; similarly seen in non-human studies such as monkeys (Cook & Mineka, 1990).

Non-human studies have shed light onto indirect fear learning by showing their abilities to

perceive and react accordingly to external signs of fear and pain. The neural workings aid in

assigning and raising awareness of a threat value to context or cue associated with the threat

(Hauser, 1996; Kavaliers, Choleris, Colwell, 2001). Social learning of aversive stimuli

through a learning model can be explained through several means but for this study the

concentration will be on imitation and empathy. In relation to empathy; Knapska, Mikosz,

Werka & Maren (2009) showed socially transmitted fear in rats through social interaction. If

a rat had previously interacted with a recently fear-conditioned partner, they exhibited

defensive behaviour during the novel task, hypothesizing the social interaction acted as a US

leading the rats to empathize.

Kavaliers et al., (2001) studied mice and their socially mediated acquisition of fear to

the aversive stimulus of biting flies. ‘Observer’ mice watched ‘demonstrator’ mice being bit

by flies and for avoidance exhibited analgesia and self-burying. When exposed to non-biting

flies, the observer mice displayed the conditioned analgesia and self-burying as observed

from before. This imitation has also been shown in humans. Self-reports from children about

their unexplained fear of the dark/animals explain they are developed from observing their

parents fearful emotions in contextual situations (Mineka & Zinbarg, 2006). This was shown

in Gerull & Rapee (2002) when mother displays negative facial expressions towards novel

stimuli, the odds of toddler imitating fearful expressions and avoidance behaviour increases.

Helsen, Goubert, Peters & Vlaeven (2011) demonstrated the effects of observational learning

on pain related fear and if it was associated with increased pain ratings. Self-reports of fear of

pain and expected pain ratings were taken after observing a learning model, whilst actual pain

and avoidance measures were taken post-exposure, and found observing another person

performing the task ended in a higher reporting of intense and unpleasant pain. Guzmán,

5

Tronson, Guedea, Huh, Gao, & Radulovic (2009) showed that even though models on

humans and non-humans have shown that observational learning instils fearful conditioning

after observing others it can also ease some forms of fear. In fear conditioning mice, they

found a social buffering process served to prevent a long-term induction of fear to stressful

events. This was also shown by Egliston & Rapee (2007) when investigating fear and

avoidance learning for toddlers in three conditions (1) modelling group – positive feedback

from mother about aversive stimuli (2) exposed to aversive stimuli alone (3) control group.

All toddlers were then exposed to observational conditioning trial where the experimenter

reacted fearfully of the aversive stimuli. Results showed toddlers in (1) were more likely to

show positive affective reactions in comparison to the other groups, showing the power of

positive parental reactions, a social buffer, to fear evoking stimuli which can over-ride

negative reactions.

All past literature shows fears learned through observational learning engage in the

same neural mechanisms as direct learning, signifying indirect means may be just as effective

and powerful as direct fear conditioning.

Another form of indirect learning is inferred/derived relational responding which can

be best explained through the relational frame theory (RFT) which proposes that humans

transfer fear and avoidance responses across stimuli (Barnes-Holmes, Barnes-Holmes,

Smeets, Cullinan & Leader, 2004; Healy, Barnes-Holmes & Smeets, 2000). An important

process which can be undermined is the verbal and language processes which contribute to

the avoidance response of indirect conditioning i.e. if a person has a spider phobia, but has

never directly had fear-evoking experiences with a spider, the word ‘spider’ can act as a CS.

Upon hearing the word ‘spider’, the individual elicits an alarmed response and seeks

avoidance. This transfer of stimulus functions can also be seen through topographically

different ‘spider-like’ stimuli i.e. real-life spiders, pictures, etc. The effect of verbal

6

information has been shown to result from associate learning (Dymond & Rehfeldt, 2000;

Field, 2006). By the functional definition of verbal events, there is the possibility of a

contribution of indirectly acquired human avoidance learning in two concepts: derived

relational responding and the transformation of functions.

Augustson & Dougmer (1997) illustrated transformation of avoidance functions

through training an ‘aversive class’ A1-B1-C1-D1 and a ‘non-aversive class’ A2-B2-C2-D2

then Pavlovian conditioning coupled B1 with shock and B2 without. A differential signalled

avoidance task was issued where participants could avoid shock by pressing a key in relation

to B1 appearing. Results showed that all participants during stimulus presentations of the

aversive class elicited avoidance responses in comparison of non-aversive class, which they

did not. This study was the first of its kind to depict how avoidance can be present in the

absence of direct conditioning. This transformation of functions, i.e. establishing a particular

behavioural response for one stimuli of an equivalence relation and then transferring the

function for its multiple relations, was also seen in Forsyth, Eifert & Barrios’ (2006) study

where by training A-B and B-C, participants derived B-A, C-B, A-C and C-A. If ‘B’ was

paired with a fear-evoking response, then participants would derive to produce fearful

behaviour for ‘A’ as well as the trained relation of ‘C’.

RFT also proposes other forms of transfer of avoidance, other than stimulus

equivalence based accounts, such as the relational frames of Same and Opposite can be learnt

by humans (Dymond, Roche, Forsyth, Whelan & Rhoden, 2007; Dymond, S., Roche, B.,

Forsyth, J. P., Whelan, R., & Rhoden, J. 2008; Whelan & Barnes-Holmes, 2004). Dymond et

al., (2008) study exposed participants to a relational completion procedure (RCP) of non-

arbitrary and arbitrary testing and training to establish Same and Opposite and derived

relations, which has been shown to be an effective procedure (Dymond & Whelan, 2010).

Participants were taught all possible Same and Opposite relations of A1-B1-C1 and A2-B2-

7

C2 which was followed by an avoidance conditioning procedure where the participants learnt

to escape the discriminative stimulus (SD) of aversive images and sounds by pressing the

space bar and not omitting the same avoidance response to B2. They were also tested for

transformation of functions for C1 and C2 and found participants that learnt the

transformation of functions also displayed correct derived relational responding, this depicts

the ability of humans to complete this task.

The aim of this study is to examine the effectiveness of observed and derived

relational learning on pathways of avoidance by exposing each participant to a fear and

avoidance conditioning phase. The observed group will watch a video of a learning model

completing a Pavlovian conditioning and avoidance acquisition phase. The video depicts a

learning model receiving a mild electric shock when a selected circle (CS+) is shown on the

screen; where the learning models facial expressions and arm movements whilst receiving the

shock will serve as the US. The stimulus will predict the delivery of shock (CS+) whilst

another contextually similar stimulus will present no negative consequences (CS-).

Participants are instructed that they can avoid the CS+ if they choose the correct button

(AV+). The inferred condition participants will be presented with RCP trials of non-arbitrary

and arbitrary training and testing of Same and Opposite relations, as shown in Dymond et al.,

(2008) study, to test for transformation of functions and derived relational responding in later

trials i.e. avoidance acquisition and extinction. The participants will have the AV+ option to

avoid the upcoming CS+ shock.

The participant’s expectancy of the shock on stimuli will be recorded through Likert

Scales and its two factors of Avoidance and Non-Avoidance, at the end of certain phases (see

Method). Bass, van Ooijen, Goudriann & Kenemans (2008) showed that a failure to learn the

CS-US contingency would result in a higher contextual fear, and then participants would self-

report higher levels of anxiety, which was confirmed by skin conductance (SCR) measures,

8

any indiscrepancies will be followed up in the data. Behavioural avoidance will be recorded

through percentage scores for each participant from eliciting an avoidance response for CS+

and CS- for each appropriate phase. An Acceptance and Action Questionnaire (AAQ-II;

Bond et al., 2011) will be undertaken by each participants. AAQ-II is a reliable psychometric

7-item scale to measure participant’s willingness to experience unwanted private events in the

intention to reach individual goals and values i.e. ‘My painful memories prevent me from

having a fulfilling life’ which participants would rate from 1=never true to 7=always true.

López et al., (2010) found that people with high AAQ-II scores displayed a higher number of

thought intrusions and higher susceptibility to a lower concentration level to the task at hand;

this will be considered for any rogue data. The results of AAQ-II will be tested against

behavioural avoidance responses to test for any correlations.

An interaction between group (observed vs. inferred) and the three dependent

variables (DV) of avoidance behaviour responses of the shock, expectancy of shock and

AAQ-II will show any superior effectiveness of each group. The IV is conditioned stimuli

(CS+, CS-). This study based upon a two-way hypothesis as it is searching for the

effectiveness of each learning group.

9

Method

Participants

A total of sixty participants (21 males and 39 females), were randomly placed in

either the observed or inferred condition, with 30 in each condition. All recruited participants

had English as their first language and were aged between 19-40 years old (M= 21.9, SD =

3.91). They were recruited via opportunistic, personal contacts and Swansea University’s

Subject Pool. All participants were given an information sheet and a consent form (Appendix

1) before participating in the study, and they were given a debrief form at the end (Appendix

2). They participated in return for either course credit or £5.

Design

The between- subject factor is Group which has two levels: observed (30 participants)

and inferred (30 participants). There are two within-subject factors (1) Conditioned Stimulus

which has two levels: CS+ and CS-. (2) Learned Behaviour which has two levels: Avoidance

(AV) and Non-Avoidance (NAV). Thus combination of the three factors created six

experimental conditions. The DerivedCS+ will be compared against the inferred response

rates of CS+ and CS-. There are three dependent variables: Avoidance behaviour responses of

the shock in each phase, expectancy of shock (taken from Likert scales at the end of each

phase), and AAQ-II results. (Appendix 3).

Apparatus

The computer system was Microsoft Windows XP Professional Version 2002 on a

17’’ LCD colour monitor, 1024 by 768 resolution. The computer program was written in

Visual Basic® 6.0 which controlled all stimulus presentations and recorded all responses for

both conditions. The electrode was 5cm in diameter, the program used for the production of

electric shocks was PowerLab 2/25 Chart 5 for Windows.

10

Stimuli

Observed Condition

A movie was produced for the observation stage of the experiment. The movie (3 min,

8s) depicted a male participant (the learning model) participating in a differential fear

conditioning experiment (Figure 1). Two coloured circles (red and blue) served as CS and

were presented in pseudo-random order for 3s on a computer screen in front of the learning

model. The appearance of the circles was incorporated with an inter-stimulus interval (ITI)

which varied between 4 and 7 s. Each coloured circle was presented six times through Phase

2 and 3. Six presentations of the colour circle serving as the CS+ were coupled with an

uncomfortable shock to the left wrist of the learning model, where the model then displayed

signs of distress i.e. tensing of the left arm, lowering of the eyebrows (US); whereas CS- was

never paired with a shock. The level of shock was selected by the learning model to be

uncomfortable but not painful prior to the recording of the video. In Phase 3, the participants

viewed the learning model being able to cancel the upcoming shocks by pressing the button

‘M’ on the keyboard.

Fig. 1. A snapshot from the movie presented to participants during the Pavlovian avoidance stage, showing the model facing a computer screen that presented the CS+ and CS-.

Inferred Condition

For the RCP trials, two stimuli in ‘wingdings’ font were used as contextual cues for

Same (i.e.) and Opposite (i.e.). Six nonsense syllables were employed as sample and

comparisons during relational training and testing, i.e. ZID, PAF, JOM, BEH, DAX, QAF.

These are clarified using alphanumerics i.e. A1, B1, B2, C1, C2, N2. An N2 stimulus, if

11

selected, was never reinforced.

In Phases 4 – 7, B1, B2, C1 and C2 were used to test derived relations. Six

presentations of the nonsense word serving as the CS+ were coupled with an uncomfortable

shock to the non-dominant arm of the participant, which they set before the trials to be

uncomfortable but not painful, whereas CS- and DerivedCS+ was never paired with a shock.

Procedure

Observed Condition

General Procedure:

On arrival to the lab, participants were given an information sheet depicting what was

expected of them during the study and the use of shocks, any ailment issues which would

mean termination of their participation, and if they wanted to opt out at any time they could

do so with no penalties. If they agreed, then a consent form was signed, the AAQ-II

questionnaire (M=17.13, SD=8.5) was counterbalanced so was either presented at this point

or at the end of the experiment. Participants were then sat in front of a computer and an

electrode was attached to the non-dominant arm and held in place by an elasticized bandage.

Shock was set by the participants to a level that was uncomfortable but not painful (M= 0.51,

SD= 0.2) starting a 0.250mA and working up in increments of 0.5. The participants were

exposed to four conditions (1) Habituation (2) Pavlovian Avoidance (3) Avoidance

Acquisition; and (4) Extinction. Phases (2) and (3) were observed by the participant watching

the pre-recorded video of the learning model.

Phase 1: Habituation

Three presentations of CS+ and CS- were presented where the participant simply

watched the screen. See Appendix 4 for computer instructions throughout both observed and

inferred conditions.

12

Phase 2: Pavlovian Avoidance (PA) and Phase 3: Avoidance Acquisition (AA)

Participants observed the learning model completing the Pavlovian Avoidance stage

where each circle was presented six times and the CS+ circle administered shocks to the

learning model. Participants then observed the learning model completing the Avoidance

Acquisition stage where the model learnt to cancel the upcoming shocks when the CS+ circle

appeared on the screen by pressing the ‘M’ key on the keyboard.

Phase 4: Extinction (EXT)

This was the testing phase, where 5 presentations of the coloured circles were

presented but in a different order to the one shown in the video. Essentially, no shocks were

administered to the participant during this phase to ensure learning was acquired through

indirect, social means only. At the end of testing, participants were presented with a Likert

Scale with four questions; (CS+ is in italics and will keep throughout the method) (1) To

what extent do you expect a shock when the red circle is presented and you do press a key?

(2) To what extent do you expect a shock when the red circle is presented and you do NOT

press a key? (3) To what extent do you expect a shock when the blue circle is presented and

you do press a key? (4) To what extent do you expect a shock when the blue circle is

presented and you do NOT press a key? Participants could rate their answers between -3

(very unlikely) to 3 (very likely). Ratings are the same for all Likert Scales in both

conditions. At the end of the experiment, subjects were given a debrief form.

Inferred Condition

On arrival to the lab, participants were given an information sheet depicting what was

expected of them during the study and the use of shocks, any ailment issues which would

mean termination of their participation, and if they wanted to opt out at any time they could

do so with no penalties. If they agreed, then a consent form was signed, the AAQ-II

questionnaire (M=15.63, SD=6.02) was counterbalanced so was either presented at this point

13

or at the end of the experiment. The participants were exposed to eight conditions (1) Non-

Arbitrary Training (2) Non-Arbitrary Testing (3) Arbitrary Training (4) Arbitrary Testing (5)

Habituation (6) Pavlovian Avoidance (7) Avoidance Acquisition; and (8) Extinction. Phase 1

– 4 used RCP to train and test same and opposite relations and derived relationships.

RCP General Procedure;

Clicking on a tick box at the bottom of the screen began the experiment for the

participants. A sample picture/nonsense words appeared on the top left corner of the screen,

after 1s the contextual cue (i.e. or) appeared in the upper centre, and after another 1s a

blank yellow square would appear on the top right corner of the screen. Three comparison

stimuli appeared on the bottom section of the screen. A selection could be made by clicking

the right mouse button on the intended picture and dragging to the black yellow square, then

releasing the mouse button. When this happened, two red buttons appeared on the bottom of

the screen that displayed the options, ‘Finish Trial’ or ‘Start Again’. Pressing the ‘Start

Again’ button refreshed the page to where all the stimuli were at the start of the trial for the

opportunity for another selection. Pressing the ‘Finish Trial’ button cleared the screen and

either produced a feedback screen during training phases, where if the correct selection was

made, the word ‘Correct’ was displayed, otherwise the word ‘Wrong’ was displayed. The

intertribal interval (ITI) of 3 seconds then happened, until the next trial began.

Phase 1: Non-Arbitrary Relational Training

Phase 1 and 2 are non-arbitrary training and testing intended to establish same and

opposite labelling for the contextual cues. The samples and comparisons stimuli were non-

arbitrary (formally related) and choosing the stimulus that was either the same or opposite of

the sample was reinforced. For instance, if a participant was shown a ‘smiley face’ as the

sample and the opposite contextual cue was shown, the selecting the ‘frowning face’ as the

comparison was reinforced, likewise if the same contextual cue was shown, and the same

14

‘smiley face’ comparison was selected. There were six stimulus sets (shown in Table 1),

presented in a random order for each participant. When participants produced eight

consecutively correct responses, they then progressed onto Phase 2.

Table 1Stimulus sets used during Phase 1 and 2, and the physical dimensions of each stimulus set.

Physical Dimension

Description End 1 End 2

Phase 1 Squares Black White

Lines Short Long

Smiley Face Frowning Smiling

Cubes Small Big

Trees Small Big

Red Disk Sections Thin Crescent Full disk

Phase 2 Columns Narrow Wide

Wavy lines Small amp. Big amp.

Snowstorm No snow Whiteout

Buildings Small Big

Pointed Star 3 points 20 points

Bowed Trees Straight Very bowed.

Phase 2: Non-Arbitrary Relational Testing

Phase 2 followed the same design as Phase 1, except the crucial removal of feedback

of responses, and was just followed by the ITI. Six new stimulus sets were used which were

presented in a random order. Participants were required to respond correct consistently across

all eight trails to be exposed to Phase 3, failure resulted in re-exposure to Phase 1.

Phase 3: Arbitrary Relational Training

Phase 3 and 4 are arbitrary training and testing intended to establish same and

opposite relations of sample and comparison stimuli of an arbitrary nature (trigrams). For this

example and throughout, it will be explained in this way; the contextual cue used will be

stated in capitals, followed by the sample stimulus, and then followed by the three

comparison stimuli options in brackets. The correct answer which would be reinforced will

be shown in italics i.e. SAME/A1 (C2-C1-N2). The example demonstrates that in the

15

presence of the contextual cue SAME and the sample stimulus A1, choosing C1 would have

been reinforced, whereas choosing C2 and N2 would not have been. All participants were

presented with the following four training trials: SAME/A1 (C2-C1-N2), SAME/A1 (N2-B1-

B2), OPPOSITE/A1 (B2-N2-B1), OPPOSITE/A1 (C2-N2-C1). Training was presented in

eight trials, with each trial presented twice. If participants made the correct answer they were

given feedback by the word ‘Correct’ appearing on the screen. When participants produced

eight consecutively correct answers they were immediately exposed to Phase 4.

Phase 4: Arbitrary Relational Testing

Phase 4 tested for the emergences of derived relations from the training of same and

opposite. Figure 2 depicts the predicated relational network. Feedback was withheld during

this phase, although the alphanumeric in italics is what the participants would choose if

responding in accordance with the predicted relational network; the test trails were as

follows; SAME/B1 (C2-C1-N2), SAME/B2 (C2-N2-C1), SAME/C1 (N2-B1-C2), SAME/C2

(C1-N2-B2), OPPOSITE/B1 (C2-C1-N2), OPPOSITE/B2 (C1-N2-C2), OPPOSITE/C1 (N2-

C2-B2), and OPPOSITE/C2 (B1-B2-N2). The author would like to note that all possible

probe trials i.e. B-C and C-B were tested during the arbitrary relational testing as in Dymond

et al., (2008) paper. Testing was carried out for sixteen trials, with each task repeated twice

per block. It was compulsory that participants produced a minimum score of 14/16 (87.5%)

correct responses to pass. Failing to meet this measure they were immediately exposed to

Phase 1 again and set to go through each trial until meeting the goal.

16

Fig. 2. The predicted relational network of trained and tested stimuli. The alphanumerics represent the nonsense syllables used in the Phase 3 and the letters S and O indicate Same and Opposite. Solid lines signify trained relations whilst dashed lines signify derived relations.

General Procedure #2

After meeting the criterion for the RCP trials, the electrode was attached to the non-

dominant arm of the participant and held in place by an elasticized bandage. Shock was set

by the participants to a level that was uncomfortable but not painful (M= 0.51, SD= 0.19)

starting a 0.250mA and working up in increments of 0.05

Phase 5: Habituation

Three presentations of CS+ and CS- were presented on to the screen, exposing

participants to the stimuli.

Phase 6: PA

The participants sat comfortably whilst observing the screen, when the CS+ appeared

a mild shock, which level was selected by the participant before, was administered to their

arm through the electrode. They were exposed to each CS+ and CS- six times. In order to

counterbalance which word served as the CSD+ and CS-, the program was edited to have

either B1 or B2 as the CS+. After the phase was finished they were asked to complete a

17

Likert Scale, the questions were as follows; (1) To what extent do you expect a shock when

PAF is presented? (2) To what extent do you expect a shock when JOM is presented?

Phase 7: AA

The aim of this trial was to observe if participants would elicit a simple avoidance

response during the presence of CS+ and not CS-. Six trials of the CS+ and CS- (B1 and/or

B2) were presented to the participant with the option to avoid the shock by pressing the ‘M’

key on the keyboard. Participants had to cancel 5 consecutive shocks to pass onto the next

phase. If they met the criterion they were passed onto Likert Scale questions; (1) To what

extent do you expect a shock when PAF is presented and you do press a key? (2) To what

extent do you expect a shock when PAF is presented and you do NOT press a key? (3) To

what extent do you expect a shock when JOM is presented and you do press a key? (4) To

what extent do you expect a shock when JOM is presented and you do NOT press a key?

Then participants are finally exposed to the final phase.

Phase 8: EXT

This was the testing phase; 6 presentations of CS+ and CS- were presented alongside

introducing the DerivedCS+ i.e. if CS+ was B1, using the predicted relational network the

DerivedCS+ would be C1, same scenario if CS+ was B2 so the DerivedCS+ would be C2., so

the aim of this phase is to test if participants have learnt the derived relations to elicit an

avoidance response. Essentially, no shocks were administered to the participants during this

phase to ensure learning was acquired through learned responses beforehand and also testing

for derived relations. After this phase, they were presented with the last Likert Scale, which is

similar to Phase 7 Likert Scale but with an important last two questions about the

DerivedCS+; (1) To what extent do you expect a shock when PAF is presented and you do

press a key? (2) To what extent do you expect a shock when PAF is presented and you do

18

NOT press a key? (3) To what extent do you expect a shock when JOM is presented and you

do press a key? (4) To what extent do you expect a shock when JOM is presented and you do

NOT press a key? (5) To what extent do you expect a shock when BEH is presented and you

do press a key? (6) To what extent do you expect a shock when BEH is presented and you do

NOT press a key? At the end of the experiment, subjects were given a debrief form.

19

Results

Table 2 demonstrates the performance of participants throughout Inferred Phases 1 –

4 (RCP). 13/30 participants passed both non-arbitrary and arbitrary relational testing on their

first exposure. P13 passed non-arbitrary testing after two exposures and passed arbitrary

testing on their first exposure. 8/30 participants passed both non-arbitrary and arbitrary

relational testing on their second exposure. 5/30 participants passed both non-arbitrary and

arbitrary relational testing on their third exposure. The final 8/30 participants passed both

non-arbitrary and arbitrary relational testing on their maximum fourth exposure. Therefore,

all 30 participants were eligible to progress on the avoidance-conditioning phase.

Table 2Trials to Criterion and Percentage of Correct Responses in Phase 1 – 4

Participant Phase 1: Non-arbitrary Relational Training (trials to criterion)

Phase 2: Non-arbitrary Relational Testing (%)

Phase 3: Arbitrary Relational Training (trials to criterion)

Phase 4: Arbitrary Relational Testing (%)

1 10 100 16 43.758 100 12 87.5

2 9 100 13 93.753 9 100 44 62.5

8 100 8 68.758 100 8 87.5

4 12 100 8 93.755 18 100 10 87.56 14 100 9 93.757 10 100 23 93.758 8 100 15 81.25

8 100 8 93.759 8 100 9 37.5

8 100 8 81.258 100 8 87.5

10 16 100 20 93.7511 9 100 12 93.7512 23 100 12 56.25

8 100 9 56.258 100 8 93.75

13 15 62.58 100 25 87.5

14 8 100 8 68.758 100 9 81.258 87.58 100 8 93.75

15 11 85.71

20

8 87.58 100 41 81.258 100 9 93.75

16 14 100 24 87.517 9 100 16 62.5

8 100 8 81.258 100 8 100

18 9 87.513 100 36 93.75

19 44 100 52 758 100 8 87.5

20 8 100 14 10021 8 100 9 87.522 12 100 31 56.25

8 100 9 56.258 87.513 100 8 93.75

23 10 100 32 81.259 100 8 93.75

24 11 100 28 7511 100 43 87.5

25 8 100 9 68.758 100 8 97.5

26 11 100 52 68.7516 100 21 2510 100 14 87.5

27 11 100 34 97.528 9 100 48 37.5

13 100 18 87.529 13 100 15 10030 8 100 8 100

Descriptive Graphs

Figure 3 displays the mean percentage of avoidance behaviour across both learning

groups, (Observed: Phase 4; Inferred: Phases 7 & 8) and the three stimuli (CS+, CS- and

DerivedCS+). The graph suggests a significant difference between CS+ and CS- for both

learning groups and also depicts the emergence of derived learning from its similarity from

the CS+ stimuli. The small standard error bars show low variability in avoidance response

results.

21

Fig. 3. Graph showing the mean avoidance behaviour (%) of participants in both learning groups.

Figure 4 displays the mean expectancy of shock if they were to elicit an avoidance

response, that was self-reported from participants through Likert Ratings across both learning

groups (Observed: Phase 4; Inferred; Phases 6-8) and three stimuli (CS+, CS- and

DerivedCS+). The graph suggests that Observed participants rated the expectancy of shock

after an avoidance response higher than the Inferred group, the standard error shows great

variability across both learning groups responses.

22

Fig. 4. Graph showing the mean expectancy of a shock when performing an avoidance response.

Figure 5 displays the mean expectancy of shock if not eliciting an avoidance response.

Results were self-reported by participants through Likert Ratings across both learning groups

(Observed: Phase 4; Inferred; Phases 6-8) and three stimuli (CS+, CS- and DerivedCS+). The

graph suggests a significant difference for both learning groups between CS+ and CS- as the

former had a greater mean expectancy rate over the latter. It also depicts the emergence of

derived learning from the similarity between the CS+ and DerivedCS+ stimuli. The standard

error bars show low variability in both groups.

Fig. 5. Graph showing the mean expectancy of shock if participants across both learning groups did not perform an avoidance response.

Aggregate Results

Two independent samples t-tests were performed to examine the statistical evidence

shown from the graphs. In each test, the percentage scores of avoidance responses of CS+

and CS- in the Observation EXT group were compared against Inferred EXT group. For these

two tests, P 16 and 22 data were excluded for failure to meet the criterion.

23

Firstly, Observed EXT CS+ was compared against Inferred EXT CS+. Considering

the group statistics, there seemed to be very little difference in the means, which is confirmed

using the top row of the t-test, as Levene’s test is insignificant (F=3.91, p=.053), showing an

insignificant difference (Observed EXT group M=98.33, SD=5.09; Inferred EXT group

M=99.40, SD=3.15; t(56)=.956, p=.343). Secondly; Observed EXT CS- was compared

against Inferred EXT CS-, considering the group statistics, there was possibly a small

difference in the means, and Levene’s test showed a significant difference (F=13.49, p=.001)

between the variances of the two groups. Using the bottom row of the t-test, it can be

concluded there is no significant difference between the means at the 5% significance level

(Observed EXT group M=1.11, SD=4.23; Inferred EXT group M=4.17, SD=8.64;

t(38.63)=1.69, p=.099). However, there is a weak significant difference at the 10% level,

showing that the Observed group preformed marginally better than the Inferred, as they had a

lower mean score. Overall, the independent t-tests showed Observed and Inferred learning

groups were both as effective for eliciting avoidance behaviour.

Six paired samples t-tests were conducted to study the percentage scores of avoidance

responses of CS+ and CS-, and to test the emergence of a derived relational response in

Inferred EXT phase. They also tested the expectancy of shock Likert Scale ratings for the

Inferred condition.

Observed EXT CS+ was compared against Observed EXT CS-. At the 10% significance

level there was a weak negative correlation (correlation=.356, p=.053) between CS+ and CS-

showing participants that elicited a avoidance response on CS+, did not on CS-, as was

expected. This is supported by the descriptive statistical means and a significant difference in

the t-test (Observed EXT CS+ M=98.33, SD=5.09; Observed EXT CS- M=1.11, SD=4.23;

t(29)=69.29, p=.000).

24

Inferred AA CS+ was compared against Inferred AA CS-. P16 data was excluded for

failure to meet the criterion. Samples statistics shows CS+ had a higher mean avoidance

response over CS-, showing a higher proportion of participants showed learnt behaviour and

cancelled the upcoming shocks, over the CS-. Also supported by the t-test showing a

significant difference (Inferred AA CS+ M=93.75, SD=12.52; Inferred AA CS- M=2.08,

SD=5.6; t(31)=40.83, p=.000).

Inferred EXT CS+ was compared against Inferred EXT CS-. P16 and 22 data were

excluded for failure to meet the criterion. The samples statistics indicate a significant

difference between the CS+ and CS- from the mean avoidance response, and this is

confirmed by the t-test showing a significant difference. This shows that participants

produced the avoidance response by learning the correct relation (Inferred EXT CS+

M=99.40, SD=3.15; Inferred EXT CS- M=4.17, SD=8.64; t(27)=56.57, p=.000).

Inferred EXT DerivedCS+ was compared against Inferred EXT CS+ to test for

emergences of derived relations. P16 and 22 data were excluded for failure to meet the

criterion. The samples statistics indicate no significant difference between CS+ and

DerivedCS+ and is confirmed by the t-test showing no significant difference. Results

illustrate derived transformation of functions emerged and were as effective as CS+ stimuli

(Inferred EXT Derived CS+ M=95.24, SD=19.7; Inferred EXT CS+ M=99.4, SD=3.15;

t(27)=1.1, p=.282).

Inferred EXT DerivedCS+ was compared against Inferred EXT CS-. P16 and 22 data

were excluded for failure to meet the criterion. Samples statistics suggested a significant

difference, and was confirmed by the t-test showing a significant difference. This shows that

participants produced the avoidance response by learning the correct derived relation to CS+

therefore learning to cancel the upcoming shock, as would be expected under CS+ (Inferred

25

EXT DerivedCS+ M=95.24, SD=19.7; Inferred EXT CS- M=4.17, SD=8.64; t(31)=40.83,

p=.000).

The final paired samples t-test was conducted on the Inferred PA Likert Scales showing

expectancy of shock when presented with the contextual stimuli. The sample statistics

illustrate a significant difference, and the t-test confirms. This shows that the participants

rated the correct CS+ higher than the CS- on expectancy of shock (PA CS+ M=5.93,

SD=.254; PA CS- M=.133, SD=.434; t(29)=57.67, p=.000).

A Pearson’s r test was conducted to investigate any correlations between AAQ-II data

and learnt behaviour. However, there were no significant correlations identified between:

Observed AAQ-II scores and Observed EXT CS+ (r= -.114, N=30. P=.548); Observed AAQ-

II and Observed EXT CS- (r= .219, N=30, p=.244); Inferred AAQ-II scores and Inferred

EXT CS+ (r= -.008, N=30, p=.966); or between Inferred AAQ-II scores and Inferred EXT

CS- (r=.200, N=30, p=.290). This shows there is no relationship between AAQ-II scores and

avoidance responses of CS+ or CS-.

The total number of responses in the Inferred AA Likert Scales for each of our within

subject experimental conditions was entered into a 2 (CS+, CS-) x 2 (AV, NAV) multi-way

within subjects ANOVA. Results showed a significant main effect of Conditioned Stimuli

(F(1,29)=108.38, p=.000), showing that conditioned stimuli had a significant effect on the

expectancy of shock. Another significant main effect of Learned Behaviour (F(1,29)=73.9,

p=.000) also showed learned behaviour effects expectancy of shock. A significant interaction

was shown between Conditioned Stimuli*Learned Behaviour (F(1,29)=126.68, p=.000).

Figure 6 shows participants understood the significance of the avoidance behaviour. The

difference in the estimated marginal means for AV and NAV for CS+ shows that participants

learnt the importance of the avoidance response, whereas for CS- where the results are

26

similar for AV and NAV, participants learnt that they wouldn’t get shocked regardless of

avoidance response.

Fig. 6. Estimated marginal means of expectancy scores displaying the interaction between

Conditioned Stimuli and Learned Behaviour.

Post-hoc tests (paired samples t-test) were conducted to further investigate these

interactions. Results showed the mean difference between CS+NAV was significantly larger

for CS+AV (t(29)=11.43, p=.000), this showed that participants learnt the effectiveness of the

avoidance behaviour to reduce the expectancy of shock. There was no significant mean

difference between CS-AV and CS-NAV (t(29)=.432, p=.669) and no significant mean

difference between CS+AV and CS-AV (t(29)=1.18, p=.247). These both show that

participants learnt when to expect a shock and the appropriate avoidance response. There was

27

a significant mean difference between CS+NAV and CS-NAV (t(29)=21.13, p=.000) which

illustrates that participants learnt the correct shock stimulus.

The total number of responses in the Observed EXT and Inferred EXT Likert Scales for

each of our six experimental conditions was entered into a 2 (observed, inferred) x 2 (CS+,

CS-) x 2 (AV, NAV) multi-way mixed ANOVA. Results showed a significant main effect of

the conditioned stimuli F(1,58)=299.34, p=.000), a significant main effect of learned

behaviour (F (1,58)=114.35, p=.000) and a significant effect on learning group

(F(1,58)=10.97, p=.002) which showed that which learning group the participant was in had a

significant effect on the expectancy of shock.

Within-subjects test showed a non-significant effect of group*conditioned stimuli

interaction (F(1,58)=.148, p=.702); Non-significant effect of group*learnt behaviour

interaction (F(1,58)=2.22, p=.141) and a significant interaction of condition stimuli*learnt

behaviour (F(1,58)=201.38, p=.000). This interaction is shown in Figure 7 participants

understood the significance of the avoidance behaviour. The difference in the estimated

marginal means for AV and NAV for CS+ shows that participants learnt the importance of

the avoidance response, whereas for CS- where the results are similar for AV and NAV,

participants learnt that they wouldn’t get shocked regardless of avoidance response.

The three-way interaction between conditioned stimuli*learnt behaviour*group was not

significant (F (1,58)=.02, p=.889).

28

Fig. 7. Estimated marginal means of expectancy scores from Observed EXT and Inferred

EXT displaying the interaction between Conditioned Stimuli and Learned Behaviour.

Figure 8 plots the group against conditioned stimuli whilst Figure 9 plots group against

learned behaviour. Observed is rated higher in expectancy of shock for both conditioned

stimuli and learned behaviour, showing that observed rated the expectancy of shock higher

than Inferred.

29

Fig. 8. Estimated marginal means of expectancy scores from Observed EXT and Inferred

EXT displaying the interaction between Group and Conditioned Stimuli.

30

Fig. 9. Estimated marginal means of expectancy scores from Observed EXT and Inferred

EXT displaying the interaction between Group and Learned Behaviour.

Post-hoc tests (independent t-test) were conducted to examine the main effects and the

conditioned stimulus*learnt behaviour interaction; results showed no significant difference at

the 5% but a weak significant difference at the 10% level between Observed CS+AV and

Inferred CS+AV. Levene’s test showed significance in variance (F=5.68, p=.020) so the

bottom row was used (t(52.87)=1.74, p=.088). A non-significant difference was shown

between Observed CS+NAV and Inferred CS+NAV (t(58)=.297, p=.768). A significant

difference was found between Observed CS-AV and Inferred CS-AV, Levene’s test

(F=11.35, p=.001) showed bottom row to be used (t(47.8)=2.19, p=.033). And finally a non-

significant difference was shown between Observed CS-NAV and Inferred CS-NAV,

31

Levene’s test (F=10.79, p=.002) so bottom row was used (t(43.84)=1.58, p=.122).

32

Discussion

The aim of this present study was to examine the effectiveness of observed and

inferred (derived) pathways of avoidance. Comparing Observed EXT CS+ and CS- and

Inferred EXT CS+ and CS-; no significant differences were found between the learning

groups and conditioned stimuli, showing the two learning groups were just as effective as

each other in producing a fear-evoked avoidance response through indirect means. The

findings suggest that observed and inferred pathways of avoidance are equally as effective as

each other, and from past supporting literature can hypothesize as effective as direct

Pavlovian fear conditioning. (Cook & Mineka, 1990; Dymond et al., 2007; Egliston & Rapee,

2007; Whelan & Barner-Holmes, 2004).

The Observed condition was shown to be effective at inducing an avoidance response

at the correct CS+ after viewing the video of the learning model in the similar task, which

mirrors Mineka & Cook (1993) study findings. This was shown by comparing Observed EXT

CS+ and CS-. There was a weak negative correlation showing participants that elicited an

avoidance response on CS+ did not on CS-, which was to be expected. It could be

hypothesized that the participants produced avoidance behaviour due to feeling empathy

(Knapska et al., 2009; Olsson et al., 2007) or just due to imitation of the learning models

avoidance response (Kavaliers et al., 2001) as 30/30 participants produced an avoidance

response, but this can’t be fully proved without an fMRI scan during the task.

There was no correlation found between AAQ-II scores and avoidance responses of

CS+ or CS-. López et al., (2010) study was inconsistent with our findings as the participants

data that had to be excluded from the Inferred condition actually had low AAQ-II scores

(P16, AAQ-II = 12 and P22, AAQ-II = 9).

33

For the inferred condition, RCP was shown to be effective in training same, opposite

and derived relations as seen by 30/30 participants passing (Appendix 5) albeit first of final

exposure. This supports Dymond & Whelan’s (2010) study of RCP being an effective

procedure.

Comparing Inferred EXT Derived CS+ against Inferred EXT CS+ no significance was

found between the two stimuli, illustrating derived transformation of functions emerged and

avoidance responses matched those of CS+. There was a significant difference shown

between DerivedCS+ and CS- illustrating that participants again, learnt the derived relations

and ignored the CS- stimuli in relation to an upcoming shock. These results support RFT, that

participants transferred fear and avoidance responses across stimuli due to learning derived

relational responding from the RCP (Barnes-Holmes et al., 2004, Healy et al., 2000).

In the 2*2 within subjects ANOVA for the Inferred AA Likert scales, there was a

significant main effect on conditioned stimuli and learnt behaviour showing both variables

had an effect on the perception of the expectancy of shock. The significant interaction

between conditioned stimuli*learned behaviour illustrates that participants learnt the correct

contingency for CS+ and showed they understood they wouldn’t get shocked for CS-

regardless of avoidance response. Post-hoc tests showed that participants understood the

effectiveness of the avoidance behaviour to reduce the expectancy of shock.

In the final 2*2*2 multi-way mixed ANOVA, results showed significant main effects

for conditioned stimuli, learned behaviour and learning group. Comparing conditioned

stimuli and learning group the only significant interaction was between Observed CS-AV and

Inferred CS-AV which could show that participants in both conditions found the understood

the significance of CS- and that they did not have to produce an avoidance response.

34

Comparing group against conditioned stimuli and learned behaviour (Figure 8 and 9)

it is clear that Observed rated higher in expectancy of shock to both, this could be attributed

to the Observed group understanding the CS-US contingency and the value of CS- more than

the Inferred group.

The Observed condition was shown to be effective at inducing an avoidance response

at the correct CS+ after viewing the video of the learning model in the similar task, which

mirrors Mineka & Cook (1993) study findings. This was shown by comparing Observed EXT

CS+ and CS-. There was a weak negative correlation showing participants that elicited an

avoidance response on CS+ did not on CS-, which was to be expected. It could be

hypothesized that the participants produced avoidance behaviour due to feeling empathy

(Knapska et al., 2009; Olsson et al., 2007) or just due to imitation of the learning models

avoidance response (Kavaliers et al., 2001) as 30/30 participants produced an avoidance

response, but this can’t be fully proved without an fMRI scan during the task.

However, a criticism of this condition is that the participants were exposed to fear-

evoking stimuli and then immediately exposed to a directly taught avoidance response; this

can be a serious limitation that could suggest the results may have been taught through direct

means. The direct observation of the learning model completing the avoidance response in

the presence of one stimuli and not the other will have affected the avoidance contingency

and consequently influenced the avoidance behaviour. Research has demonstrated that

explicit instruction is not needed to generate indirect observational fear learning (Guzmán et

al., 2009; Helsen et al., 2011; Olsson et al., 2007). To improve on this condition, the

participants would only be exposed to the learning model experiencing the Pavlovian

Avoidance phase, then be exposed to Likert ratings and finally, the participant will

experience avoidance acquisition phase with the same set of instructions given to the Inferred

group avoidance acquisition phase (Appendix 4). This would ensure that the participants are

35

being indirectly taught the CS-US contingency and then eliciting an avoidance behaviour

response from that.

The inferred condition results illustrate that participants showed transformation of

functions for the DerivedCS+ from the CS+. This condition supports the ideas for

improvement on the observed condition that explicit instructions were not needed for derived

transformation or avoidance behaviours to occur (Dymond et al., 2008; Field, 2006). The

derived extinction phase mirrors clinical procedures in Acceptance and Commitment Therapy

(ACT; Hayes, Follette & Linehan, 2004; Hayes, Luoma, Bond, Masuda & Lillis, 2006) called

cognitive diffusion. This is a where patients illiterate negative self-statements for a few

minutes i.e. ‘I have no friends’ until generalized of the extinction is met, i.e. vocally created

stimuli in relational to real-life environmental events, not within the therapy setting.

The observed condition would have to be retested against the inferred condition to test

for any real significant difference, as the limitations of this study hinders the potential

findings reflected in all past literature. The emergence of transformation of functions and

derived relational responding shows the effectiveness of Same and Opposite relational frames

in a RCP, and illustrates a starting point for more procedures to aid in indirect anxiety and

phobias to be built. To improve the effect of indirect learning over direct, a further study

would be to test the observed group and inferred group against an instructed (Pavlovian fear

condition) group. This will affirm the effectiveness of indirect learning of fearful stimuli.

An issue of external validity must also be raised, as it does not necessarily stand that

as participants showed promise in avoidance functions across stimuli in a lab setting with

novel stimuli, that in a real life setting the results would be the same. To conclude, the results

demonstrated participants transferred a fear response to a novel stimuli through observation

36

and that derived relational training can effectively train participants to transfer learned fear

avoidance response across novel stimuli in a lab environment.

37

References

Augustson, E. M., & Dougmer, M. J. (1997). The transfer of avoidance evoking functions through stimulus equivalence classes. Journal of Behaviour Therapy and Experimental Psychiatry, 3, 181-91.

Barnes-Holmes, D., Barnes-Holmes, Y., Smeets, P. M., Cullinan, V., & Leader, G. (2004). Relational frame theory and stimulus equivilance: Conceptual and procedural issues. International Journal of Psychology and Psychological Therapy, 4, 181-214.

Barot, S. K., Chung, A., Kim, J. J., & Bernstein, I. L. (2009). Functional Imaging of Stimulus Convergence in Amygdalar Neurons during Pavlovian Fear Conditioning. PLoS ONE, 4(7):e6156.

Barot, S. K., Kyono, Y., Clark, E. W., & Bernstein, I. L. (2008). Visualizing Stimulus Convergence in Amygdala Neurons during Associative Learning. Proceedings of the National Academy of Sciences of the United States of America, 105(52), 20959-63.

Bass, J. M., van Ooijen, L., Goudriaan, A., & Kenemans, J. L. (2008). Failure to condition to a cue is associated with sustained contextual fear. Acta Psychologica, 127(3), 581-92.

Bond, F. W., Hayes, S. C., Baer, R. A., Carpenter, K. M., Guenole, N., Orcutt, H. K., Waltz, T., & Zettle, R. D. (2011). Properties of the Acceptance and Action Questionnaire – II: A| revised measure of psychological flexibility and experiential avoidance. Behavior Therapy, 42, 676-88.

Cook, M., & Mineka, S. (1990). Selective associations in the observational conditioning of fear in rhesus monkeys. Journal of Experimental Psychology: Animal Behaviour Processes, 16, 372-89.

Davies, M., & Whalen, P. J. (2001). The Amygdala: Vigilance and Emotion. Molecular Psychiatry, 6, 13-34.

Delgado, M. R., Jou, R. L., Ledoux, J. E., & Phelps, E. A. (2009). Avoiding negative outcomes: Tracking the mechanisms of avoidance learning in humans during fear conditioning. Frontiers in Behavioural Neuroscience, 3(33).

Delgado, M. R., Li, J., Schiller, D., & Phelps, E. A. (2008). The role of the striatum in aversive learning and aversive prediction errors. Philosophical Transactions of the Royal Society of London. Series B: Biological Studies, 363(1511), 3787-800.

Dymond, S., & Rehfeldt, R. (2000). Understanding complex behaviour: The transformation of stimulus functions. The Behaviour Analyst, 23, 239-54.

Dymond, S., & Roche, B. (2009). A contemporary behaviour analysis of anxiety and avoidance. The Behaviour Analyst, 32, 7-27.

38

Dymond, S., Roche, B., Forsyth, J. P., Whelan, R., & Rhoden, J. (2007). Transformation of avoidance response function in accordance with Same and Opposite relational frames. Journal of the Experimental Analysis of Behaviour, 88(2), 249-62.

Dymond, S., Roche, B., Forsyth, J. P., Whelan, R., & Rhoden, J. (2008). Derived avoidance learning: Transformation of avoidance responses function in accordance with Same and Opposite relational frames. The Psychological Record, 58, 269-86.

Dymond, S., & Whelan, R. (2010). Derived relational responding: A comparison of Match-to-Sample and the Relational Completion Procedure. Journal of the Experimental Analysis of Behaviour, 94(1), 37-55.

Egliston, K. A., & Rapee, R. M. (2007). Inhibition of fear acquisition in toddlers following positive modelling by their mothers. Behaviour Research and Therapy, 45(8), 1871-82.

Forsyth, J. P., Eifert, G. H., & Barrios, V. (2006). Fear conditioning research as a clinical analog: What makes fear learning disorderd? In M. G. Craske, D. Hermans & D. Vansteenwegen. (Eds.). Fear and learning: Basic science to clinical application (pp. 133-53). Washington DC, USA: American Psychological Association.

Field, A. P. (2006). Is conditioning a useful framework for understanding the development and treatment of phobias? Clinical Psychology Review, 26(7), 857-75.

Gerull, F. C., & Rapee, R. M. (2002). Mother knows best: Effects of maternal modelling on the acquisition of fear and avoidance in children. Behaviour Research and Therapy, 40, 279-87.

Guzmán, Y. F., Tronson, N. C., Guedea, A., Huh, K. H., Gao, C., & Radulovic, J. (2009). Social modelling of conditioned fear in mice by non-fearful conspecifics. Behavioural Brain Research, 201(1), 173-78.

Hauser, M. D. (1996). The Evolution of Communication. Massachusetts, USA: MIT Press.

Hayes, S. C., Follette, V. M., & Linehan, M. M. (2004). Mindfulness and Acceptance: Expanding the Cognitive-Behavioural Tradition. New York, USA: The Guilford Press.

Hayes, S. C., Luoma, J. B., Bond, F. W., Masuda, A., & Lillis, J. (2006). Acceptance and Commitment Therapy: Model, processes and outcomes. Behaviour Research and Therapy (22), 1-25.

Healy, O., Barnes-Holmes, D., & Smeets, P. M. (2000). Derived relational responding as generalized operant behaviour. Journal of the Experimental Analysis of Behaviour, 74, 207-227.

39

Helsen, K., Goubert, L., Peters, M. L., & Vlaeven, J. W. (2011). Observational learning and pain related fear: An experimental study with coloured cold pressor tasks. The Journal of Pain, 12(12), 1230-39.

Kavaliers, M., Choleris, E., Colwell, D. D. (2001). Learning from others to cope with biting flies: social learning of fear-induced conditioned analgesia and active avoidance. Behavioural Neuroscience, 115, 661-74.

Kim, J. J., & Jung, M. W. (2006). Neural circuits and mechanisms involved in Pavlovian fear conditioning: A critical review. Neuroscience and Bio-behavioural Reviews, 30(2), 188-202.

Knapska, E., Mikosz, M., Werka, T., & Maren, S. (2009). Social modulation of learning in rats. Learning and Memory, 17(1), 35-42.

LaBar, K. S., LeDoux, J. E., Spencer, D. D., & Phelps, E. A. (1995). Impaired fear conditioning following unilateral temporal lobectomy in humans. Journal of Neuroscience, 15, 6846-55.

LeDoux, J. E., & Gorman, J. M. (2001). A call to action: Overcoming anxiety through active coping. American Journal of Psychiatry, 158, 1953-55.

López, J. C., Ruiz, F. J., Feder, J., Rubio, A. B., Suárez-Aguirre, J. J., Rodríguez, J. A., & Luciano, C. (2010). The role of experiential avoidance in the performance on a high

cognitive demand task. International Journal of Psychology and Psychological Therapy, 10(3), 475-488.

Lovibond, P. F., Chen, S. X., Mitchell, C. J., Weidemann, G. (2011). Competition between an avoidance response and a safety signal: Evidence for a single learning system. Biological Psychology, In press.

Mineka, S., & Cook, M. (1993). Mechanisms involved in the observational conditioning of fear. Journal of Experimental Psychology, 122, 23-38.

Mineka, S., & Zinbarg, R. (2006). A contemporary learning theory perspective on the etiology of anxiety disorders: It’s not what you thought it was. American Psychologist, 61, 10-26.

Olsson, A., Nearing, K. I., & Phelps, E. A. (2007). Learning fears by observing others: the neural systems of social fear transmission. SCAN, 2, 3-11.

Olsson, A., & Phelps, E. A. (2007). Social learning of fear. Nature Neuroscience, 10(9), 1095-102.

Phillips, R. G., & LeDoux, J. E. (1992). Differential contribution of amygdala and hippocampus to cued and contextual fear conditioning. Behavioural Neuroscience, 106(2), 274-85.

40

Schultz, W. (2006). Behavioural theories and the neurophysiology of reward. Annual Review of Psychology, 57, 87-115.

Seligman, M. E. P. (1968). Chronic fear produced by unpredictable electric shock. Journal of Comparative and Physiological Psychology, 66(2), 402-11.

Siddle, D. A., & Bond, N. W. (1988). Avoidance learning, Pavlovian conditioning, and the development of phobias. Biological Psychology, 27(2), 167-83.

Watson, J. B. & Rayner, R. (1920). Conditioned emotional reactions. Journal of Experimental Psychology, 3, 1-14.

Whelan, R., & Barnes-Holmes, D. (2004). The transformation of consequential functions in accordance with the relational frames of same and opposite. Journal of the Experimental Analysis of Behaviour, 86, 317-35.

41

Appendix 1 – Information Sheet (Observed)

Information Sheet

In this study, you will be attached to a bar electrode which will deliver electric shocks to your forearm, you will set the intensity of the shocks at the beginning of the experiment to be ‘uncomfortable but not painful’.

A set of instructions will appear on the screen before each phase begins and time will be allowed for any questions to be asked to the researcher. There are 4 phases in total one of which requires you to watch a video. At the end of the study you will be given a debrief form explaining the nature of the study and the opportunity to find out more about this research and to ask any unanswered questions you may have.

This study is conducted in accordance with the British Psychological Society and Departmental ethic guidelines, your participation in this study is completely voluntary and you may withdraw at any point, without penalty. You can also choose to have any data that you provided in this study to be disregarded at any time, during or after the study. Otherwise, your results will be kept anonymous and confidential to only the researchers or supervisors working on this study. All data will be password protected and not accessed for any other reasons than this study. You must not participate in this study if you have a pacemaker fitted; suffer with heart palpitations or currently talking psychotropic medication.

If you agree to take part in this study we would appreciate it if you agree to not discuss your participation or any details of this study with others.

Please now complete the consent forms and return them to the researcher.

42

Appendix 1 – Information Sheet (Inferred)

Information Sheet

In this study, you will be attached to a bar electrode which will deliver electric shocks to your forearm, you will set the intensity of the shocks at the beginning of the experiment to be ‘uncomfortable but not painful’.

A set of instructions will be presented on the screen before each phase begins and time will be allowed for any questions to be asked to the researcher. There are 8 phases in total, phase 6 onwards may entail up to 12 electric shocks. At the end of the study you will be given a debrief form explaining the nature of the study and the opportunity to find out more about this research and to ask any unanswered questions you may have.

This study is conducted in accordance with the British Psychological Society and Departmental ethic guidelines, your participation in this study is completely voluntary and you may withdraw at any point, without penalty. You can also choose to have any data that you provided in this study to be disregarded at any time, during or after the study. Otherwise, your results will be kept anonymous and confidential to only the researchers or supervisors working on this study. All data will be password protected and not accessed for any other reasons than this study. You must not participate in this study if you have a pacemaker fitted; suffer with heart palpitations or currently talking psychotropic medication.

If you agree to take part in this study, we would appreciate it if you agree to not discuss your participation or any details of this study with any others.

Please now complete the consent forms and return them to the researcher.

43

Appendix 1 – Consent Form (Observed & Inferred)

Project title: The effectiveness of observed and inferred pathways of avoidance on learning.

Name of Supervisor: Dr Simon Dymond, ([email protected])

Name of investigators: Katie Davies & Ellyn Moore

Name of participant:

Age:

Sex:

I consent to take part in this study and am satisfied with the instructions I have been given so far. I have also been reassured that any further information I request regarding this study will be supplied to me at the end of this experiment.

I have been informed that the data I provide will remain confidential and be password protected. I am free to ask any questions at any time before, during and after the study and have been provided with a copy of this form and a participant information sheet. I also understand I will be provided with a debrief form at the end of the study.

I understand that during this study I will be required to place a bar electrode on my forearm which will be held in place by a small bandage. I will then be subject to a series of electric shocks, of which I am able to set the intensity at the beginning of the experiment. I understand the shock level will be “uncomfortable but not painful” and am free to stop the study at any time.

I am at least 18 years of age and have not been coerced in any way to participate in this study and can withdraw from the study at any point.

I understand I cannot participate in this study if I have any medical history of cardiac problems, or have been fitted with a pacemaker.

Data protection: I agree to the university processing personal data that I have supplied. I agree to the processing of such data for any purposes connected with the research project as outlined to me.

Participant signature:

Date:

Name of researcher:

Signature:

Date:

44

Appendix 2 – Debrief Form (Observed)

Thank you for taking part in this study. The aim of this study was to examine the effectiveness of observed learning and derived learning pathways of avoidance.

Throughout this study we were testing whether you had learnt to avoid the correct stimulu just from observing someone completing a similar task, so at no point during the experiment, other than the calibration of the shock strength, were you in danger of receiving an electric shock based on the procedures of Olsson et al (2007).

We would like to remind participants that if you feel uncomfortable with this study, then you are free to withdraw your data from this study at any time, although we would also like to remind you that all results are confidential and at no time will any data be identifiable to you or your name.

If this study has resulted in any discomfort to yourself, we would like to inform you of the Student Counselling Service provided on campus at Horton House or alternatively, you can contact them on +44 (0) 1792 295592 or email [email protected]

We would like if you would refrain from talking about this experiment to others whilst it is still running as any details you reveal could affect future results.

If you would like to be informed of the results of this study once it is completed, feel free to contact the experimenters on;

[email protected] or; [email protected]

Supervisior; Dr Simon Dymond ([email protected]) Telephone: 01792 295602

If you would like more information about this area of research, please follow these references;

Dymond, S., Roche, B., Forsyth, J. P., Whelan, R., & Rhoden, J. (2007). Transformation of avoidance response functions in accordance with the relational frames of same and opposite. Journal of the Experimental Analysis of Behaviour, 88, 249-262.

Olsson, A., Nearing, K. I., & Phelps, E. A. (2007). Learning fears by observing others: the neural systems of social fear transmission. SCAN, 2, 3-11.

45

Appendix 2 – Debrief Form (Inferred)

Thank you for taking part in this study. The aim of this study was to examine the effectiveness of observed learning and derived learning pathways of avoidance.

Throughout this study we were testing whether you had learnt the correct same and opposite pairings for this experiment phases, so some phases had the threat of electric shocks whilst others did not and was just to test you.

We would like to remind participants that if you feel uncomfortable with this study, then you are free to withdraw your data from this study at any time, although we would also like to remind you that all results are confidential and at no time will any data be identifiable to you or your name.

If this study has resulted in any discomfort to yourself, we would like to inform you of the Student Counselling Service provided on campus at Horton House or alternatively, you can contact them on +44 (0) 1792 295592 or email [email protected]

We would like if you would refrain from talking about this experiment to others whilst it is still running as any details you reveal could affect future results.

If you would like to be informed of the results of this study once it is completed, feel free to contact the experimenters on;

[email protected] or; [email protected]

Supervisior; Dr Simon Dymond ([email protected]) Telephone: 01792 295602

If you would like more information about this area of research, please follow these references;

Dymond, S., Roche, B., Forsyth, J. P., Whelan, R., & Rhoden, J. (2007). Transformation of avoidance response functions in accordance with the relational frames of same and opposite. Journal of the Experimental Analysis of Behaviour, 88, 249-262.

Dymond, S., & Whelan, R. (2010). Derived relational responding: A comparison of matching to sample and the relational completion procedure. Journal of the Experimental Analysis of Behavior, 94, 37-55.

46

Appendix 3 – AAQ-II (Example)

47

Appendix 4 – Computer instructions for Observed and Inferred

Observed

Phase 1: Habituation.

“Thank you for agreeing to participate in this study. In this phase, on every trial you will be presented with one of two coloured circles. Each circle will appear for 3 seconds. Your job is to simply watch the screen. If you

have any questions please ask the experimenter now.”

Phase 2: Pavlovian Avoidance and Phase 3: Avoidance Acquisition

“You will now watch a short film of a person doing an experiment similar to the one you yourself are going to be doing afterwards. The person in the film is going to receive shocks paired with one of two coloured circles

presented to him. Pay close attention to which coloured circle is followed by shocks and which is not. The person in the film is then going to learn to press one of two marked keys on the keyboard to cancel the

upcoming shock. Again, it is important you pay close attention to the film, because, in the experiment you are going to do afterwards, you are going to have to learn which one of the marked keys cancels the upcoming

shock. When you are ready to view the film, press ‘Continue’.”

Phase 4: Extinction

“Now you are going to take part in an experiment similar to the one you have just watched. You will be presented with the same number of coloured circles as the person in the film, but in a different order.

Importantly, you will receive shocks paired with the same colour as the person in the film. However, you will be able to cancel upcoming shocks by pressing one of the marked keys on the keyboard. When you are ready to

proceed, press ‘Continue’.”

Derived

RCP General Procedure;

“Thank you for agreeing to participate in this study. You will be presented with a series of images or nonsense words on the top half of the screen from left to right. Then you will be presented with 3 images or nonsense

words on the bottom of the screen.Your task is to observe the images or words that appear from left to right and drag one of these images or words from the bottom to the blank, yellow square. Click and hold the mouse over the image or word to drag it to the blank square. To confirm your choice, then click ‘Finish Trial’. If you wish to make another choice, then click

‘Start Again’.Sometimes you will receive feedback on your choices, but at other times you will not. Your aim is to get as many

tasks correct as possible. It is always possible to get a task correct, even if you are not given feedback.”

Phase 5: Habituation

“Thank you for agreeing to participate in this study. In this phase, on every trial you will be presented with one of two words. Each word will appear for 3 seconds. Your job is to simply watch the screen. If you have any

questions please ask the experimenter now.”

48

Phase 6: Pavlovian Avoidance

“In this phase, on every trial you will be presented with one of two words. Each word will appear for 3 seconds followed by either a 250ms shock or no shock. The shock will be set at the level you have just selected. At the

end of this phase, you will be asked to make some ratings as to how much you expect shock to follow each of the words. Please follow the onscreen instructions, and if you have any questions please ask the experimenter

now.”

Phase 7: Avoidance Acquisition

“In this phase, you will again be presented, on every trial, with one of two words. When a word appears on the screen, the marked keys on the keyboard will be available to be pressed. Pressing the correct key when one of the words appears will cancel a pending shock. The same key will cancel the shock for the remainder of the

study. You can learn which key is correct by paying close attention to the screen. At the end of this phase you will be asked to make some ratings as to how much you expect a shock to follow each of the words. Please

follow the onscreen instructions, and if you have any questions, please contact the experimenter now”

Phase 8: Extinction

“This phase is important. Remember that all of the phases you have completed are interrelated and that the words seen here are the same seen earlier.”

49

Appendix 5 – Excel spreadsheets

50

Appendix 6 – SPSS outputs

51