Virtual Reality Exposure Therapy for Spider Phobia

Alexander Miloff The overall aim of this thesis has been the study of an automated virtual reality exposure therapy treatment for spider phobia. Using Virtual reality exposure therapy inexpensive hardware and not requiring a human therapist to deliver, such treatments may help ensure that the benefits of exposure therapy for spider phobia can be disseminated widely. A rigorous methodology was designed to evaluate the efficacy and safety of the novel virtual reality treatment as compared to gold-standard in-vivo exposure therapy. Effort was made Virtual exposure reality therapy for spider phobia Alexander Miloff to provide sufficient justification necessary to allow for future effectiveness trials to be conducted in community settings outside the laboratory, such as in local clinics or a home environment. A process evaluation focusing on the virtual therapist component of the treatment and requiring the development of a new questionnaire was undertaken to better understand how the treatment imparted its therapeutic effects. Finally, quantitative and qualitative baseline data concerning what participants fear about spiders was analyzed to explore how future generations of exposure therapy treatment could be improved.

ISBN 978-91-7911-114-4

Department of Psychology

Doctoral Thesis in Psychology at Stockholm University, Sweden 2020

Virtual reality exposure therapy for spider phobia Alexander Miloff Academic dissertation for the Degree of Doctor of Philosophy in Psychology at Stockholm University to be publicly defended on Friday 29 May 2020 at 10.00 in David Magnussonsalen (U31), Frescati Hagväg 8.

Abstract Exposure therapy for specific phobia involving systematic and repeated presentation of an aversive stimuli or situation is a highly effective treatment for reducing fear and anxiety. Dissemination of this evidence-based treatment has proved challenging, however, and for over 20 years an alternative method of delivery using virtual reality technology has been explored with positive results. This thesis consists of three empirical studies examining a new generation of virtual reality exposure therapy (VRET) that by using automation, inexpensive hardware, and downloadable software aims to ensure that a highly efficacious exposure therapy can be made available to almost anyone. Study I evaluated the efficacy of this novel automated VRET for spider phobia as compared to gold-standard in-vivo one-session treatment (OST) using a randomized non-inferiority design. Results indicated that large effect size reductions in self-reported fear were evident at post-assessment in both treatments and the automated VRET was not inferior to OST at 3- and 12-months follow-up according to behavioral approach test, but was significantly worse until 12-month follow-up. No significant difference was noted on a questionnaire measuring negative effects of treatment. Study II conducted a process measure evaluation of patient alliance towards the virtual therapist used in the VRET treatment with a purpose-built questionnaire entitled the Virtual Therapist Alliance Scale (VTAS). Exploratory factor analysis indicated a sound two-factor solution composed of a primary task, goal and co-presence factor and a secondary bond and empathy factor. Psychometric evaluation of the VTAS suggested good internal consistency, and a moderate correlation between the VTAS and change in self-reported fear over follow-up. Study III assessed what individuals with a fear of spiders found most frightening about spiders. Both quantitative ratings and qualitative descriptions indicated that movement characteristics were reported as most fear provoking and to a lesser extent appearance characteristics, however factor analysis of scores in these categories did not find a correlation with participant baseline self-reported fear. Overall, the above findings suggest that VRET is a potential alternative to OST for the treatment of spider phobia also with respect to therapist alliance, and spider movement characteristics should be emphasized in future VRET treatments.

Keywords: Specific phobia, spider phobia, spider, fear, anxiety, virtual reality, exposure therapy, automated, alliance, psychometric, factor analysis, virtual therapist, RCT, randomized, non-inferiority, clinical trial, dissemination, 12-month follow-up.

Stockholm 2020 http://urn.kb.se/resolve?urn=urn:nbn:se:su:diva-180746

ISBN 978-91-7911-114-4 ISBN 978-91-7911-115-1

Department of Psychology

Stockholm University, 106 91 Stockholm

VIRTUAL REALITY EXPOSURE THERAPY FOR SPIDER PHOBIA

Alexander Miloff

Virtual reality exposure therapy for spider phobia

Alexander Miloff ©Alexander Miloff, Stockholm University 2020

ISBN print 978-91-7911-114-4 ISBN PDF 978-91-7911-115-1

Cover illustration Rythme n°2, décoration pour le Salon des Tuileries Robert Delaunay (Victor Félix) Reprinted with permission Musée d’Art moderne de la Ville de Paris AMVP 2575 CCØ Paris Musées / Musée d'art moderne de la Ville de Paris

Printed in Sweden by Universitetsservice US-AB, Stockholm 2020 To my Great Aunt Helen and Uncle Peter

“VR researchers have to acknowledge the reality of inner life, for without it virtual reality would be an absurd idea… there you are, the fixed point in a system where everything else can change.” - Jaron Lanier, Father of VR

Abstract

Exposure therapy for specific phobia involving systematic and repeated presentation of an aversive stimuli or situation is a highly effective treatment for reducing fear and anxiety. Dissemination of this evidence-based treatment has proved challenging, however, and for over 20 years an alternative method of delivery using virtual reality technology has been explored with positive results. This thesis consists of three empirical studies examining a new generation of virtual reality exposure therapy (VRET) that by using automation, inexpensive hardware, and downloadable software aims to ensure that a highly efficacious exposure therapy can be made available to almost anyone. Study I evaluated the efficacy of this novel automated VRET for spider phobia as compared to gold-standard in-vivo one-session treatment (OST) using a randomized non-inferiority design. Results indicated that large effect size reductions in self-reported fear were evident at post-assessment in both treatments and the automated VRET was not inferior to OST at 3- and 12-months follow- up according to behavioral approach test, but was significantly worse until 12- month follow-up. No significant difference was noted on a questionnaire measuring negative effects of treatment. Study II conducted a process measure evaluation of patient alliance towards the virtual therapist used in the VRET treatment with a purpose-built questionnaire entitled the Virtual Ther- apist Alliance Scale (VTAS). Exploratory factor analysis indicated a sound two-factor solution composed of a primary task, goal and co-presence factor and a secondary bond and empathy factor. Psychometric evaluation of the VTAS suggested good internal consistency, and a moderate correlation between the VTAS and change in self-reported fear over follow-up. Study III assessed what individuals with a fear of spiders found most frightening about spiders. Both quantitative ratings and qualitative descriptions indicated that movement characteristics were reported as most fear provoking and to a lesser extent appearance characteristics, however factor analysis of scores in these categories did not find a correlation with participant baseline self-reported fear. Overall, the above findings suggest that VRET is a potential alternative to OST for the treatment of spider phobia also with respect to therapist alliance, and spider movement characteristics should be emphasized in future VRET treatments.

Sammanfattning

Exponeringsterapi är en effektiv behandling för att minska rädsla och ångest vid specifik fobi, och innebär att ångestväckande stimuli och situationer på ett systematiskt sätt presenteras upprepade gånger för patienten. Det har emeller- tid visat sig vara en utmaning att implementera denna evidensbaserade metod i klinisk praktik, och därför har studier de senaste 20 åren undersökt möjlig- heten att genomföra behandlingen med hjälp av virtual reality. Denna avhand- ling består av tre empiriska studier som undersöker en ny generation av virtual reality-behandlingar som är automatiserad, använder sig av billig hårdvara, och mjukvara som enkelt laddas ner då den behövs, vilket förhoppningsvis tillgängliggör exponeringsterapi för fler. Studie 1 jämförde Virtual Reality Exponerings-Terapi (VRET) vid spindelfobi med standard ensessionsbehandling, med hjälp av en non-inferiority-design. Resultaten visade att båda be- handlingarna hade stora effekter på självskattad rädsla vid avslutad behandling, och att i ett närmandetest var VRET inte signifikant sämre än standard- behandling vid 3- och 12 månader efter avslutad behandling, men var betyd- ligt sämre vid 12 månaders uppföljning. Studie 2 var en process-studie, som undersökte patientens upplevelse av allians till den virtuella terapeuten i VRET, vilket undersöktes med ett nytt mått: Virtuell Terapeut Allians Skala (VTAS). En explorativ faktoranalys fann en tvåfaktorslösning: en faktor be- stående av uppgift-, mål- och delad närvaro-skattningar, och en faktor bestå- ende av relation och empatiskattningar. Psykometrisk analys av VTAS visade god intern konsistens, och en måttlig korrelation mellan VTAS och förändring av självskattad rädsla mellan behandlingsstart och uppföljningsmätning. Stu- die 3 undersökte vad personer med spindelfobi upplevde som mest skräm- mande med spindlar. Både kvantitativa skattningar och kvalitativa beskriv- ningar pekade ut spindelns rörelser som mest skrämmande, och till lägre grad också spindelns utseende. Korrelation mellan dessa två faktorer, tagna från en explorativ faktoranalys på skattningarna, fann inget samband med deltagarnas självskattade rädsla innan behandling. Sammantaget antyder ovanstående re- sultat att VRET är ett potentiellt alternativ till ensessionsbehandling av spindelfobi också med avseende på terapeutallians och spindlars rörelseegen- skaper bör beaktas i framtida VRET-behandlingar.

List of studies

The present doctoral thesis is based on the following studies, which will be referred to by their Roman numerals:

I. Miloff, A., Lindner, P., Dafgård, P., Deak, S., Garke, M., Hamil- ton, W., Heinsoo, J., Kristoffersson, G., Rafi, J., Sindemark, K., Sjölund, J., Zenger, M., Reuterskiöld, L., Andersson, G., & Carlbring, P. (2019). Automated virtual reality exposure therapy for spider phobia vs. in-vivo one-session treatment: A randomized non-inferiority trial. Behaviour Research and Therapy, 118, 130–140.i

II. Miloff, A., Carlbring, P., Hamilton, W., Andersson, G., Reuter- skiöld, L., & Lindner, P. (2020). Measuring Alliance Toward Em- bodied Virtual Therapists in the Era of Automated Treatments With the Virtual Therapist Alliance Scale (VTAS): Development and Psychometric Evaluation. Journal of Medical Internet Re- search, 22(3):e16660

III. Lindner, P., Miloff, A., Reuterskiöld, L., Andersson, G. & Carlbring P. (2019). What is so frightening about spiders? Self- rated and self-disclosed impact of different characteristics and associations with phobia symptoms. Scandinavian Journal of Psy- chology, 60, 1–6.ii

* Reprinted with permission from the publishers

Contents

Abstract ...... 1

Sammanfattning ...... 3

List of studies ...... 5

Contents ...... 7

Abbreviations ...... 9

Introduction ...... 10

Fear, anxiety and avoidance ...... 12 Normal and abnormal fears ...... 12 Fear versus anxiety ...... 13 Symptomology ...... 14 Information processing biases ...... 15 Impact on quality of life ...... 17

Prevalence of specific phobias ...... 19

Etiology ...... 23 Historical overview ...... 23 Environmental and biological explanations ...... 25 Evolutionary perspective ...... 27

Theory-based exposure treatments ...... 30 Historical overview ...... 30 Theory ...... 30 Clinical strategies ...... 32 Pharmaceutical and other cognitive enhancers ...... 35 Exposure therapies ...... 36 Dissemination problems of exposure therapies ...... 38 Alternatives to exposure therapies ...... 39 Efforts to improve exposure therapies ...... 40

Virtual Reality ...... 44 Historical overview ...... 44 Theory and technology ...... 46

7 Early specific phobia treatments ...... 48 Treatments for spider phobia ...... 51 Modern reviews and process measures ...... 58

Alliance ...... 64

Aims of the thesis ...... 69

Overview of empirical studies ...... 70 Study I Automated virtual reality exposure therapy for spider phobia vs. in-vivo one-session treatment: A randomized non-inferiority trial ...... 71 Aims ...... 71 Method ...... 72 Results and discussion ...... 73 Study 2 Measuring Alliance Toward Embodied Virtual Therapists in the Era of Automated Treatments: The Virtual Therapist Alliance Scale (VTAS) ...... 75 Aims ...... 75 Methods ...... 76 Results and discussion ...... 76 Study 3 What is so frightening about spiders? Self-rated and self-disclosed impact of different characteristics and associations with phobia symptoms ...... 78 Aims ...... 78 Methods ...... 79 Results and discussion ...... 79

General Discussion ...... 81 The efficacy of automated VRET for spider phobia ...... 81 Theoretical implications of VRET efficacy ...... 83 Potential threats to validity of results and limitations ...... 85 The challenges and future of automated VRET ...... 88

Acknowledgements ...... 92

References ...... 95

Abbreviations

ACT Acceptance and Commitment Therapy BAT Behavioral Approach Test BII Blood-Injection-Injury CBT Cognitive Behavioral Therapy CI Confidence Intervals EMDR Eye Movement Desensitization and Reprocessing FSQ Fear of Spiders Questionnaire HMD Head Mounted Display IAPS International Affective Pictures System ICBT Internet-Based Cognitive Behavioral Therapy IQR Interquartile Range MRI Magnetic Resonance Imaging OR Odds Ratio OST One-Session Treatment PTSD Post-traumatic Stress Disorder RCT Randomized Controlled Trial SD Standard Deviation SPQ Spider Phobia Questionnaire SUDS Subjective Units of Distress Scale VR Virtual Reality VRET Virtual Reality Exposure Therapy VTAS Virtual Therapist Alliance Scale

9 Introduction

Efficacious treatments exist for many mental health problems, but there is perhaps no better example than that of exposure therapy for specific phobia. However, access has proved problematic for a variety of reasons pertaining to both patient and therapist. Due to the learned behavior of withdrawing from anxiety provoking stimuli and therefore failing to disconfirm catastrophic fears, specific phobias will typically take a chronic course unless treated, and may contribute to the development of additional co-morbid conditions.

The development of treatments for specific phobia have followed a trajec- tory similar to other medical procedures, using theory-based hypothesis testing and experimentation. Knowledge of the disorder has been recognized as far back as Hippocrates who described two individuals with unreasonable and excessive fear. Resolving repressed aggressive tendencies towards a parent figure was the solution undertaken by Freud for treating a young patient with a fear of horses. It was really cognitive-behavioral treatments, and exposure therapy specifically, however that have enabled large numbers of people to be treated successfully, systematically, and with relatively few or even one session.

This thesis summarizes the development of a new generation of treatments for specific phobia, in this case for fear of spiders. Using low-cost virtual reality technology, now easily accessible to the average consumer in middle- and high-income countries, and downloadable software accessible to anyone with an internet connection, we show that spider phobia may be treated with similar efficacy to the gold-standard treatment for specific phobia but without requiring patients to interact with a real-life spider or having to pay a psychologist for their services.

The example of automated treatments for specific phobia offers a course of action and a suggested set of steps and requirements that any therapy for a mental health care problem must meet before inexpensive and wide-spread access to a psychological treatment can be enabled: identify the active ingredients for treatment efficacy, distill those ingredients into a time-limited treatment format, use technology to reduce costs and increase accessibility, and finally incorporate supportive components in the therapy to enable patients to

complete the intervention on their own, for example by using a virtual therapist.

The development of a truly automated treatment for specific phobia is not yet complete. While we have identified an efficacious treatment for spider phobia that can be conducted by oneself, this research will need to be replicated in other populations outside Sweden, ideally using real-world effectiveness trials in outpatient clinics or people’s own homes. In the future, a deeper understanding of underlying mechanisms related to specific phobias may be identified and used to improve upon the innovations reported here. Neverthe- less, the future appears bright in so far as the work of countless researchers and psychologists have before us, when combined with a novel technology, enabled an evidence-based treatment to be accessible at low-cost to all.

11 Fear, anxiety and avoidance

Normal and abnormal fears

Fears and anxieties are a persistent part of the human condition, and yet for the most part they are not pathological. They feature prominently during the course of childhood and adolescence, are typically developmentally normative, and can even be considered a welcome sign of increasing cognitive mat- uration and sophistication (Davis, Ollendick, & Öst, 2019). At the earliest, this begins to occur from 6 to 8 months old when infants display anxiety when removed from an attachment figure (Lavallee & Schneider, 2019), and from 8 to 10 months when infants recoil when placed in the arms of a stranger, avoid visual cliffs while crawling, and rapidly detect the presence of a snake or spider (LoBue & Adolph, 2019). In the case of separation anxiety, early detection of anxiety is followed by a mounting intensity (over approx. 1 year), to a peak, even to concerning levels (at 15-18 months), and then a typical gradual decline (between 2-4 years; Lavallee & Schneider, 2019). In adulthood, healthy anxiety and fear is familiar to anyone who has had to take an exam (or completed a dissertation!), the added anxiety helpful for increasing performance as long as it remains at moderate levels (Barlow, 2002). Thus, fears and anxieties in healthy children and adults are largely transient and fleeting experiences, and dependent on environmental conditions.

Fears and anxieties become abnormal according to diagnostic manuals such as the DSM-5 (American Psychiatric Association, 2013) and ICD-11 (World Health Organization, 2018) when they interfere in daily functioning, with negative effects on social relationships, work and leisure activities. Unlike the temporary experiences of anxiety associated with particular environmental challenges, clinically diagnosed anxiety disorders must be persistent, lasting for at least 6 months in adults, and typically involving the avoidance of situations or specific objects. Fears must be actively endured, marked and be out of proportion to the real danger posed by a situation or salient object (whether that is recognized by the individual or not). At the same time anxiety and fear are dimensional constructs and while categorical diagnostic criteria can be useful to clinicians, what is normal or abnormal requires clinical judgement balancing across multiple factors, such as an understanding of the motivations and origins of the patient’s emotional response (Craske & Stein, 2016).

Fear versus anxiety

There are multiple theories for what discriminates fear from anxiety but Barlow’s (2002) definition is popular, based on whether the threat is imminent (fear) or an anticipation of a future event (anxiety). Another theory suggests that fear emphasizes alarm and physiological hyperarousal, while anxiety has a greater emphasis on mood-oriented subjective symptoms. An analogy of this difference might be an animals state during potential versus actual predator contact (Craske et al., 2009). While the DSM-5 (APA, 2013) characterizes 12 separate anxiety disorders, only one is categorized as a specific phobia, defined as fear towards a particular object or situation. Previously, the ICD-10 formally made a distinction in category based on whether symptoms fell within an anxiety or fear grouping, however the updated ICD-11 now deter- mines diagnosis based on the individual’s focus of apprehension (Reed et al., 2019). That is, classification is based on the content of the disorder, for example spiders or other highly constrained content in the case of specific phobias, as compared to worry, for example, in the case of generalized anxiety disorder.

The focus of apprehension, in the case of specific phobias, can be identified towards nearly any situation or object, and lists are available online (e.g., pho- bialist.com) totaling nearly 530 apparently distinct and referenced types (as of July 2019). In order to be a defined as specific phobia, the individual must be fearful to that object or situation almost all the time. Despite the diversity of fears possible, the vast majority of specific phobias fall into one of five categories including fear of animals (e.g., spiders), natural environment (e.g., lightning), situational (e.g., fear of flying), blood-injection-injury (BII, e.g., needles or blood) and a miscellaneous “other” category (e.g., situations that might lead to chocking or suffocation). Factor analyses of specific phobia symptoms tend to group into three types (not five), however; including animal, BII and a combined natural environment-situational phobia (Muris et al., 1999). An example of the challenge of differentiating the later combined sub- grouping is given by Barlow (2002) for an individual fearful of bridges: is it a natural environment-type phobia (i.e., fear of heights) that the person is afraid, or a situational-type phobia (i.e., something unique to bridges)? As a result of challenges like this, researchers continue to suggest these categories be col- lapsed into one (Davis et al., 2019). Identifying the phobia by name rather than type communicates more and is better supportive of planning treatments effectively according to Barlow (2002). However, the groupings can also be very useful. For example, specific phobias of similar types tend to occur together in individuals and within families, possibly aiding preventive efforts, and certain groupings have different recommended treatment approaches than others (e.g., applied tension in the case of BII; LeBeau et al., 2010). Im- portantly, the sub-divisions may be related to differences in symptomatology

13 with unique patterns of physiological and cognitive responding common to each (Muris, Schmidt, & Merckelbach, 1999).

Symptomology

Specific phobia symptoms, have in general, been described as including a “fast heartbeat, tight muscles, an urge to run, rapid breathing, feeling of doom, feeling fidgety, trembling, shortness of breath, cold hands or feet, and a pound- ing sensation in the chest” (Barlow 2002, p. 394). This description is based on the top 10 most commonly reported specific phobia symptoms (in rank order) reported by 20 patients with the disorder (Thyer & Himle, 1987). These symptoms are also similar, by some accounts, to the experience of a spontaneous panic attack, however, these patients tend to report more intense symptoms, and an emphasis, for example, on a fear of dying or losing control and somatic sensations such as dizziness (Rapee et al., 1992). The tripartite model of fear, anxiety and depression, first proposed by Lang (1968), and based on physiological indicators (i.e., somato-visceral activity such as muscle tension), subjective feelings (i.e., verbal-subjective symptoms such as worry), and behavioral actions (i.e., overt motor acts such as avoidance), can provide a helpful framework for clarifying what constitutes a specific phobia and how it is different from other anxiety disorders (see Craske et al., 2009).

Across anxiety disorders, physiological measurements of heart rate and blood pressure have tended to be consistent with subjective symptoms that individuals report (Lewis & Drewett, 2006). Anxiety disorders are indicative of an “overpreparedness” that involves a higher resting and variable heart rate as compared to healthy peers, greater electrodermal reactivity, and a resting hyperarousal not dissimilar from that of healthy subjects during anxiety in- duction experiments (i.e., participants given painful shocks during a task completion; Barlow 2002). However, there are also differences in patterns of responding between the different anxiety disorders. According to Barlow (2002), two older studies provide an example of the prototypical physiological responding by phobic versus anxious patients when confronted by aversive noise stimuli. In Lader and Wing (1964), clinically anxious and control patients were exposed to high intensity (100 db) audio while measured for heart rate and electrodermal response. Clinically anxious patients had on average 15 beats per minute higher heart rate than controls, as well as greater spontaneous skin conductance fluctuations, and a slower habituation of electrodermal responses. In a follow-up study (Lader 1967), 6 groups of individuals with different anxiety and depression diagnoses were included and again exposed to the audio stimuli. Those with specific phobia were indistinguishable from healthy controls in regards to speed of electrodermal habituation (both were quick to habituate), had few spontaneous skin conductance fluctuations, and

were generally less overtly anxious. Those in the other anxiety disorder groups, however, clustered tightly together in their responding with typical anxiety symptoms; based on cut-offs defined by the authors only 16% (3/19) phobia participants and 11% (8/71) other participants, were misclassified according to their pattern of responding.

When confronted with their unique phobic stimuli, however, those with specific phobia do share a pattern of physiological responding common with other anxiety disorders (Sarlo et al., 2002). This includes sympathetic activity, such as cardiac acceleration, increased blood pressure, vasoconstriction with increased skeletal muscle blood flow, enhanced electrodermal activity; that is symptoms consistent with escape-type behaviors. BII phobic responding is somewhat different, with a similar animal sub-type phobic response at first, followed by an abrupt relative decrease in heart rate. Previous research has shown that the decrease is sufficient to cause fainting in as many as 70% of blood phobics and more than 55% of injection phobics (Öst, 1992). These are the “typical” phobic responses, however individual unique variations also do occur. In a study of spider phobic (n = 46) and healthy female controls (n = 44) shown images of tarantula as well as neutral and positive images from the International Affective Picture System (IAPS), as expected, higher average heart rate was found for spider phobic patients toward the spider images as compared to healthy participants (Knopf & Pössel, 2009). At an individual level, only those categorized as high response phobic, however (i.e., those with the greatest heart rate difference scores between neutral and spider images) were significantly higher in heart rate toward spider images than non- phobic individuals. Those spider phobic patients with low responsiveness (one-third of spider phobic participants) actually had significantly lower heart rates than medium and high response non-phobic participants. That is on the basis of individual unique response in heart rate, a portion of spider phobic individuals were indistinguishable from healthy participants. Electrodermal activity on the other hand was consistently higher in phobic than non-phobic individuals and did not seem affected by response level to the neutral versus spider images.

Information processing biases

In addition to physiological differences as compared to healthy populations, those with anxiety disorders tend to have a variety of information processing biases that contribute to the development and maintenance of their disorder. As summarized in a recent volume of the Clinical handbook of fear and anxiety (Abramowitz & Blakey, 2020), these include selective attention, confirmation (or interpretation) bias and memory bias. Selective attention in healthy adults functions in such a way as to reduce cognitive workload by

15 prioritizing relevant stimuli and minimizing irrelevant information, and is adapted towards identifying potential threats (or potential appetitive signals like finding a mate or food). In individuals with anxiety disorders, however, such attention can become consistently focused toward even minor threats, involve a tendency toward scanning or hypervigilance, and may result in difficulty disengaging (Azriel & Bar-Haim, 2020). Those with anxiety disorders may have confirmation (or interpretation) biases which lead to erroneously attributing danger to benign, neutral or ambiguous evidence, and failing to disconfirm evidence to the contrary. For example, this may include a covaria- tion bias which indicates a propensity to see two unrelated events as being related, such as spiders always being found in particular places (Williams et al., 2009). Memory serves to support an individual’s healthy functioning by enabling recollection and recombining of past learned information to support present moment tasks, goals, self-care as well as effective planning for the future, but memory biases can also serve to sustain psychopathology, for example by the preferential encoding and recalling of threatening information (Romano et al., 2020). These memory biases tend to be related to autobio- graphical episodes or semantic knowledge about the self, reinforcing negative beliefs one has about their person and the probable outcome of an event.

Other cognitive biases in anxiety disorders exist and one such example that has received considerable attention in specific phobia literature is related to the emotion of disgust (Knowles et al., 2019). Although it is normative to experience the negative feelings of disgust when presented with certain stimuli, in those with anxiety disorder this may be connected with a fear of contami- nation, or a “looming vulnerability,” a related information processing bias that may construe a contagion as spreading, rapidly approaching and intensifying in danger and risk (Williams et al., 2009). In a recent meta-analysis evaluating disgust proneness (Olatunji et al., 2017), high anxiety symptom individuals had a large, significant difference in disgust proneness as compared to low anxiety individuals (g = 1.15, 95% CI: 0.95 to 1.34, p < .001) and this relationship was significantly stronger in studies of disgust disorders in which contagion concerns are an issue (g = 1.24, 95% CI: 0.99 to 1.49, p < .001). For this reason, disgust has frequently been discussed in terms of small animal phobias (those capable of spreading pathogen), spider phobias particularly (related possibly to hairiness and crawling movements) and BII phobias (where blood, injuries or mutilation is involved). Although it is still unclear what underlying cognitive processes contribute to psychopathology from being disgust prone, attentional avoidance such as maladaptive visual avoidance, i.e., visual attending aimed at moderating one’s anxiety rather than focusing on the threat, may be a contributing factor (Knowles et al., 2019).

In general, information processing biases lead individuals to overestimate the danger of situations or stimuli and may result in compensatory behaviors

such as a reliance on safety signals or experiential avoidance to cope with distress. In the short-term, avoidance strategies and safety behaviors can be helpful for immediate relief of anxiety such as leaving a situation or putting on special clothing to minimize risk of exposure to a spider, however in the long-term they can become habitual, rigid and automatic responses to stimuli or situations causing impairment (Hayes-Skelton & Eustis, 2020). Acceptance and commitment therapy (ACT), commonly used in treating anxiety disorders (but not typically specific phobias), nevertheless offers a potentially valuable model for framing how avoidance develops, is reinforced, and leads to psychopathology. Hayes, Strosahl and Wilson (2012), suggest that avoidance first begins at the cognitive level (one’s own inner thoughts) and attempts to suppress, modify or eliminate those that are distressing. Since most thoughts are not amenable to voluntary control, individuals have no alternative but to resort to modifying behavior (and emotions if possible). Behavior, as the theory suggests, comes under aversive (fear-based) rather than appetitive (goal-based) control, and rather than emphasizing valued-actions, individuals learn to emphasize anxiety reduction strategies. Traditional cognitive-behavioral theories of avoidance further suggest that such unproductive behaviors may continue because individuals never disconfirm their incorrect beliefs (Davis, Ollendick, & Öst, 2012). Parents may also play a role in the development and maintenance of avoidant behavior (Ollendick et al., 2012). Children with parents present during a behavioral approach test for animals were able to perform 58% of steps with a parent present and only 38% of steps without a parent present, with performance on the former predicted based on parent warmth and involvement during the test.

Impact on quality of life

The combination of pathological cognitive strategies, and physiological and behavioral responding, have broad implications for the quality of life of individuals. Prospective samples of young people (aged 14 to 24 years old) exploring the development of mental health disorders using standardized diagnostic criteria, identified significant co-occurrence between meeting criteria for an anxiety disorder and impairment in professional career and personal relationships (Wittchen et al., 1998). Self-reported severe psychosocial disa- bility during the lifetime peak of the episode was most frequently associated with specific phobias, followed by generalized anxiety disorder and panic disorder. Whereas panic disorder with agoraphobia was most often associated with impairment at work, education and household, specific phobia affected leisure time most frequently, and in the case of generalized anxiety disorder social contacts. In regards to those with specific phobias particularly, bivariate correlations between demographic variables and 30-day prevalence of specific phobias from a US National Comorbidity Survey (Magee, Eaton, Wittchen,

17 McGonagle, & Kessler, 1996) indicate that specific phobias have a significant relationship with unemployment, including being a student (OR 1.49, 95% CI: 1.01 to 2.21), being a homemaker (OR 2.48, 95% CI: 1.44 to 4.26) or other (OR 2.91, 95% CI: 1.66 to 5.07), as well as living with parents (OR 2.41, 95% CI 1.46 to 3.98), and being in the lowest income bracket (OR 2.02, 95% CI: 1.16 to 3.50). More than a third (34.2%) of the sample of those with specific phobias indicated their disorder “interfered a lot” with their “life and activities”.

Detailed research on the relationship between marriage, divorce and mental disorders was explored in a cross-national study using epidemiological data from 19 countries (Breslau et al., 2011). Specific phobias along with major depression and alcohol abuse, of the eighteen disorders included in the analysis, were associated with the largest population attributable risk proportions for increased divorce (1.3, 4.0 and 2.9% increase, respectively). Specific phobia (but not any other disorder) was also significantly and positively associated with increased prevalence of early marriage (3.6%; marriage before 18 years). In addition to problems related to professional and leisure activities, as well as personal relationships, phobias are associated with multiple negative repercussions from failure to complete required medical tests such as the use of imaging technology, for example magnetic resonance imaging (MRI) due to claustrophobia, completing blood tests as a result of BII phobias, and fears related to and preventing dental procedures (as reviewed in Choy, Fyer, & Lipsitz, 2007; Wolitzky-Taylor, Horowitz, Powers, & Telch, 2008)

In summary, fear and anxiety are a persistent part of the human condition from early life until adulthood, however individuals will meet criteria for one or more specific phobias only when the fear is persistent and impairing. Spe- cific phobias suggest an overestimation of the real danger of a situation or stimuli. The pattern of responding when encountering the object suggests imminent threat, with consistent patterns of physiological, emotional and behavioral responding. These patterns of responding are sustained in part by pathological forms of information processing and use of avoidance and safety behaviors as a coping strategy. The repercussions for the health of the individual are negative, including deficits in a wide-ranging number of activities from personal life, to career and education.

Prevalence of specific phobias

Unreasonable fears are remarkably common. In a book section describing how to conduct exposure therapy, the clinician and researcher Lars-Görgan Öst (2012, p. 83) catalogued a list of the common catastrophic fears patients report when entering therapy. A spider phobic patient states that a “spider will crawl underneath my clothes and I will get a strong panic attack and die of heart failure.” An individual suffering from fear of enclosed spaces reasons that upon entering an elevator that it “will get stuck between two floors, no one will hear the alarm signal, the air will run out, and I will die from suffocation.” Despite the obvious discomfort these thoughts might have, individuals who report them will not be expected to meet criteria for specific phobia unless their avoidance and distress leads to impairment. According to a recent Lancet Psychiatry review which summarized data across three nationally representative surveys in the United States and the Netherlands, approximately 20% of individuals report a lifetime below clinical threshold fear of heights, 10% a fear of enclosed spaces, and 13 to 22% a fear of animals (Eaton, Bienvenu, & Miloyan, 2018). Of these individuals, the conditional probability of meeting diagnostic criteria given the presence of a fear (irrespective of which sub-category) is 25-30% and similar between the United States and the Netherlands.

Among mental disorders, specific phobia is considered the most prevalent condition along with depression and alcohol-use disorder (Eaton, Kessler, Mortensen, Rebok, & Roth, 2019). For those meeting diagnostic criteria for specific phobia, the lifetime prevalence of the disorder was 7.2% in adults (IQR: 4.0-10.4), based on 25 community-based studies included in the Lancet Psychiatry article (from 1984 and 2016). This is similar to a separate analysis of 25 cross-sectional World Mental Health surveys conducted between 2001 and 2011 which found a lifetime global prevalence of 7.4% across 22 countries (Wardenaar et al., 2017). Among specific phobia subtypes globally, lifetime prevalence figures range from 2 to 6% for any one sub-type, with animals (3.3 - 4.7%) and heights (4.5 - 5.3%) being consistently among the most frequently reported (Eaton et al., 2018). Half of all individuals with a lifetime specific phobia diagnosis reported a fear of heights or animals according to the United States National Epidemiological Survey on Alcohol and Related Conditions (Stinson et al., 2007). As would be expected, 12-month prevalence

19 rates for specific phobias are lower than lifetime prevalence with a recent systematic review identifying a median 12-month prevalence figure of 4.8% (IQR: 3.5 - 7.3%; Eaton et al., 2019).

For policy decisions, lifetime prevalence and 12-month figures are important measures, however they may suffer from recall bias and only report whether an individual has met criteria up until that point (Kessler et al., 2012). Therefore, lifetime morbid risk of a disorder, a measurement predicting the proportion of people that will eventually develop the disorder at some point in their lives based on combined observed lifetime prevalence data and projec- tions about future onset, can also be useful. Kessler et al. (2012) conducted such an analysis using US household population data (individuals aged 13 and older) to construct survival models that predicted 18.4% of Americans would meet criteria for specific phobia in their lifetime; a higher number than reported lifetime prevalence figures provided in the same study (15.6%). Ac- cording to the authors, this puts specific phobia second only to major depression for greatest lifetime prevalence and lifetime morbid risk for anxiety or mood disorders in the United States.

In the Lancet Psychiatry review, regional population characteristics were noted and summarized. East Asian populations tended to figure lower in prevalence than European and American surveys. In China, Japan and Korea, prevalence rates were all below 4%, whereas in 12 Europe and North American surveys (with the exception of Italy at 1.5% and Poland at 3.4%) national figures were above 6.5%. New Zealand was categorized with Asia but had a prevalence figure closer to Europe and North America at 10.8%. In general, prevalence figures for specific phobia are higher in high- and middle-income countries (8.1 and 8.0%) versus low-income countries (5.7%), according to the World Mental Health Surveys (Wardenaar et al., 2017). Despite relatively consistent figures at the global and regional level in prevalence, there can be high variability at the local level. As reported in Lancet Psychiatry, the City of Oslo for example, had a prevalence figure of 14.4% versus rural Norway at 6.5% (total respondents), and the city of Baltimore (23.5%) versus New Ha- ven, Connecticut (8.5%; female respondents only). The authors of the review suggest that as yet unknown risk factors might be responsible for prevalence differences among nearby localities.

Related to the incidence of specific phobias, age-of-onset can be an important source of information for prevention efforts and to ensure early treatment. The onset of specific phobias is considered one of the earliest of any mental health disorder. Kessler et al. (2007) identifies a median age-of-onset for specific phobias of between 7 and 14 years (IQR: 4 - 20). Given the relative lack of prospective cohort studies similar to research for cancer and cardiovascular disease (the authors note), these figures have been based primarily

on retrospective World Mental Health Surveys. An updated World Mental Health Survey analysis (Wardenaar et al., 2017) indicates 8 years as the median first onset (IQR: 5 – 13). As compared to other mental disorders such as impulse-control disorders (median onset age of 7 to 9 years), these figures for specific phobia are similar but perhaps slightly later than that very early and narrow age band, but almost identical to the estimated distribution for separation anxiety disorder (Kessler et al., 2007). Given the early onset of specific phobias and the condition being known as an important predictor of panic disorder (relative risk = 4.38), generalized anxiety disorder (relative risk = 4.10) and other serious conditions (Lieb et al., 2016), with adequate and prompt access to care, young patients with specific phobias could potentially be pre- vented from developing additional co-morbid conditions.

It is important to note, however, that rates of incidence may fluctuate with specific phobias, and in a way that is more varied than other disorders such as depression, where the predominant incidence is early to mid-adulthood (three- quarters before age 42; Eaton et al. 2019). Specific phobia incidence may occur at any time in the lifespan and is not only isolated to the early years. A longitudinal cohort study of 175 000 Baltimore residents, interviewed at four time periods starting in 1981 and ending in 2005 (Eaton et al., 2007), indicated that although peak incidence occurred below age five (often before individuals could remember) there were additional peaks of incidence in later years (Eaton et al., 2019). These variations appeared to be largely gender-specific. After age five, peaks in new incidence occurred at age 35, 55 and 75 years among female respondents. The authors suggest that changes in incidence rates may be related to environmental and social factors at different developmental stages. Overall though, women are more likely to report specific phobias then men. According to the Lancet Psychiatry article (Eaton et al., 2018), this was the case for all community-based samples included in their systematic review. In North America and Europe, approximately twice as many women report lifetime prevalence of a specific phobia, however there have been reports of isolated differences. For example, in rural Norway approximately four times as many women report with specific phobia than men.

Once an individual receives a diagnosis for specific phobia the disorder tends to follow a chronic course. Mean duration of an episode has been reported as 20.1 years in those with a lifetime specific phobia diagnosis (Stinson et al., 2007). This can also be assessed by interviewing individuals receiving a diagnosis in the year preceding a follow-up interview. In the first year after diagnosis, according to United States Epidemiological Catchment Area data reported in Eaton et al., (2018), fear of heights was the most persistent disorder with as many as 20% of individuals continuing to receive a diagnosis. This was followed by fear of water (17%), fear of storms (14%), and fear of flying and animals (12%). At the other extreme (12-years follow-up), fear of crowds

21 for which only 6% of individuals at 1-year follow-up continued to receive a diagnosis, was the most persistent at 28%. Next, however was fear of animals at 14% indicating little fluctuation over the more than decade time period, and fear of heights, formally the most persistent at 1-year follow-up, which also continued to have high chronicity at 12-years (11%). In comparison, a separate nationally-representative United States survey (National Epidemiological Survey on Alcohol and Related Conditions; Eaton et al., 2018), identified a narrower band of persistence (11 - 19%) at 3-years follow-up among 10 separate specific phobias diagnoses, and according to National Comorbidity data, at 10-years follow-up, a much higher overall but still narrow band of persistence across seven specific phobias (25 - 38%).

In summary, specific phobias are very common, similar in prevalence to the other most frequently reported mental health disorders, but current figures in the range of 7% are still likely an underestimate. A large proportion of individuals have a subclinical threshold fear resulting in distress but never report this as a disorder, and some may fail to recall previous episodes during surveys. Overall women are more likely to report a specific phobia than men. Prevalence of phobias are higher in some populations than others (European and North American as compared to East Asian) but high variation is found even at the local levels, for example between cities and rural areas. Fear of heights and animals are reported by half of all individuals with a specific phobia. The median age of onset is among the earliest of all disorders but first incidence may occur at any time, possibly in relation to social and environmental changes. Given that specific phobias can become a chronic condition and early onset may predict other serious conditions the early treatment of phobias should be prioritized.

Etiology

Historical overview

The story of Little Albert is synonymous with research on classical conditioning in humans and is among the most widely cited of all psychology studies (Harris, 2002); it also happened to be a powerful early demonstration that fear can acquired and the mechanism tested empirically (Lipp et al., 2019). Sometime during a four-month period in 1919, John B. Watson, then a Pro- fessor of Psychology at Johns Hopkins University, and his assistant Rosalie Rayner, conducted a series of conditioning experiments with a healthy 9- month-old boy referred to as Albert B. The founder of the influential Behav- iorist School of Psychology, Watson in a 1913 publication entitled Psychology as the Behaviorist Views it, stated his aim to ensure psychology became an “experimental branch of natural science which needs introspection as little as do the sciences of chemistry and physics” (Watson, 1913), and suggested that “within normal biological limits, children were made, not born” (Bigelow & Morris, 2001). The Little Albert study was a natural extension of his theoretical model of emotional development and had as its goal an empirical test of whether these characteristics could be conditioned by events, persons and objects in their environment and therefore whether a person’s emotional life could be brought under experimental control (Harris 2002).

In the series of experiments, Albert was first presented with a number of inanimate objects (such as cotton, human masks and burning papers) as well as live animals (such as a rabbit, dog, monkey and rat), apparently showing little emotional reaction to these. As one would expect, however; the child was startled and would begin to cry when a metal bar was struck behind him. Sub- sequently, Albert was presented with the loud noise while touching a white rat. The initial phase of conditioning was replicated seven times across two sessions (conducted one week apart), and following these pairings Albert re- acted with crying and avoidance whenever the rat was presented. This occurred even without the simultaneous presentation of a loud sound. In sessions conducted after five days, Albert continued to display a strong fear response but now also to other furry objects (such as a rabbit) indicating a generalization to similar objects. The same generalization did not occur to playing blocks, however, and other non-furry objects.

Today these experiments would be considered highly unethical, and they also had numerous flaws in experimental design, such as reliance on a single subject, overly subjective interpretations of the child’s behavior, confounding on some trials (such as the barking of a dog frightening the child), and a seem- ingly reduced and possibly inconsistent phobic response in later sessions after which the child was reconditioned (Harris, 2002). Nevertheless, the study had an important impact on later generations of psychologists. A recent review of novel approaches to strengthening fear extinction in humans (Lipp et al., 2019) suggests the work provided an underlying foundation for fields as diverse as development psychology, human emotional learning, experimental psychology, and experimental psychopathology. Perhaps, the most important effect of this work, however was inspiration for the development of treatments for anxiety disorders using behavior therapy. Prior to this, psychodynamic therapists from Freud onwards asserted that irrational fears toward an object were based on activities of the id (such as anger or hostility). These intense feelings were liable to become repressed and projected onto symbolic objects in the environment, requiring awareness of the underlying impulse to over- come (see Little Hans horse phobia case; Thorpe & Salkovskis, 1999). Both Joseph Wolpe and Hans Eysenck, founders of Behavior Therapy provided a new alternative framework for understanding fear, used the little Albert story to emphasize that phobias could be “induced,” that a neutral stimulus could acquire the ability to evoke fear depending on how it was presented, and in time they pioneered novel and powerful ways to extinguish it (Harris 2002).

According to Wolpe, it was Ivan P. Pavlov and his student M.N. Eroféeva who deserve primary credit for elucidating the principles that underlie how fears develop (Wolpe & Plaud, 1997). Wolpe’s research emphasized Pavlov’s experimental conditioning procedures and based on laboratory experiments with cats would go onto develop Systematic Desensitization, the first empirically supported behavior therapy treatment. In this paradigm, specific phobias and other anxiety disorders are progressively treated using a counterconditioning methodology (i.e., reciprocal inhibition), weakening the relationship between stimulus and fear by incorporating an appetitive reinforcer (in the case of cats, food, but in the case of humans, progressive muscle relaxation), as well as other components such as reducing negative reinforcers (i.e., safety behaviors). Hans J. Eysenck, on the other hand, emphasized another area that Pavlov had pioneered which was his system of personality types, the basis for which he assumed were natural variations in excitatory and inhibitory components of the nervous system. Pavlov had identified differences in personality types among dogs such as friendliness or aggression, but Eysenck related these to human differences, such as introversion and extroversion for which baseline cortical arousal levels were either high or low, respectively. Rather than the

acquisition of fear being entirely environmentally-derived, this research suggested as an interplay between conditioning and what Eysenck assumed was a genetic component. This ensured that some individuals would be more likely to develop a specific phobia than others.

Environmental and biological explanations

Today the view of the etiology of specific phobias and other anxiety disorders continues to be that their development is based on both environmental conditioning and genetic predisposition. In terms of environmental conditioning there are numerous ways that early upbringing such as family life plays a role in determining the health of children and providing for the critical skills needed for their resilience. This includes everything from parenting style (harsh or supportive), neighborhood security (safe or violent), as well as financial, nutritional and institutional access (Yoshikawa et al., 2012). Child- hood maltreatment such as physical, sexual, emotional abuse and neglect have all been linked to internalizing and externalizing disorders that underlie many mental health problems (Vachon et al., 2015). Whereas previously there was an assumption that certain forms of maltreatment were more deleterious to mental health outcomes than others, findings by Vachon et al., (2015) aggre- gating data over nearly three decades suggests that exposure to the diverse types of abuse described above (even emotional) have roughly equivalent negative outcomes for the child.

There may be a range of explanations for these results, and evidence of the biological effects of stress from childhood onwards provide one rationale (McEwen, 2007). This includes hormonal changes, such as dysregulation of cortisol production, and metabolic changes such as reactivity to hormones including insulin, leptin and ghrelin. Structural remodelling of the hippocampus, pre-frontal cortex and amygdala, all responsible for cognitive and emotional processing, can be observed in animal models. Recently correlations have now been made to deprivation in childhood and human brain development (Hair et al., 2015). Findings from this longitudinal cohort study indicates that children in low-income households scored significantly lower on academic and cognitive test scores, with maturational lags in frontal and temporal lobes based on MRI scans explaining as much of 20% of variation.

Socio-economist status indicators such as educational attainment and income status have been identified as particular risk factors for the development of anxiety disorders. In samples collected in the United States (as reviewed in Eaton et al., 2019), education level defined as failure to complete high school (vs. college completion) was associated with a 2.8 times more common diagnosis of an anxiety disorder, and annual income less than $20,000 (vs. >

25 $70,000) a 2.1 times more common diagnosis. In regards to specific phobias, particularly, they are more prevalent in those without high school completion than those with (11.4 vs. 8.64%), and are 1.73 times more commonly diagnosed in high school educated as compared to those with college education. While in general anxiety disorder differences between urban and rural populations are considered largely minimal, this is to a lesser extent the case for specific phobias which show that those living in rural areas have a 1.21 greater chance of receiving a diagnosis.

There is a clear relationship between parental anxiety and the development of anxiety within children: anxiety “runs in families” as described by Beesdo- Baum & Knappe (2012). As discussed previously, exposure to deprivation, neglect and stress within the family is a risk factor for the development of mental illness, however there is increasing evidence that genetics rather than environment may predominate in the transmission of anxiety disorders. As reviewed in Shimada‐Sugimoto et al., (2015), risk of developing an anxiety disorder is heightened (but still moderate) in children if their parent has the disorder, with estimates of heritability in the range of 30 to 50%, and based on a data from twin studies, 30 to 60% for phobias (including agoraphobia and social phobia). There may also be differences in heritability between specific phobia sub-types (Van Houtem et al., 2013). According to authors, the two highest reported fear and specific phobia sub-types (fear of animals and BII phobias) had a heritability of 45% and 33%, respectively. In providing an overall estimate of heritability, Eaton et al. (2018) summarized data across 10 independent twin studies finding a mean heritability for animal, situational and blood-injection sub-types of approximately 30%.

Genetic influence in the development of anxiety disorders could be related to how the brain processes fear. Those more prone to anxiety disorders have a greater reactivity in fear processing areas such as the amygdala and insula than anxiety-normal individuals, i.e., those scoring in the top 15th percentile on trait-anxiety versus those in 40 to 60th percentile (Stein et al., 2007). Those individuals diagnosed with an anxiety disorder such as specific phobia (as well as PTSD and social anxiety disorder) also have greater baseline activation in the same brain regions as compared to healthy controls (Etkin & Wager, 2007). Healthy controls will, however, show patterns of activation similar to those diagnosed with an anxiety disorder when confronted with a fear conditioning task.

Evolutionary perspective

Given the genetic component for the development of specific phobias, and that certain specific phobias predominate over others such as spider and snake phobia, it is natural to assume there is an evolutionary rationale for the development of these disorders (“evolutionary fear-relevant” stimuli; Soares, Lindström, Esteves, & Öhman, 2014). Snakes still cause widespread harm to humans with as many as 100 000 annually dying as a result of snake bites, up to 400 000 a year facing lasting impairment, and share an estimated 100 mil- lion years of evolution in which they mutually interacted with us in a predator- prey relationship. While only 0.5% of spider species are harmful to humans (200 of hundreds-of-thousands), mortality is limited (approx. 200 deaths annually), and spiders primarily prey on insects as compared to snakes which are carnivorous, they do have a long-shared history with our early ancestors, and can still be harmful. Research by New and German (2015) give the example of Latrodectus (the widow spiders) as an evolutionary rationale for the continued fear of spiders in modern humans. The Widow spiders have verte- brate-specific neurotoxins that cause severe nausea, muscle cramps and pain upon envenomization, and prior to antivenom could be lethal in healthy adults (estimated at 4 to 8%) and even more so in certain populations (the old, young, those who are pregnant and ill). This particular species now exists on every continent except Antarctica and has a particularly high-density population in Southern Africa (Garb et al., 2004) where it, like recent research suggests, anatomically modern humans originated approximately 200 thousand years ago prior to first migrations (Chan et al., 2019). Given the shared recent history (relatively speaking) and likelihood that populations of hominid lived side-by-side for many millions of years earlier with poisonous spiders, it would not be surprising that our ancestors developed highly relevant adapta- tions for dealing with them (of which Latrodectus is just one example). Simi- lar sets of hypotheses could be generated for why fear of heights, storms and other small animals are common types of specific phobia in modern populations.

Research by Arne Öhman at Karolinska University suggest that evolutionary relevant threats may receive special “preferential access” to a threat detection system based on their low-level sensory features and cues, triggering arousal (Öhman & Mineka, 2001). Unlike familiar slow, preferential and con- scious attention, in order to keep the individual safe at risk of death, the detection system would need to be pre-attentive, automatic, resistant to cognitive

27 control, and capable of rapid, reflexive avoidance of danger. Over many generations of cue and consequence understandings, as brought about by natural selection, conditioned responses to fear could develop into innate defensive responses, and be physically located in an increasingly specialized and advanced nervous system as a module (e.g., amygdala; the “fear module”). Öh- man and Mineka (2001) remind readers that evolutionary rationale for psychological principles are not unlike any other scientifically derived argument for a psychological theory requiring empirically testable hypotheses and rul- ing out of competing theories, however researchers in this area have begun to do just that.

As an important first step, adequate defense requires reliable and early detection of a predator or source of harm, and therefore evolutionary rationale for the development of phobias requires exploration of perceptual processes (Öhman and Mineka, 2001). This detection system would not just be relevant for specific phobias like snakes and spiders in which relatively simple shapes or types of movement could trigger responses, but also be relevant for social anxiety disorders in which faces or physical features are identifiable. New and German (2015), suggest that spiders unlike snakes are a particularly good test of the visual cognitive systems inherited (rather than learned) attentional pri- orities given that spiders are not currently a significant survival risk although they may have been historically. In addition, unlike certain complex shapes which might have to be learned (such as a polar bear or wolf), the template of a spider is relatively invariant, and typically referred to as a central mass from which multiple segments radiate. Using an inattentional blindness task in which unexpected stimuli are presented peripherally to a central task-oriented stimulus, New and German (2015) found that the prototypical template shape of spiders presented without prior warning were reliably detected, located and identified by a majority of participants, irrespective of their level of fear of spiders, whereas the shape of a housefly and hypodermic needle were not, despite the former also being an animate object (and insect) and the later con- ceptualized as being a threatening object. This reliable detection occurred both for spiders with curvilinear and rectilinear forms (natural curved shapes or straight legged spider images) as well as spider shapes in which the prototypical central mass and radiating segments were scrambled.

Previous research has shown similar faster identification of spiders and snakes (fear-relevant) as compared to fear-irrelevant items (flowers and mush- rooms) in a grid-pattern with and without distractors (Öhman et al., 2001). Participants here were told to simply watch for discrepant category images in this study (requiring top-down attention and vigilance) versus the New and German (2015) study which used an attentional blindness task (asking participants to compare the length of two bars) after which stimuli were briefly flashed in the periphery of their vision. This attentional blindness task was

particularly geared towards recreating the type of situation in which a spider might actually be found in one’s environment i.e., without foreknowledge, warning, and unrelated to one’s current task. Nevertheless, in the Öhman et al. (2001) study, participants here who met self-report cut-offs for fear of snakes or spiders were more rapid at identifying the stimuli that they feared than the alternative. A more recent study by Öhman and colleagues (Soares et al., 2014) compared those with variable levels of fear of spider and fear of snakes, reflecting the general population. Their findings suggest that spiders were more efficiently detected when presented in central vision, but snakes were detected with relative invariance across the visual field suggesting a human visual field capacity for orienting toward the ambush hunting-type scenarios that snakes are known for, and would make sense in the context of our shared evolutionary history with these fear-relevant creatures.

In summary, the etiology of specific phobias and anxiety disorders have biological, environmental and evolutionary rationale for their occurrence. These include natural genetic predisposition in populations and families, neighborhood environments emphasizing safety or insecurity, access to education and support systems or deprivation and abuse, and an evolutionary rationale that posits an innate threat detection system (“fear module”) providing reflexive, automatic detection and avoidance of danger. It has also been shown that fears can be acquired and extinguished. Given how effective systematic counterconditioning effects such as exposure therapy are, the classical conditioning model has proven particularly popular for understanding the etiology of specific phobias, and continues to be developed upon today.

29 Theory-based exposure treatments

Historical overview

Exposure therapy-based treatments conducted alone or with cognitive elements are considered a modern, highly effective approach to the treatment of many anxiety disorders (Norton & Price, 2007), particularly specific phobias (Choy et al., 2007), and have been developed over decades of careful research and experimentation (Barlow, 2002). However, the first use of exposure therapies for the treatment of anxiety appears to be quite old. While it took until the 17th and 18th Century for the clinical description of a phobia first described by Hippocrates to take root as a unique medical condition in Europe (distinct from melancholy; Crocq, 2015), already in the 9th Century Aby Zayd Ahmed ibn Sahl al-Balkhi of Khorasan (an Arab polymath) proposed a nosology of phobia quite similar to the DSM-5 (Awaad & Ali, 2016). The definition contained in his book the Masalih al-Abdan wa al-Anfus (Sustenance of the Body and Soul) differentiated present fear from future anticipatory anxiety and identified physiological symptoms of fear such as trembling hands and feet or pale skin from the withdrawal of blood towards the organs. The exposure therapy approach suggested by al-Balkhi was also similar to what would be prescribed nowadays. A series of examples were used to convince readers of his approach, including the suggestion that the anxiety experienced by sailors towards large waves and storms is reduced as a result of continued time and exposure to the sea, and whereas a horse may fear an object at first, training the horse to become accustomed to the object by gradual presentation eventually reduces its anxiety.

Theory

This basic description of a behavioral treatment by al-Balkhi shares much with our modern understanding of fear reduction (Maples-Keller & Rauch, 2020). A horse will tend to naturally reduce its fear to a novel object by continued exposure to that object in the absence of negative consequences (extinction learning). The same underlying mechanism of fear reduction is found in humans (the sailor, for example) by continued and repeated exposure to the presence of waves (exposure therapy). However, if a sailor, fearful of the

waves avoided the inevitable but temporary physiological and cognitive discomfort by staying on land, he or she would not have the opportunity to disconfirm their beliefs about its danger. The mechanisms underlying habituation to discomfort, required for all exposure therapy based-treatments of anxiety disorder, are still based on theories originally proposed by Foa and Kozak (1986).

These authors suggest that fear is held in memory as a program intended to escape danger, composed of a network of three parts including information about the aversive stimuli itself, physiological, verbal or behavioral responses to the stimuli, and interpretative meaning about the stimulus-response elements. In order to modify the fear structure, it must be elicited (by being exposed to waves, for example), and because it is an escape-oriented program it must inevitably involve consequent physiological activation, without which the fear structure could not have been elicited. The proverbial sailor on the sea or patient in the clinic must withstand the effects of anxiety until it subsides in order for new more adaptive information about the stimuli (the absence of catastrophic effects of the stimuli, for example) to be added to the memory structure. This process is referred to as emotional processing (Foa & Kozak, 1986). Our understanding of how the fear structure is modified, however, is currently an active area of research, the consequences of which has broad implications for how exposure therapy treatments are conducted.

In the emotional processing theory of fear reduction, modification of the fear structure occurs through the incorporation of novel information incom- patible with the original structure: as a result, preexisting links between stimulus and response are weakened or replaced. Certain implications arise from this theory. For example, because the fear structure must be elicited in order to be modified, the greater the fear experienced by the patient (i.e., the greater the fear structure elicited), the greater the potential corrective learning that can be obtained (Weisman & Rodebaugh, 2018); and the longer the period of time the patient spends with the phobic object in a habituated state (i.e., weakened stimulus-response links) the greater the beneficial outcomes that can be expected. Also, those with a more coherent fear structure or rather those with fear structures containing less contrary information about the nature of the phobic stimuli (information not reflecting reality) are hypothesized to be those patients with the more intense phobias (Lang 1977 in Foa and Kozak, 1986).

New research, however points to an alternative hypothesis about the mechanism of extinction learning and therefore how exposures should be carried out. Inhibitory learning theory, proposed by Michelle Craske and colleagues (Craske et al., 2014), while still founded on basic extinction learning principles may provide a model more in keeping with neural mechanisms of how fear extinction occurs and the realities of clinical experience (Weisman &

31 Rodebaugh, 2018). For example, top-down processing related to executive function of the medial pre-frontal cortex innervate and appear to inhibit the amygdala, an emotion center key to anxiety disorders (Milad et al., 2014). Evidence collected by Craske et al. (2008) indicate that initial fear activation and habituation within-session are not necessarily related to eventual outcomes (McMillan & Lee, 2010). The reacquisition of fear that occurs sometimes months or even years after a patient was successfully treated suggest that the integrity of the fear structure remained intact (Vervliet et al., 2013). In the inhibitory learning model, fear structures are not erased but rather new competing learning is developed capable of inhibiting the fear structure. In addition to being a more parsimonious theory for explaining the effects of exposure-based treatments than those proposed by earlier researchers (at least the classical model by Foa and Kozak, 1986), the theory also provides certain potentially fruitful predictions and strategies for improving the efficacy of treatments (Weisman & Rodebaugh, 2018).

Clinical strategies

Inhibitory learning theory strategies for augmenting exposure therapy can be summarized as emphasizing: a) strategies for developing non-threat associations, and b) the successful retrieval and accessibility of new learning. A recent review by Weisman and Rodebaugh (2018), summarized work in the area by identifying a variety of specific approaches for maximizing exposure therapy. According to the authors, the strategy with perhaps the strongest basis in theory and evidence is maximizing expectancy violations, and the authors suggest this can even be considering a guiding principle for improving exposure therapies. One practical implementation of this is framing exposures in terms of behavioral tests (i.e., first identifying the patient’s pre-existing negative beliefs). Although, exposure alone appears to be effective (Norton & Price, 2007), the addition of the cognitive element prior to each exposure may improve outcomes (McMillan and Lee, 2010). Certain kinds of cognitive therapy performed prior to exposure treatments should be minimized, however; these include efforts to lessen the probability of overestimation (i.e., cognitive restructuring), as this could reduce the mismatch of expectancy and actual outcome (Craske et al., 2014). Less negative appraisals are a natural outcome of exposures in any case as suggested by a clinical handbook (Sewart & Craske, 2020) and effort should instead be placed on cognitive restructuring during the post-exposure consolidation phase of treatment.

Acquisition of new learning may be aided by limiting distraction during exposure tasks to improve awareness of the nonoccurrence of expectancy and increase salience of the excitatory stimuli. This could involve the removal of mobile phones, for example. A parallel concept to eliminating distraction

would be increasing attentional focus (Sewart & Craske, 2020). Performing actions in direction violation of one’s natural fear tendency (fear antagonistic actions) is one means of increasing the salience of tasks, such as the example given by Weisman and Rodebaugh (2018) of an acrophobic patient running (not walking) towards a railing. Certain kinds of distraction can prove beneficial to treatment outcomes, however (e.g., humorous chats with the therapist) by helping patients to see the phobic object from a new perspective (Ventis et al., 2001).

Second to maximizing mismatch with expectancy, another overarching strategy recommended by review authors is increasing the variability of exposures. Unlike the previous strategy that aimed to develop non-threat associations, the benefits of exposure variability are geared toward improving retrieval of newly learned information. Sewart and Craske (2020) suggest that exposing phobia patients to multiple types of spiders, for example, or randomly ordering the intensity of heights, may improve the capacity for new learning by enhancing attending to novel and unexpected stimuli, generating schema that can be applied across a range of fear provoking situations, and facilitating retrieval by increasing the number of stimuli cues stored in memory. Other methods of increasing variability include random spacing of exposure trials (as compared to massed schedules), and conducting exposures in multiple contexts (or even times of day), and in unfamiliar environments (e.g., with or without a therapist present). Weisman and Rodebaugh (2018) suggest that given the apparent success of these new strategies for improving outcomes, the traditional, slow graduated increase in fear intensity may be worth abandoning in exchange for more variability in exposures.

Other related and potentially important strategies emphasized by Weisman and Rodebaugh (2018) include deepened extinction learning and affect labelling. Deepened extinction learning involves extinguishing individual aversive stimuli before incorporation in future trials with non-extinguished stimuli. In order for this to be successful, all stimuli should be predictive of a similar outcome. For example, extinguishing fear towards one type of spider can be performed prior to incorporating that same spider in a trial with a novel spider type. The benefits of this approach may involve the increased expectancy associated with multiple negative stimuli combined with subsequent violation (Sewart & Craske, 2020). Affect labelling (also called linguistic processing) involving verbalization of one’s moment-to-moment emotional experience either in written or spoken form may be useful for enhancing inhibitory regula- tion (perhaps via cortical inhibition of the amygdala). It remains uncertain, however, whether labelling emotional material or labelling any kind of ongoing process is most effective, for example stating one’s fear when interacting with a snake versus verbalizing the texture of the snake’s skin being held in hand.

Efforts to eliminate safety signals and behaviors during treatment have been incorporated since the early days of exposure therapy (e.g., systematic desensitization; Wolpe & Plaud, 1997), and they continue to be important in inhibitory learning paradigms of treatment. They aim to reduce misattribution of safety to the safety indicator thereby strengthening behaviors used to avoid aversive stimuli and increase the salience of excitatory stimuli; these also ensure cognitive resources be devoted instead to disconfirming beliefs. Removal of safety signals could relate to particular individuals or reducing use of anti- anxiety medications (Sewart & Craske, 2020). Some safety behaviors can be quite subtle, however, and it is unclear whether the availability of the safety indicator or its utilization is what interferes with learning (Powers et al., 2004). There are also potential advantages to the use of specific safety behaviors in terms of improving the tolerability of exposure therapy (Rachman et al., 2008), however recent clinical guidelines suggest eliminating safety behaviors as soon as the client is willing (Sewart & Craske, 2020).

Other strategies reviewed by Weisman and Rodebaugh (2018) include those that rely on relatively complex memory-based mechanisms but that so far lack empirical evidence. These include occasionally reinforced extinction, in which an aversive outcome is paired with a feared stimulus only sometimes. Using this strategy, a patient would be less likely to predict when a negative outcome would occur (as opposed to seeing all outcomes as negative), and it may increase saliency and expectancy associated with the stimuli, thereby enhancing expectancy violations. A clinical handbook suggests the use of this method only in later phases of treatment and recognizes it may not always be ethical to implement, for example, in the case of exposure to a snake bite (Sewart and Craske 2020).

Use of retrieval cues have had some positive laboratory success, by using cues from an early successful extinction context in a novel context, and therefore helping patients activate earlier memories of extinction but in a new environment. A retrieval cue could include use of a particular wristband, for example. A more practical use of cues (particularly outside the laboratory) may include cognitive mental reinstatements. Unfortunately, evaluation of this approach may be limited by the difficulty of empirically manipulating a mental cue (Weisman & Rodebaugh, 2018). There is also the risk that cues acquire inhibitory value and become a safety signal; as a result, in general use of cues should be used sparingly and towards the end of treatment to prevent relapse (Sewart & Craske, 2020). Finally, strengthening reconsolidation of newly- learned associations by introducing feared stimuli for short periods (10 to 30 min) prior to normal exposure tasks has been proposed (referred to as capital- izing on reconsolidation). This may be a method of enhancing the accessibility

of new information and strengthening its storage in long-term memory, however much effort would appear to be required to identify optimal presentation and timing of pre-exposures and clarifying the risks of reactivation serving instead as an ineffective exposure.

Pharmaceutical and other cognitive enhancers

Although not specific to inhibitory learning strategies, other theoretical paradigms may play a role in improving exposure therapy treatments. Cognitive enhancers such as D-cycloserine (DCS), a pharmaceutical and partial agonist of N-methyl-d-aspartate (NMDA) receptors located in the amygdala, has been shown to be involved in fear extinction in animal studies (Davis, Ressler, Rothbaum, & Richardson, 2006) and tested in human patients to augment exposure therapy to positive effect (Mataix-Cols et al., 2017). Using orally administered DCS (50 or 500mg), treatment benefits were found to predominate from pre- to post-treatment (d = 0.25), rather than pre- to follow-up (d = 0.19), and researchers have hypothesized that individual differences may explain some of the benefits, which will require further trials (Weisman & Rodebaugh, 2018). Another cognitive-enhancer includes glucocorticoids (cortisol in humans), which research suggests functions by suppressing amygdala activity toward phobic stimuli to similar levels as compared to healthy controls, possibly by reducing aberrant inputs to the amygdala from non-visual cortex path- ways (e.g., fusiform gyrus; Nakataki et al., 2017). Its efficacy at improving exposure therapy has been explored for fear of heights, finding that in those administered 20mg cortisol orally 1 h before treatment (as compared to placebo), treatment effects were significantly better at post-treatment and over follow-up (de Quervain et al., 2011). Participants also had lower in-session anxiety while performing exposures. A similar methodology was employed for those treated for fear of spiders, finding a similar beneficial result from cortisol treatment (Soravia et al., 2014).

Another strategy may rely on combining exposure therapies with physical activity in order to help patients benefit from the neural plasticity, learning and positive memory effects of exercise (Powers, Medina, et al., 2015). This could be a particularly low-cost and easily disseminatable way to augment exposure therapy given its accessibility to many individuals (Powers, As- mundson, et al., 2015). However, the optimal timing, type and intensity of exercise is as yet unknown. A recent study identified no improvement from pre-training during extinction learning in four studies of rats (in which exercise was manipulated), or a subsequent study in which humans were prescribed either 30 min of rest or 30 min of exercise prior to an exposure treatment for fear of heights (Jacquart et al., 2017). The lack of benefit may be due to the superior benefits of memory consolidation when exercise is prescribed

35 after rather than before treatment, as some have suggested (Tanner et al., 2018). Further clinical work in humans in regards to length, intensity and type of exercise will be necessary to confirm this (Weisman & Rodenbaugh, 2018).

The importance of memory consolidation effects from sleep have been explored in conjunction with exposure therapy. Kleim et al., (2014) conducted a spider phobia exposure therapy treatment in which half of patients were randomized to follow treatment with sleep and the other half to watch a neutral video (the control group). Following the single short exposure treatment of 45 min duration, participants were given 90 min to sleep but on average slept approximately 50 min. Subjects were moved to the wake condition if they failed to sleep (n = 2) and vice-versa (n = 1). Whereas, there was similar effects at post-treatment in both groups on a range of variables such as behavioral avoidance, and self-reported anxiety, at 1-week follow-up there were stronger improvements in self-reported anxiety and reduction of catastrophic fears in the sleep group compared to the control group.

Another novel theoretical construct relies on timing presentation of exposure therapy stimuli in accordance with cardiovascular systole and diastole events (cardiac timing; Watson et al., 2019). The suggestion is that the brief afferent signals that occur during systole and diastole, and are based on baro- receptors transmitting stretch and pressure information from the arteries, also provide cues to the brain regarding emotional states. Previous research on cardiac systole afferent information suggests that these signals may enhance the processing and detection of phobic related fear and threat information. This theory was tested by high spider fearful participants randomized to groups in which exposure to computer images of spiders were timed to coincide with ventricular systole or diastole event as well as a random display group. At the end of treatment, participants in the systolic group had significantly lower fear of spider self-report as compared to those in the random or diastolic groups, and these effects were mediated by interoceptive accuracy (ability to detect one’s own heartbeat) on the part of participants. Those in whom exposure was timed with heartbeat (irrespective of condition) also had improved behavioral avoidance as compared to the random group. These results should still be considered preliminary, given a relatively small sample size, but nevertheless may provide a novel means by which exposure therapy could be enhanced in the future.

Exposure therapies

Today, two broad classes of exposure therapy are available for the treatment of anxiety disorders, one involves direct exposure to the aversive stimuli by visual, touch or other sensory interaction (direct exposure), and the other

involves imagining the object or situation (imaginal exposure). A third-class called interoceptive exposure involves reproducing internal physical sensations such as choking in the patient (Boettcher et al., 2016), but this tends to be most useful for patients who fear the physical consequences of their somatic sensations (Boettcher & Barlow, 2019). Given the complexity of recreating traumatic memories, imaginal exposure has typically been conducted in those with post-traumatic stress disorder but also specific phobias toward more involved situations or stimuli such as phobia of accidents and choking (De Jongh et al., 1999). Variations on imaginal exposure such as those pioneered by Wolpe (i.e., systematic desensitization) include an appetitive reinforcer (training in progressive muscle relaxation) in order to suppress anxiety symptoms (Choy et al., 2007), but this tends not to be used in modern treatments (Wolitzky-Taylor et al., 2008). Another variation called Eye Movement Desensitization and Reprocessing (EMDR) includes imaginal exposure plus therapist directed lateral eye movements toward an external stimulus such as therapist’s hand (Wolitzky-Taylor et al., 2008), however evidence although promising requires more high-quality studies to prove its efficacy (Cuijpers et al., 2020).

Direct exposure may involve not only exposure to the real-life stimuli (in- vivo exposure) but also exposure to digital or other recreations of the stimuli (video, photos or virtual reality) including observing others being treated (i.e., vicarious exposure). Direct exposures are typically conducted by gradually, systematically and continuously exposing a patient to increasing intensities of the phobic object or situation. In the treatment of BII phobias, applied tension can be useful to incorporate to reduce the risk of fainting (Choy et al., 2007). Like systematic desensitization that relies on imaginal exposures and includes an appetitive reinforcer, applied relaxation is sometimes used in conjunction with in-vivo exposures (Wolitsky-Taylor et al., 2008). Although the efficacy of both direct and imaginal exposure have been explored in the treatment of anxiety disorders, evidence suggests the improved efficacy of direct exposure for specific phobias (at least in the short-term; Wolitzky-Taylor et al., 2008) and although there is good evidence for imaginal exposures in symptom reduction, there is mixed evidence in regards to avoidance behavior (Choy et al. 2007). Direct methods of exposure have therefore tended to be the treatment of choice. They’re also the method that benefits the most from inhibitory learning-based exposure strategies such as maximizing expectancy violations and increasing the variability of exposures, advances in technologies to facilitate exposures, and decades of effort to create manualized intervention that can be conducted in relatively few (or even one session) and with high repro- ducibility and efficacy.

37 Dissemination problems of exposure therapies

Despite the efficacy of direct exposure treatments and their continued refinement there have been problems with its dissemination. Lars-Göran Öst in discussing the development of a new exposure therapy treatment for specific phobia noted that it took 4 years for the first 20 patients to seek help for their disorder (Davis et al., 2012). For the most part data on the treatment seeking behavior and refusal of phobic patients is limited, however; reports from large epidemiological surveys in the United States suggest that only 8% of those with a specific phobia sought treatment specifically for their disorder (Stinson et al., 2007). Choy et al. (2007) reported a refusal of 13.6% for those offered treatment in one study (Emmelkamp et al., 2002) and Garcia-Palacios et al. (2001) reported that 27% of students with high fear of spiders if offered in- vivo treatment would be unwilling to receive it. Despite limited data, there is nevertheless good reason to believe that patients are highly resistant to seeking and receiving exposure therapy. The symptoms related to specific phobias, unlike many anxiety disorders, can be effectively countered by avoidance behaviors, and in certain climates (such as Sweden) phobic stimuli such as spiders are seasonal and therefore most problematic only during certain limited times of the year.

There is evidence also of therapist factors in the limited accessibility of these evidence-based treatments. Practical challenges can include maintaining a ready supply of diverse stimuli, particularly live animals and insects. Treat- ments may need to be conducted for longer than the traditional psychotherapy session length of 45 min (often 2-3 h) and insurance companies many not be willing to pay (Reuterskiöld & Öst, 2012). Conducting exposure treatments in real-world naturalistic settings can be beneficial for clients but complicated in regards to therapist insurance for working outside the clinic, maintaining client confidentiality, and risking boundary crossings in the therapist-client relationship (Olatunji, Deacon, & Abramowitz, 2009). A larger and more complicated issue to address is that therapists themselves may have negative views about providing the treatment (Olatunji et al., 2009). This can include concerns that it evokes fear rather than soothes it, with potential physical or emotional risk to the client, or at the very least that it may exacerbate symptoms. Exposure treatments are frequently accompanied by high levels of physiological and emotional response in the patient, and some therapists may fear vicarious traumatization, patient dropouts, or worse malpractice litigation. Even when exposure therapy is conducted it may be done incompetently, without sufficient intensity levels or duration to achieve habituation, or therapists may continue to allow ready use of safety behaviors.

For these reasons, exposure therapy is underutilized by therapists in the US and Europe, used about as frequently as dream analysis and art therapy, and by only about 19 to 33% of CBT practitioners (as summarized in Reid et al., 2017). There is evidence, however, that therapist didactic training can reduce therapist negative perceptions. A full-day training of community therapists on the theory and practice of exposure therapy, for example, included a 50% reduction in therapist reservations about conducting the treatment (Deacon et al., 2013). Olatunji et al., (2009) suggests that there is little data to support many therapist concerns, and provides multiple recommendations on how ethical risk can be minimized. Alternatively, if treatments for specific phobia were available that were as effective as exposure therapy but did not require direct contact with the phobic object or situation, than perhaps these treatments could be offered in their place, unfortunately the evidence for good alternatives appear to be lacking.

Alternatives to exposure therapies

There is some evidence that cognitive-only approaches to the treatment of specific phobias may be efficacious, even as compared to in-vivo exposure therapy (Choy et al., 2007). The evidence is mixed, however, over follow-up. Cognitive-only treatments are typically conducted with cognitive restructuring, using logical empiricism, for example, to reduce the overestimation of threat in phobic patients. This may be helpful in the short-term, however according to inhibitory learning theory, minimization of expectancy violation may actually be deleterious to treatment goals as discussed earlier (Weisman and Rodenbaugh, 2018). There is also the possibility that the benefits of cognitive-only treatments are not as a result of the treatment itself. One study by Szymanski and O’Donohue (1995b) found no significant difference between cognitive-restructuring and a placebo treatment (i.e., lecture on facts about spiders), with one meta-analysis suggesting the benefits of cognitive-only treatments may be non-specific factors alone (Wolitzky-Taylor et al., 2008).

One study treating patients with either supportive therapy (plus tricyclic antidepressant) or systematic desensitization (with or without tricyclics) over 26 weeks with a mix of specific phobia and other phobias (e.g., panic attack), identified no benefit of treatment type (Lipsitz et al., 1999). According to authors, results indicated good efficacy for supportive therapy and not because outcomes were poor using the behavioral approach. Over long-term follow- up (10-16 years post-treatment), however; nearly half of all patients reported clinically significant symptoms in the intervening period (45%). Although sample size was limited with just over one-third reporting (35%), there was a small non-significant advantage for the behavioral group (p = .12).

39 It should be noted that the efficacy of placebo treatments such as “pulsed audio/photic stimulation” and the use of subliminal messages intended to pro- voke fear, are moderate as compared to wait-list control, at least at post-treatment (d = 0.57, p < .01; Wolitzky-Taylor et al., 2008). This indicates the minimum efficacy which should be obtained in an alternative active treatment. In general, the efficacy of non-exposure treatments is relatively high, with large effect sizes overall (d = 0.98, p < .001) in six studies comparing non-exposure treatments to a wait-list condition, but as compared to exposure treatments across 10 studies, non-exposure treatments did not lead to as great improvement (Wolitzky-Taylor et al., 2008). At post-treatment exposure treatments had a moderate advantage over non-exposure treatments (d = 0.44) and at follow-up a slightly smaller advantage (d = 0.35).

Pharmacological agents are commonly prescribed as a first-line treatment for a range of anxiety disorders (e.g., antidepressants; Craske & Stein, 2016), but have shown limited efficacy in the treatment of specific phobias. Benzo- diazepines have shown utility as a short-term acute treatment, such as for a dental appointments or on a flight, but fear returns if the drug is not used on subsequent occasions (Choy et al., 2007), and there are a range of negative health effects from long-term use of the drug making its prescription problematic (Craske & Stein, 2016). There is evidence that nitrous oxide can be useful as a general anesthesia during dental appointments but it is not as effective when compared to behavioral treatments (Choy et al. 2007).

Efforts to improve exposure therapies

Given the advantage of exposure therapy over alternatives, effort has been made to develop exposure-based interventions that don’t require as intensive or as lengthy therapist contact (or even none at all) and can therefore be more cost-effective and easily disseminated. In the 1970s exposure therapies were conducted by behavior therapists for specific phobias like treatments devised by psychodynamic therapists, typically of 1 h duration over a period of 8 or more weeks (Davis et al., 2012). This began to change, however with Öst’s development of One-Session Treatment (OST). As the name suggests, this treatment uses a single massed session of 2 to 3 h duration, and incorporates graded, prolonged in-vivo exposure, participant modelling, reinforced practice, and a cognitive-element consisting of behavioral experiments to focus the treatment tasks and eliminate catastrophic cognitions. The first clinical trial was published in 1989, consisted of 20 patients treated for an average of 2.1 h, and with excellent results (Öst, 1989). At the follow-up period approximately four years after treatment, still 90% of patients were much improved or completely recovered. A review of the literature for this short-term treatment of specific phobias published just less than 20 years later identified 21

separate studies making direct reference to Öst’s treatment procedures (Zlomke & Davis, 2008). The treatment was found to be efficacious for a range of specific phobia types in adults, such as fear of spiders, injectables, flying, small animals and claustrophobia, with an average effect sizes reduction in self-reported fear of Cohen’s d = 1.98 (in the range of 0.91 to 3.05).

Use of treatment manuals to conduct self-exposures were then attempted to reduce therapist time and involvement in treating specific phobias and thereby improve dissemination (Öst et al., 1991). Manualized self-exposures had been used effectively for assisting patients to conduct exposure therapy in the past but mostly for agoraphobia. In testing this approach, Öst et al., (1991) participants with spider phobia were provided 30-page manuals composed of treatment procedures (similar to therapist-directed OST) and instructed to schedule at least two 2 h sessions during a 2-week period in which self-exposures would occur. Patients were asked to call on the assistance of friends or family (or the clinician if needed) in order to gather a number of spiders in glass jars for their treatment. A follow-up to this study also with spider phobic patients (Hell- ström & Öst, 1995), used either the spider phobia specific-manual used in the earlier study or a more general manual (not spider specific) of 23-pages including questions for patients and descriptions for how to conduct self-exposures (Marks, 1978). Self-exposures were performed either at home, similar to the previous study, or in a clinic but without the help of a therapist. In both studies, the manual-based treatments were compared to a single session of therapist facilitated OST.

In the first study at post-treatment, evidence for the efficacy of OST was notable, with clinically significant change found in 71% of those completing the therapist-led treatment. In comparison, only 6% in the self-exposure group met the same criteria. It should be noted that stringent cut-off scores were required on three criteria to qualify for clinically significant change, including self-reported fear, a clinician rating of patient anxiety levels, and completing full scores on the behavioral approach test (12-points) concluding with a spider crawling on a patient’s hand for 20sec. In the second study, using the same criteria, the authors found that the therapist-delivered treatment was superior to self-administered treatments, as expected, but the clinic-based manual condition using a spider-specific handbook was also quite effective. At post and follow-up, 80% of participants met clinically significant criteria in the former group and 36% at post in the later, but rising to 63% after 1-year follow-up. Only between 9 and 10% of those conducting one of two self-administered treatments at home reached clinically significant improvement at follow-up, however. One of the challenges with self-exposures, as identified in the first study, was that despite organized arrangements with the therapist, participants delayed conducting the self-exposures as long as possible and may therefore have not have absorbed their training sufficiently. The use of a clinic-based

41 self-exposure appeared to improve treatment outcomes, at least in the case of the spider-specific handbook, but the general handbook even in the clinic setting returned only 10% of participants meeting clinically significant change.

Group treatment formats were next attempted as a method to reduce therapist time, with the first study working with spider phobic patients during a single session in either small groups of 3 to 4 patients or large groups of 7 to 8 patients (Öst, 1996). Participants worked with the assistance of a single therapist by directly interacting with four spiders of increasing size. Prior group treatments for specific phobia had been completed previously but primarily with dental phobic patients and to a lesser extent acrophobic and fear of flying patients. Results according to clinically significant change identified positive outcomes for both the small and large groups at post-treatment (82 and 70%, respectively) as well as over 1-year follow-up (95 and 75%, respectively). A second group-based study (Öst, Ferebee, & Furmark, 1997) made an effort to reduce therapist involvement even further by using only large groups of 8 participants and randomizing them to either a direct interaction group (similar to the previous study), an observation group where one individual receives treatment while the others watch (but don’t interact with spiders), or watching a videotape of a direct observation session in a clinic without the presence of a therapist (but who returned following the treatment to answer questions). The videotape group was referred to as an indirect observation group. Direct group treatment was the method of choice with as many as 75% of participants at post-treatment in that condition reaching clinically significant improvement. Lesser benefits were found for both observation groups, with 7% direct observation and 31% indirect observation participants meeting clinically significant change at post-treatment and 14 and 44% at follow-up over one year, respectively. The authors noted a problem reported by participants in the direct observation group in that physiological and emotional activation occurred quite strongly during the session, however the participant did not have an opportunity to habituate before the session moved onto a more challenging stage of treatment. This is made it difficult for the participant to pay attention and learn from what was occurring during the treatment

A further study used a stepped-care model to move participants to more psychologist demanding treatments in the event that they did not meet clinically significant change in an earlier therapy (Öst et al., 1998). Participants were initially given a self-help manual specific to spider phobia for use at home, followed by video tape of 120 min duration for watching at home, a group-treatment format with 8 patients, or individual one-on-one treatment with a therapist. A total of 27% of participants met clinically significant change conditions following the first treatment type, followed by 10% in the video condition, 68% in the group treatment, and 100% in the individual treatment condition. Lower clinically significant change percentages were found

in the video condition in this study compared to the previous study, however earlier participants sat in a clinic with 8 other patients, whereas in this study patients viewed the videos at home. The authors report that the videotape condition could also be improved by using more professionally developed videotapes, combining them with better pedagogical instruction, and the videotape condition could also best be combined with a self-exposure manual. It is worth noting that an initially large number of patients dropped out of the manual treatment phase of this study (n = 38) with the most frequently cited rationale being their need for a therapist during the treatment, or their inability to find spiders to conduct exposures with.

The Öst et al. (1997) and Öst et al., (1998) studies incorporating videotapes were some of the first uses of a technological solution for lowering the cost and improving access to evidence-based exposure therapy treatments. Since then multiple other technological solutions have been evaluated: from using computers to conduct internet-based cognitive behavioral therapy (ICBT), incorporating text-modules and online videos of modelled exposure sessions (Andersson et al., 2009a), to viewing short clips of action hero Spiderman videos (Hoffman, Pitcho-Prelorentzos, Ring, & Ben-Ezra, 2019). One of the most promising technological treatments, however uses virtual reality (VR) to conduct exposure therapy by simulating three-dimensional environments, stimuli and even virtual embodied therapists, and enabling patients to interact with the stimuli without having to face the real-life object or situation (Free- man et al., 2017). Given the recent development of modern consumer-available VR and easily accessible downloadable software, the use of VR may become as common as playing a videotape was in the 1990s, enabling widespread access to direct exposure therapy treatments.

In summary, exposure therapy using systematic and repeated exposure to anxiety provoking stimuli and situations is a powerful means of reducing fear, and research continues to better understand its underlying mechanism and optimize how it is conducted. Cognitive enhancers such as pharmaceuticals, physical activity, and use of sleep following treatment, have been incorporated into exposure therapies to improve clinical results but more research is required to validate their efficacy. Today two broad classes of exposure therapy exist, those that favor imaginal exposure and those that favor direct exposure. There have been problems with dissemination of exposure therapy approaches, however; and despite the development of highly efficacious one- session treatment (OST) which can be conducted in a single session, effort continues to be made to deliver the benefits of exposure therapy as widely as possible.

43 Virtual Reality

The first surviving image to have ever been made was taken in 1826 or 1827 on a pewter plate covered in tarlike light-sensitive material called bitu- men of Judea (Barger & White, 2000). The image is of a view from the upper window of the inventor Nicéphore Niépce’s workroom overlooking a garden on his estate in Burgundy, France. Less than 200 years later, image making devices are ubiquitous, carried around nearly constantly by the billions of individuals who own a smartphone (Statistica, 2019), and the number of cameras is expected to reach 45 billion by 2022 (LDV capital, 2017). For those who view and take images, it is clear that the pictures are capable of not just documenting events and moments in time, but transmitting emotion and feeling. The development of the capacity to record an image may have been made as history suggests as a result of a few able inventors seeking French govern- ment financial prizes (Société d’Encouragement des Arts et Métiers) for improvement to an early printing method (lithography), however it has become a means by which joy and happiness, fear and terror, and every emotion in between can be transmitted across time and place. So too is this true of the most recent technological advance in imagery. And unlike the rectangular picture produced by cameras now commonplace on everything from smartphone to cinema screens, VR promises to not just allow us to view an image from the outside but step into and interact with the world within it. The full extent of what this technology means for providing psychological treatments and exposure therapy specifically are as yet unknown, but the benefits are increasingly on a sound scientific footing.

Historical overview

VR had a humble beginning in the 1960s, based on ideas developed within the aerospace industry. According to Ivan Sutherland, referred to as the “god- father of virtual reality” (SIGGRAPH, 2018), Bell Helicopters was experi- menting with placing infrared cameras on the bottom of helicopters synced with pilot head movement and intended to transmit visual information during nighttime landings in narrow clearings. During testing however, researchers identified an unexpected property of the system: a ball thrown towards a camera on the roof would cause the viewer sitting in their office to duck. The

viewer experienced themselves at the camera rather than at their desk. Suth- erland and students went on to develop the 1968 Sword of Damocles system at Harvard and University of Utah, which instead of a camera substituted the visual information people were viewing as computer graphics (Sutherland, 1968). This is largely understood to have been the first 3D VR head mounted display (HMD) ever created (Scarfe & Glennerster, 2019). The system consisted of individual 40-degree field of view head-mounted cathode ray tubes for displaying graphics (refreshing at 30 Hz), hung from the ceiling by a tele- scoping mechanical arm (thus the name of the system), shaft encoders and ultrasonic sensors for measuring motion, and the users head fastened securely to the device with straps. Those experiencing VR were able to move with 6- degrees of freedom (i.e., rotational and movement on x, y and z axis) and able to view wireframe objects from any angle in a 6ft x 6ft x 3ft volume of space.

Following development of the 1968 3D HMD, the technology continued to be refined in specialist laboratories and research institutions in the 1970s including a system at the University of North Carolina which enabled force- feedback touch, and in the 1980s the first US Air Force flight simulator called the Visually Coupled Airborne System Simulator that allowed pilots to practice flying missions (Cipresso et al., 2018). It wasn’t until the mid to late 1980s, however, that commercial companies began selling VR hardware and software. These included VPL Research which was the first VR startup to sell an HMD (the EyePhone in 1988). According to Jaron Lanier the founder of the company, who is considered today the “father of virtual reality,” given the limited computer processing power of the time, development of their VR system started first with touch (the DataGlove first sold in 1985). This device enabling users to move their hands within a 2D virtual world, and was followed by research into 3D audio (enabling spatial orientation of sound), and only sometime later vision (Lanier, 2017).

The HMD and input technologies of the era were limited, but still allowed users to visualize complex virtual objects (e.g., inside a hydroelectric dam), manipulate these objects with their hands (e.g., organic molecules), and allow architectural companies to show homes, gaming companies to publish games, and even multiple people to be in VR together and interact. The VR technology of that generation, however; even in the most advanced systems allowed just 12-15 frames per second refresh rates (versus the 60 to 90 expected today), had poor resolution screens, and primitive graphics (Silverman, 1992). As a point of reference, it was in 1993 that the familiar (but now obsolete) original Intel Pentium CPU became available (running at 60-67 Mhz; Computerhis- tory.org, 2019). VR technology was also extremely expensive. Mel Slater, a researcher using VR devices to conduct psychological research since the early 1990s reported that a system of that generation could cost more than $100 000, with half being computer hardware alone (Slater, 2018). Even the VR system

45 purchased for his lab he reports, shortly before the Oculus Rift was released in 2016 cost $35 000, plus an additional $15 000 for the head tracking setup. Since then the cost of modern VR hardware, which initially was around $650 with the Oculus Rift has continued to drop, as of course has the price of the computers needed to run them.

Today, the most recent devices available are fully-contained systems requiring no additional computer to purchase. They cost as little as $200 for systems that track head rotation (3-degrees of tracking; e.g., Oculus Go), with controls provided by a small hand-held device that function like a laser pointer. For more advanced systems (e.g., Oculus Quest) they cost approximately $400, but also track movement in x, y and z directions (6-degrees of freedom), and include hand controllers for grasping and interacting with objects in VR. These are not much different in capability than the recent Oculus Rift systems requiring a physical computer except for reduced graphics capacity. The ability to deliver treatments at home or a community health centers at a reasonable cost, but with technology superior to those only recently available at advanced university facilities has now become possible (Lindner et al., 2017; Slater, 2018).

Theory and technology

To VR pioneers such as Jaron Lanier who coined the term, the ultimate expression of a VR device would be to offer sufficient stimulus generators and actuators for every part of the body that has a sensor, visual displays for the eyes, scent producer for the nose, speaker for the ears and so on, that a person could experience anything (Lanier, 2017). While the senses like touch and sound may be important enhancement to VR simulations, vision is considered critical (Scarfe & Glennerster, 2019). We may listen all day to music or other sounds while wearing noise-cancelling headphones, but we don’t necessarily feel we have been immersed into another world.

The way our human visual system perceives the world created by a 3D HMD is not dissimilar from how it perceives the real world. A single-eye viewing real or computer-generated scenes gathers rich visual cues from both, such as heights, texture gradients, perspectives, and shading (Scarfe and Glen- nester, 2019). Knowledge of the statistical structure of the world, such as as- suming light comes from above reduces uncertainty about what an individual is viewing and provides interpretations of how patterns of shading are related to a 3D structure. Although the world outside is made up of three-dimensions, the human visual system captures information in two-dimensions, and uses other attributes such as binocular disparity (the horizontal separation from each eye) to generate depth from the resulting offset images one eye from the

other. Similarly, the visual display provided by modern headsets provides a two-dimensional rendered scene but with correct perspective for each eye, and is reconstructed by our visual system as a three-dimensional space (Freeman et al., 2017).

Visual information alone is not sufficient to provide the cues necessary to build up a three-dimensional space, however (even when viewing the real world). Motion parallax or the difference in viewpoint cause by movement, is considered a cue more important than binocular disparity for VR applications (Scarfe and Glennerster, 2019). Movement information is collected in a 3D HMD using inertial sensors and compasses on the headset for the first 3-degrees of tracking (pitch, roll and yaw) and magnetic fields, lasers, ultrasonic sensors or cameras at each corner for the second 3-degrees of tracking (the additional x, y and z coordinates). To be useful, this movement information must be updated in real-time (Scarfe & Glennerster, 2019) and some consider accurate tracking even more important to a good experience than the quality of visuals (Lanier, 2017). When the users head moves right, the scene must move left to compensate. Other cues useful for understanding the dimensions of a scene include occlusion (objects hidden by another object) and optical blur (minimization of blur at the fixated part of the scene relative to the edge). The ability to capture movement related information in VR is one of the most important differences between VR and a 3D television, for example. Together these elements give the impression of a world outside of you, stationary and independent (Lanier, 2017).

This feeling of being in another world has been referred to as “presence” (Slater, 2009). Some differentiate the presence of being in the illusion (place illusion; PI) from the experience that events are really happening (plausibility illusion; Psi). Mel Slater, the researcher referred to previously and a professor at the University of Barcelona and University College London, has been publishing research on VR and its relevance to social and clinical psychology for nearly 30 years. He suggests that the first requirement for presence to occur is that sufficient sensorimotor actions are available from the real world in the hardware of the VR system. This is referred to as the immersion of the system. A system that has head-tracking will always offer a superior experience of immersion than a system without head-tracking, as will a system that enables interaction between the user’s hand and the virtual environment, or one that provides a more accurate and richer view of the virtual world (e.g., higher resolution or wider field of view screen) from one that is more impoverished. Unlike presence, which is referred to as qualia and cannot be measured directly, immersion is an objective measure of the system (Slater, 2018).

47 A person experiencing a system with relatively poor immersion may still experience presence but according to Slater (2009) this is a fundamentally different experience than the presence experienced by someone in a high immersion system. That is, in the low immersion system, the inability to conduct an action in VR as one would in real life requires “creative mental processing” to fill in the gaps experienced by the breaking of immersion (Slater, 2009; p. 3556). Possible modes of gap filling could be heightened emotion and arousal, or top-down conceptual information about the nature of the environment and the stimuli which Diemer, Alpers, Peperkorn, Shiban, and Mahlberger (2015) differentiate from bottom-up perceptual-only processes. Also, the virtual worlds responsiveness may play a role in maintaining presence. Jaron Lanier refers to this responsiveness as a sensorimotor loop, which provides expectations of what the next moment of reality will be like:

“the body and the brain are constantly probing and testing reality. Reality is what pushes back. From the brain’s point of view, reality is the expecta- tion of what the next moment will be like… once, the nervous system has been given enough cues to treat the virtual world as the world on which to base expectations, VR can start to feel real, realer than it ought to… the nervous system is holistic, so it chooses one external world at a time to believe in.” (Lanier, 2017, p. 50-52)

While the experience of presence may never to be measured directly (unlike immersion), if an individual does experience that events are really happening and that they are actually in the illusion, than an individual will “respond-as-if-real” (referred to as RAIR; Slater, 2009). This is not so much a “belief” as Slater (2018) suggests in a recent commentary, but the actual degree to which individuals respond realistically to the virtual environment (Slater 2009). As a user approaches a cliff in VR, they don’t forget that they are in a virtual world (it remains an illusion), but their cognitive system is slower to catch up than their reactions. They will respond as if the event was really happening to them.

Early specific phobia treatments

While VR fell out of the public’s attention in the 1990s and was rumored to be “dead,” the technology of that generation continued to be useful for conducting research on a range of topics, from testing car ergonomics and design in simulations (Lanier, 2017), to uses in the medical sciences and as will be discussed shortly psychotherapy (Slater & Sanchez-Vives, 2016). An early book on VR in medical care by Weghorst, Sieburg and Morgan (1996) called Medicine Meets Virtual Reality: Health care in the information age included

chapters on research themes as varied as using VR to simulate surgery on an- eurisms, aid needle biopsies of the breast, diagnose and treat sensorimotor dis- turbances, view 3D ultrasounds and provide interactive orthodontic care.

Applications for psychotherapy in the 1990s and early 2000s, despite the expense of computer systems of that generation, had a range of uses that were worth exploring even then. For one, exposure to VR stimuli unlike the real world can never truly harm an individual, and it was therefore deemed safer and less threatening, as reported in an early systematic review (Glantz et al., 1996). It was suggested, therefore that this could improve the acceptability of treatments as compared to an in-vivo exposure (Garcia-Palacios, Botella, Hoffman, & Fabregat, 2007). Computers are also highly flexible and program- mable, and therefore theoretically adaptable to the needs of almost any patient (Glantz et al., 1996). Unlike imaginal exposures, a therapist could view the same environments that a patient was experiencing. They could also exert high control over the stimulus, adjusting the level of fear inducement (more or less realistic), as well as its range of behavior (Carlin et al., 1997). Researchers during that era discussed creating virtual situations that were more fear inducing than real situations (enabling “overlearning”; Botella et al., 1998). Even during this period of expensive HMD hardware and computers, VR could provide exposure scenarios more inexpensively than conducting them in the real- world, for example for treating fear of flying (Rothbaum, Hodges, Watson, Kessler, & Opdyke, 1996). Instead of having a patient conduct multiple in- vivo flights and paying for their own and quite likely a psychotherapist’s ticket, such exposures could be conducted with many repetitions for no more than the cost of a therapist’s time.

The first randomized controlled treatment for a psychological disorder using VR was conducted in the city of Atlanta, Georgia by Barbara Rothbaum and her group for fear of heights (Rothbaum et al., 1995a). Prior to this, a pilot test was published by the group that included promising initial evidence for the utility of VR with patients (Hodges et al., 1994). The authors reported patient symptoms consistent with actually travelling to a great height such as “the higher I get, the more worried I get,”, “I was terrified,” or “I’ve got to get off this bridge” (Hodges et al., 1994, p.g. 6). The authors also reported that participants avoided looking down, that VR bridges with gaps in the flooring induced more fear than bridges with solid floors, and this is despite the participant never having left a platform just four inches off the ground and the very limited graphics and processing speed of the time. In terms of performance characteristics, the authors reported that the “bottom line” performance goal they had was just ten frames per second refresh rates and this using monoscopic images with no binocular disparity. Other environments included an open elevator on a 49th story hotel, balconies and wooden platforms over can- yons with rivers up to 80m below. Footbridges were visible at lower levels to

49 increase the sense of height. Novel still by today’s standards, the virtual environments included virtual walkways with railings mapped to physical walkways and railings. The use of a tracker was accessible to the patient’s right hand, and followed its movements, appearing as a virtual hand to control ele- vators buttons. These experiences by a few pilot testers appeared to confirm the value of using graded virtual reality exposure therapy (VRET) as a treatment for fear of heights.

A case study of one 19-year old participant’s experience using the VR application was then published by the group alongside preliminary results data (Rothbaum et al., 1995b). This individual met diagnostic criteria for fear of heights, and in addition to self-reported fear measures, the study included a behavioral approach test in a glass elevator rising within a 147-metre inner hotel atrium at the Marriot Marquis Convention Hotel (an iconic building in Atlanta). The individual received multiple 35 - 45 min sessions (twice per week) for a total of 5 sessions. In the sessions, it was possible for the therapist running the treatment session to view what the patient was seeing. They were able to encourage the patient to get closer to the edge, let go of railings, and perform many of the other clinical skills common to an in-vivo exposure treatment. Raw scores on an acrophobia questionnaire indicated that self-reported anxiety and avoidance reduced from 66 to 36 from pre to post, and subjective units of distress (SUDS) reduced from 65 to 10 when at the 40th floor in a glass in-vivo elevator. The authors noted that there may have been some confounding on the initial pre-treatment behavioral avoidance task (BAT), however; given an express elevator was used initially, rising more quickly and to a greater height than they would have preferred.

That same year the American Journal of Psychiatry published the first VRET randomized-controlled trial (RCT), totaling 20 participants (Rothbaum et al., 1995a). In this trial, 12 participants who were students (not treatment seeking patients) were treated for a total of 7 weekly sessions on each of the graded virtual exposures, and another 8 participants completed a wait-list condition. The authors reported significant decreases in the VRET group from pre- to post-treatment, but not in the wait-list group. Data was also provided from another study which conducted a fear of heights treatments using systematic desensitization (Cohen, 1977). The studies had comparable self-report and avoidance data at baseline, and yet scores improved by 37.3 and 13.3 in the VR study, respectively, and by just 26.6 and 6.7 in the older exposure therapy study. These initial studies provided further indications of the potential of VRET for treating anxiety disorders and particularly specific phobias.

Following this series of ground-breaking studies on fear of heights, the field of VRET treatments began to proliferate, with a fear of flying case report

(Rothbaum et al., 1996), and a controlled agoraphobia treatment with 60 participants (North et al., 1995). The Rothbaum study included exposure to audio and visual material consistent with flying in an airplane, such as sitting on a plane (engines off), takeoffs (rough and smooth), landings (as well as missed landings), and weather such as rainstorms, thunder and calm weather (as well as turbulence). The North treatment included balconies of different heights, an empty room (no windows), a dark barn (including a black cat), covered bridges, rope bridges (similar to the previous canyon bridges), and sitting in a floating balloon. In both studies, self-report symptoms were reported to be significantly reduced following treatment, and in the case of the fear of flying study by Rothbaum and colleagues, the authors reported a complete cross- country roundtrip flight by the lone participant with minimal anxiety. In addition to work completed in Georgia, in the following years researchers in other parts of the world began conducting psychotherapy research using VR, such as for fear of driving (Andrew Berger in New York), for eating disorders using VR to reduce body dissatisfaction (Giuseppe Riva in Italy; Glantz et al., 1996), and for treating claustrophobia (Cristina Botella in Spain; Botella et al., 1998). These were the first research reports on therapies incorporating VR for the treatment of psychological disorders.

Treatments for spider phobia

Treatments for fear of spiders have been particularly emphasized in VRET research, and after fear of heights, agoraphobia, and fear of flying were one of the earliest treatments performed successfully with the then new technology. The first treatment for spider phobia was published in the form of a case study by Hunter Hoffman and colleagues at the University of Washington (Carlin et al., 1997). In this study, a 37-year old patient with an incapacitating fear of spiders had approached the authors for help after having seen a documentary about the Rothbaum research groups work with fear of heights (see HITLab, n.d.). Treatment was conducted over 12 sessions, once weekly, for a period of three months with the patient. After the second month, the researchers were successful in linking animations in VR to the patient’s hand and a physical furry spider toy, a technology they referred to as tactile augmentation. When the patient reached out to touch the virtual spider, her hand made physical contact with the toy. According to the authors, SUDS scores stayed high for the patient during most sessions. Compared to a group of 280 undergraduate students tested for fear of spiders, just one student had a higher or equal fear of spiders at the start of treatment but at the end of treatment 29% of students (i.e., 80) had an equal or higher fear. Symptoms continued to reduce following treatment, and according to the authors the patient began to conduct in-vivo exposures that were self-initiated, indicating a generalization of VR exposures to the real-world.

The first controlled study was then performed using 23 individuals meeting diagnostic criteria randomized to treatment or a wait-list control (Garcia-Pa- lacios et al., 2002). This trial and all future RCT using VRET for treating spider phobia are presented below (see Table 1). In addition to self-reported fear, outcome measures included a behavioral avoidance task using a tarantula. Over an average of four (1 h) sessions, participants were exposed to virtual spider stimuli, first by navigating towards the virtual spider, then touching the spider with their virtual hand, and in the last sessions touching the virtual spider using tactile augmentation. Treatment was considered complete only after the tactile augmentation session while holding a virtual spider with little to no anxiety. According to the authors, 83% percent of patients treated achieved clinically significant outcomes (plus or minus 2 SD) on all main outcomes including improvements on the BAT, self-reported fear, and an 8-point clinician rated phobic severity scale. The wait-list control condition was un- changed at post-treatment and no patients dropped out. One of the strengths of this study (similar to Rothbaum et al. 1995a) was the use of a behavioral approach test, however, patients were not followed up after treatment (post- assessment only) and a relatively small sample size was used.

Overall outcome Overall 2 1 > 3 2 > 1 > 3 2 > 1 = 2 1 < 2 1 > FSQ ES FSQ 2.19 0.21 2.74 3.98 0.64 2.91 4.07 NA 0.71^ 0.74^ 2.07 0.97 BAT ES BAT 1.75 0.17 2.25 0.56 0.42 2.03 3.39 NA 0.94 1.17 1.42 0.98 Sessions 4 sessions x 1hr 3 sessions x 1 hr 8 x hrs 1.5 4 sessions x 1 hr 1 x 30 mins Treatment 1. VRET 2. Wait-list Tactile 1. VRET Non-tactile 2. VRET 3. Wait-list 1. VRET therapy exposure 2. In-vivo 3. Wait-list 1. VRET therapy exposure 2. In-vivo 1. VRET 2. Psychoeducation N ± 23 36* 36 31 77 participants were not evaluated for meeting diagnostic criteria diagnostic meeting for evaluated wereparticipants not

) 2018 ( .

Table 1. Overview and efficacy at post-assessment of randomized-controlled trials using VRET to treat fear of spiders of fear treat to VRET trials using randomized-controlled of post-assessment at Table efficacy 1. Overview and Study et al. Garcia-Palacios USA) (2002; et al. Hoffman USA) (2003; et al. Michaliszyn Canada) (2010; & Bouchard St-Jacques, Canada) (2010; et al. Minns USA) (2018; Available. Not = NA d; Cohen's Size Effct = ES Spider Fear Questionnaire; of = FSQ stimuli); as the a tarantula (all used Test studies Approach Behavioral = BAT NOTE. criteria are data presented meeting not n=28 criteria spider only for diagnostic phobia, met al.participants 36 (2003), et 8 of Hoffman In * al et Minns ± In was also (SPQ-Children) different self-report measure the and adults, using studies other the unlike 8-15 years, aged children with was conducted study This ^

Given the perceived increases in SUDS as a result of tactile augmentation, enabling participants in the VR condition to feel physical texture and force feedback when interacting with a virtual spider, a new study tested out the hypothesis that this would offer increases in treatment efficacy relative to VR with no tactile cues (Hoffman, Garcia-Palacios, Carlin, Furness, & Botella- Arbona, 2003). A total of 28 non-clinically spider phobic students (plus an additional 8 meeting criteria) were randomized to either of these two conditions as well as a wait-list control group, with those in the VR conditions receiving a total of 3 one-hour VRET sessions. Similar to the previous randomized-controlled trial, tactile augmentation only occurred in the last treatment session, with the earlier sessions restricted to getting close to the virtual spider or picking up a virtual bucket with a spider inside using a virtual hand (controlled by a computer mouse-like device). Nevertheless, despite only one-session difference in conditions between the two VRET groups, those in the tactile augmentation group showed greater post-treatment improvement relative to the no tactile augmentation group on behavioral avoidance with a live tarantula (p < .003) and SUDS during the BAT test (p < .005). A similar pattern of results was found in the group meeting diagnostic criteria but were too few to make a statistical comparison. On average, the behavioral avoidance test distance prior to and after treatments with tactile and no tactile augmentation was from 5.5 ft to 6 inches and from 5 ft to 2.5 ft after treatment, respectively, indicating the potential benefits of adding the ability to touch physical objects in VR. Despite the benefits in regard to behavioral avoidance, there was no difference between groups on self-reported fear.

In 2006, one of the first studies attempting to improve the cost of dissemi- nating VRET was published as a within-group efficacy study (Bouchard et al., 2006). The method selected in this study by Stéphane Bouchard, a professor at the Université de Québec en Outaouais (UQO), was not to use lower cost HMD hardware or computer equipment but to reduce the development costs of the VRET application by using existing computer 3D game editors (in this case from Half-life). In addition to being free to distribute, and enabling the use of existing, high quality 3D environments, the authors suggested that due to the ease of modifying these environments they could be done by almost anyone with relatively little computing experience. It was yet uncertain, however, whether they could deliver the same therapeutic benefits as more specialized software. A total of 11 participants meeting diagnostic criteria were included in the study and met over 5 weekly, 90 min sessions. Only the last three sessions, however, included time completing graduated VRET, while the earlier sessions were composed of a diagnostic interview, a familiarization process in the VR environment (without spiders), and traditional CBT psychoeducation and treatment rationale. The last session focused on relapse pre-

vention. During the treatment, patients explored an underground basement virtual environment composed of 6 rooms using a joystick. All rooms contained different combinations of spider types (some striped, with different colors and sizes) as well as room purposes (e.g., kitchens and living rooms). Patients had the opportunity to explore the rooms in any order that they wished, having previously been given a list of what they contained. However, they were instructed to stay in each room until their SUDS levels reduced and the room no longer elicited anxiety. Therapists were with the patients at all times and were trained to help them through exposure tasks and to provide encouragement. Overall, patients experienced significant reductions in self-reported fear on the fear of spider questionnaire (FSQ) and improvements in behavioral avoidance at post-treatment. The avoidance task, however, used a daddy long legs spider (Pholocus) not a tarantula as in the previous studies.

The only study to compare VRET to in-vivo exposure therapy for spider phobia prior to the research conducted in this thesis was also by the Stéphane Bouchard research group (Michaliszyn, Marchand, Bouchard, Martel, & Poirier-Bisson, 2010). In this trial, a total of 36 participants were randomized to either an in-vivo condition, VRET condition, or wait-list control, and were followed for 3 months. A total of 11 were included in the wait-list control group but joined treatment groups after being re-evaluated (8 weeks later). The majority but not all participants met criteria for spider phobia, with 94% (15/16) in the VRET condition and 82% (13/16) in the in-vivo group, but with no significant difference on outcome (or demographic) measures at baseline. A total of 32 completed post-assessment. Treatment consisted of 8 (1.5 h) sessions, with the first session dedicated to psychoeducation and cognitive restructuring, and the last session dedicated to relapse prevention (therefore 6 sessions of exposure therapy). In the VRET condition, participants used an HMD that projected monoscopic (not 3D images), measured rotational head- movement (3 degrees of freedom), and used a virtual environment built with the Max Payne 3D game editor. Participants travelled through the environment using a handheld wireless mouse and the simulation was stopped every 20 to 30 min to prevent nausea from the use of a VR device. In the in-vivo condition, graded exposures to local tegeneria and pholcus spiders were used. In the VRET condition treatment was considered complete after interaction with a large virtual black widow spider (with low levels of anxiety) and in the in-vivo condition after being able to handle both types of live spider in each hand. At post-treatment, both VRET and in-vivo groups were significantly improved according to self-reported fear (p < .001), with no significant difference between them (p = .445). On a behavioral avoidance task (12-steps) using a tarantula spider, a significant reduction over time was also found (p < .011), with no significant difference between groups (p = .12). A significant reduction was found at post-treatment on diagnostic criteria (p < .001), with 1 VRET patient meeting full criteria, another 2 patients meeting partial criteria,

55 and no in-vivo patients meeting criteria. Whereas on self-reported fear over follow-up the in-vivo group continued to improve, the VRET group did not. High-end state functioning required 2 standard deviation improvements on all self-report measures, full remission on diagnostic criteria, and achieving a score of 10 or higher on the BAT. There was found to be no significant difference between treatment groups on this criteria at post-treatment (p = .76) and follow-up (p = .62). The waiting-list group did not improve over both pre- treatment evaluations on self-report questionnaires (p’s < .871).

Recently, a randomized-controlled trial incorporating 3D videos instead of computer-generated stimuli and environments were used to conduct VRET (Minns et al., 2018). The videos used a live Chilean rose hair tarantula filmed with ortho-stereoscopy to create images that are full-sized, 3D and mirror the real world in terms of depth, size and perspective. The authors suggest that this could lower the cost of producing VRET content, as well as being more realistic. In order to test the efficacy of this new method, participants with a fear of spiders (N = 77) scoring 1 standard deviation above the mean on self- reported fear (FSQ) and not able to physically interact with a tarantula during the BAT (scoring no more than 10 out 14), but not tested according to diagnostic criteria, were included in the study and randomized to a psychoeducation-only control group (n = 39) or the VRET group (n = 38). There were no differences on baseline demographic or clinical outcome measures (p’s > .05). In the treatment group, participants watched an initial short-video explaining the treatment rationale, how the treatments were conducted, and this was followed by viewing a 5-min video of spider content a total of 6 times on repeat (30 min total). Over the course of the video, the spider stimuli became increasingly intense, with first a spider being held by a model, then positioned closer to the camera, and then even closer to the camera, and looming towards the viewer. Research personal were not directly involved in providing the treatment but sat next to the participant and instructed them to refrain from safety behaviors and interact with the stimuli as much as possible. The control group viewed the same psychoeducation video as the VRET group, but instead of the 3D spider content, watched a neutral 30-min 2D documentary about music genres. All participants provided clinical outcome measures immediately following treatment, and then the FSQ specifically one week following that. On the measure of self-reported fear, a large significant effect size difference between intervention and control was identified (d = 0.85, 95% CI: 0.384 to 1.32). A medium effect size difference was also found on the BAT (d = 0.47, 95% CI: 0.014 to .92), and VRET participants had significantly lower SUDS scores while completing the behavioral approach measure.

A number of smaller process studies involving spider fearful individuals have been conducted that also include outcome data. In 2005, a within-group efficacy study by the Bouchard group (Côté & Bouchard, 2005) used stimuli

created with a computer 3D game editor (from Max Payne) to conduct exposures. In addition to collecting self-report data and in-vivo behavioral approach data before and after treatment, the authors collected cardiac response (heart rate) and evaluated automatic processing of threatening stimuli (using an emotional stroop task). The treatment portion of the study included 5 VRET sessions (60 min each) involving 28 participants meeting criteria for spider phobia. On the BAT assessment, while no patient was able to achieve all 10 steps at pre-test, following treatment, 46.4% of patients were able to complete the last step on the BAT, a significant improvement over pre-test scores (p < .001). A statistically significant reduction on self-reported fear (FSQ) was also found (p < .001). On the process measures, the authors indicated that heart rate of participants when facing a live tarantula from 173 cm at post-assessment was slower than at pre-assessment, and results on the emotional stroop task indicated that response times for threat interference (time to push color- coded button for threat versus non-threat stimuli) was reduced following treatment.

The Kleim study (Kleim et al., 2014) which was discussed earlier in the context of improving exposure therapy using the memory consolidation effects of sleep also included clinical outcome data. All 40 participants included in the study met criteria for spider phobia and were provided one-session of VRET that lasted for approximately 45 min and followed a structured protocol requiring participants to view various spiders in virtual rooms of an apartment (staying in each location for 60s or until SUDS reduced), after which half were randomized to a sleep condition (for 90 min) or viewing of a neutral documentary video (also for 90 min). Following treatment, significant reductions in self-reported fear were reported in the overall sample, on the Spider Beliefs Questionnaire, and on approach distance (in cm) towards a terrarium with a tarantula, both immediately after treatment and 1-week following treatment. Unfortunately, the measures used in this trial were not comparable to the other trials making comparison difficult.

A study by Shiban et al., (2015) was conducted in order to evaluate the effect of multiple contexts and multiple stimuli on fear reduction in spider phobic patients. This study was based partly on a similar previous study that found significantly reduced return of fear when participants were treated only in multiple contexts as compared to a single context (Shiban et al., 2013). In the follow-up study, total of 58 patients were randomized to one of four VRET conditions that totaled 20 min of exposure: either a single stimuli or multiple stimuli in a single context or multiple context. The different virtual stimuli amounted to one basic spider shape and basement room but with four color and texture variations, respectively. Prior to a behavioral approach test with in-vivo tarantula or household spider, a 20 s return of fear exposure was conducted to a fifth spider in a novel fifth room (both rated by the authors as

57 scariest in previous research). At post-assessment and follow-up, BAT scores using the household spider were significantly less reduced in the single exposure than the multiple stimuli exposure group (p = .04 and p = .022, respectively), however not when using the tarantula. On self-reported fear measures (German version of the FSQ) at follow-up, a significant reduction in fear was noted (p < .001), but no interaction effect for single or multiple stimuli. In the context of clinical treatments, these results indicate that multiple types of stimuli may be beneficial for reducing behavioral avoidance at both post-treatment and follow-up, however interestingly this outcome was only found when the context remained the same (not in the multiple context with multiple stimuli condition).

Modern reviews and process measures

Much progress has been made since the early days of VRET for the treatment of specific phobias. A recent review of exposure-based treatments for anxiety disorders (including PTSD) summarizes the evidence (Maples-Keller, Yasinski, et al., 2017). As of publishing, the review identified 10 RCTs conducted for fear of flying, and an additional 7 trials made up of within-group and other experimental formats. For fear of heights, 4 RCTs have been conducted, 5 within-group comparisons, and 2 additional RCTs incorporating VRET with cognitive enhancers such as DCS. Treatment for storm phobias has included a single subject design study, a claustrophobia study evaluating fear response to the perception of a closing door, and a case study for fear of driving. A separate meta-analysis published in the same year (Maples-Keller, Bunnell, et al., 2017) reviewed the evidence for treating social anxiety disorder (8 total studies), including two RCTs using virtual social interactions, three RCTs for fear of public speaking, and two studies undertaking a preliminary evaluation of school related anxieties. At least one additional RCT, produced by our research for the treatment of fear of public speaking has been published since the Maples-Keller reviews (Lindner et al., 2019).

Indications even since the early days of VRET treatments were that they were effective, however it was not until the 2000s that a series of meta-analyses helped confirm this. In 2008, two meta-analyses were published including only randomized-controlled trials (Parsons & Rizzo, 2008; Powers & Em- melkamp, 2008). The stated goals of these meta-analyses were to provide pooled effect size estimates for treating a range of conditions with VRET (mostly specific phobias; Parsons & Rizzo, 2008), but also to compare VRET to wait-list control and in-vivo treatments (Powers & Emmelkamp, 2008). Meta-analytic evaluation was deemed particularly important because although a total of 300 patients with anxiety disorders were included in the Parsons and Rizzo (2008) meta-analysis, for example, sample sizes were small in the 21

studies included (between 4 and 30 participants). Results of this meta-analysis indicated average random effect sizes were Cohen’s d = 0.97 across all pooled participants for the treatment of anxiety disorders. In the Powers and Em- melkamp (2008) meta-analysis, taking into account 13 studies (n = 397 participants), effect sizes compared to control conditions were Cohen’s d = 1.11 (95% CI: 0.82 to 1.39), and VRET treatments were found to be significantly better than in-vivo treatments (d = 0.35, CI: 0.05 to 0.65), although only 3 studies had directly made this comparison. A later meta-analysis by Opriş et al., (2012) took a slightly different approach than the previous studies, by only comparing evidence-based treatments for anxiety disorders enhanced with a VRET component, against other classic evidence-based treatments (without a VRET component). A total of 23 studies (n = 608 participants) were included. Overall, a large and significant effect size was found against wait-list control (d = 1.12, 95% CI: 0.71 to 1.52). In the studies comparing VRET against classic evidence-based treatments, there was no overall significant difference in effectiveness on primary outcome and behavioral measures at post-treatment or follow-up (p’s > .05). In this analysis there appeared to be a dose-response relationship, such that more sessions resulted in greater improvement (p < .01). Despite these promising results, the authors acknowledge the relatively few studies comparing classic evidenced-based treatments to VRET were available; just 15 studies evaluating self-report at post-treatment, 8 studies behavioral approach, and fewer still over follow-up.

More recently, a series of new meta-analyses have been conducted. A 2019 meta-analysis by (Carl et al., 2019) included 1057 participants and 30 studies comparing VRET for a range of anxiety disorders against in-vivo treatments and controls. Similarly, with this much larger sample size, effect size estimates as compared to wait-list were large (g = 0.90) and there was no significant difference between VRET and in-vivo treatments (g = -0.07). Compared to psychological placebo controls, the effect size for VRET was g = 0.78. For specific phobias particularly, effect sizes compared to wait-list or placebo control were large g = 0.95 (95% CI: 0.63 to 1.28), and not-significantly different as compared to in-vivo exposures (g = -0.08, 95% CI: -0.35 to 0.19). Im- portantly the greater number of participants and studies in this meta-analysis allowed moderator evaluations not possible earlier. Larger sample size studies were associated with smaller effect sizes in VRET versus control condition studies (p < .05), but not with VRET versus in-vivo exposures study comparisons (p = .70). More VRET treatment sessions were not associated with better effect sizes as compared to either controls or in-vivo treatments (p’s > 0.67) and publication year was also not predictive of effect sizes (p’s > 0.20). Fi- nally, Fodor et al., (2018) conducted a meta-analysis for VRET anxiety but also depression studies (n = 39), finding a similar g = 0.70 (95% CI: .57 to 1.02) effect sizes for anxiety as compared to controls at post-treatment and no

59 significant difference with active controls. The authors noted that a predominant number of studies used wait-list controls. The meta-analysis noted a risk of publication bias as a result of potential small study effects, using a number of tests, such as a visual inspection of funnel plots and a significant Egger’s regression intercept test (p = .04).

Other meta-analyses have completed more specialized evaluations of the VRET literature. Morina, Ijntema, Meyerbröker and Emmelkamp (2015) conducted an assessment intended to explore specifically whether treatment using VRET could have effects that generalized to real world circumstances (i.e., behavioral changes related to contact with in-vivo phobic stimuli). The authors therefore selected only specific phobia studies that included in-vivo behavioral approach tests and confined their analysis to studies with patients meeting diagnostic criteria (e.g., unlike Opriş et al., 2012). A total of 14 studies were included and all were either treatments for acrophobia (n = 9) or fear of spiders (n = 5). Uncontrolled (within-group) behavioral approach effect sizes were large both from pre to post (g = 1.23, 95% CI: 1.00 to 1.46) and pre to follow-up (g = 1.63 95% CI: 0.68 to 2.84). In regards to controlled (between-group) evaluations at post-treatment, VRET versus wait-list controlled studies returned large significant effect size improvements (g = 1.41 95% CI: 0.82 to 1.99) and no significant difference as compared to active controls (g = -0.13 95% CI: -0.43 to 0.17). At follow-up, VRET versus active controls also returned no significant difference (g = 0.44 95% CI: -0.12 to 1.00). The effect size results between self-report and behavioral approach outcome measures were reported by the authors to be similar. As a result of non-significant difference between VRET and active treatment results in this and past meta-analyses, researchers have increasingly begun to suggest that VRET might be non- inferior to in-vivo gold-standard exposure therapy treatment, which is a focus of this thesis.

A recent systematic review attempted to answer this question using a met- analysis. Wechsler, Kümpers, and Mühlberger (2019) identified 9 studies (n = 371 participants) using strict criteria to evaluate only those protocols ensuring no differences between amount of exposure between VRET and in-vivo groups, and requiring that both VRET and in-vivo groups were conducted by therapists (i.e., no automated treatments). These treatments were for agoraphobia, specific phobia and social anxiety disorder. Similar to the previous meta-analyses that conducted a VRET versus in-vivo evaluation (e.g., Powers & Emmelkamp, 2008; Carl et al., 2019), this study identified a non-significant difference in effect sizes but this time slightly in favor of in-vivo exposure across anxiety disorders (g = -0.20 95% CI: -0.55 to 0.16) and in a subset of four specific phobia studies (g = -0.15 95% CI: -0.47 to 0.16). Treatments using in-vivo methods for social anxiety disorder were significantly better, however, than VRET treatments (g = -0.50, 95% CI: -0.83 to -0.16). Effect

sizes for VRET treatments had an overall random-effect model g = 1.00 (95% CI: 0.65 to 1.35) effect size and for in-vivo exposure therapy g = 1.07 (95% CI: 0.34 to 1.10). In the case of the specific phobia subset of studies, fixed effect models gave an effect size of g = 0.68 (95% CI: 0.32 to 1.05) for VRET treatments and g = 0.72 (95% CI: 0.34 to 1.10) for in-vivo treatments. It should be noted that although meta-analyses are useful for exploring comparisons between treatments by combining underpowered studies, they are not a substi- tute for a properly powered RCT (in this case non-inferiority trial) because of the heterogeneity of included studies and unknown statistical biases (Mauri & D’Agostino, 2017).

Evaluation of other aspects of VRET (beyond efficacy) have been conducted as meta-analyses recently, including evaluations of attrition in VRET studies and evidence for potential negative effects as a result of the treatments. Benbow and Anderson (2019) undertook an evaluation to understand the question of dropouts (i.e., discontinuation of treatment prior to treatment completion), an important issue to resolve because VRET treatments have in particular been cited as less aversive as compared to traditional in-vivo exposure therapies. The authors included data from 46 VRET studies (n = 1057 participants) to identify overall attrition rates, and a smaller sample of 13 studies (n = 756 participants) to evaluate the likelihood of dropouts from VRET as compared to in-vivo treatments. Overall attrition from VRET studies was a weighted prevalence rate of 16% (95% CI: 12.9 to 19.7) and ranging from 1.7 to 41.4%. The odds of attrition from VRET as compared to in-vivo treatments indicated no significant difference in likelihood of dropping out (odds ratio of 1.09, p = .66). Moderators of attrition were evaluated and only one (homework) was found significant, indicating that those studies that included homework had 17.1% less dropouts as compared to those studies that discouraged between-session homework. While it is positive that VRET studies were not worse than traditional in-vivo treatments in regards to attrition, they were also not better. Nevertheless, it should be remembered that patients were randomly assigned to VRET versus in-vivo exposures and the authors suggest that had patients selected VRET the number of patients dropping out could have been fewer.

In regards to the evaluation of deterioration in VRET studies, this was completed recently using an individual patient data level meta-analysis (Fernán- dez-Álvarez et al., 2019). The study included 15 datasets and a total of 810 patients. In total, deterioration was found to occur in 4% (n = 14) of patients treated with VR, and 15% (n = 27) patients on wait-list. This translated to 4.87 greater odds of deteriorating if assigned to a wait-list condition (95% CI: 0.05 to 0.67). In regards to moderators of negative effects, being married was associated with a significantly lower probability of deterioration as compared to

61 being unmarried in the wait-list condition (OR 0.19, 95% CI: 0.05 to 0.67), but no other sociodemographic moderators were significant.

Overall, VRET offers many important advantages as compared to gold- standard in-vivo and other evidence-based treatments, as summarized earlier, and based on over a decades of meta-analyses similar levels of efficacy, similar levels of dropout as conventional exposure therapy, and low deterioration rates. Criticisms of study quality remain after many years of VRET treatment, however (McCann et al., 2014; Page & Coxon, 2016; Parsons & Rizzo, 2008). Critiques include small sample sizes (Parsons et al., 2008), inadequate information on dropouts (Opris et al. 2012), high dropout rates, lack of long-term follow-ups (Valmaggia et al., 2016) and heterogeneity of outcome measures and reporting of results (Parsons 2015). Few studies have included behavioral avoidance measures to evaluate real-world treatment effects (Powers et al., 2008) and evaluations of assessor and therapist competence have been minimal (Morina et al., 2015). Recent systematic reviews have suggested VRET for fear of spiders is one of the few interventions whose study quality has improved in the past decade (Page & Coxon, 2016). However, previous studies have not been sufficiently powered to determine non-inferiority (Morina et al. 2015). Many studies have shown preliminary support for new treatments rather than conducting methodologically rigorous research on long-standing VRET treatments (McCann et al., 2014). Other limitations of VRET interventions as they currently are, include that that therapists may have reservations about conducting VRET interventions for various reasons as outlined in a recent study by our research group (Lindner, 2019), technical glitches are possible when conducting VRET reducing efficacy, training may be needed for therapists conducting VRET sessions but be unavailable, initial cost of VR is increasingly low for conducting VRET but may still be significant for some, and the quality of VRET software may not always be adequate and would likely benefit from improved design and usability (Maples-Keller, Bunnell, et al., 2017). More mechanistic and dismantling studies are needed and would be beneficial for evaluating for whom VRET treatments benefit and how (Ma- ples-Keller, Yasinski, et al., 2017).

In summary, VR is a technology that was first developed in the 1960s and despite periods of widespread public interest followed by perceived irrele- vance has generally undergone continual refinement since then both to its basic technology and how it is implemented in healthcare and clinical practice. By simulating computer-generated stimuli visually, responsive to movement and sometimes incorporating touch, sound and other senses, VR can present to the individual immersive situations and environments that they might otherwise be too fearful of encountering in real-life and with learning and knowledge that translates to real-world experience. Today VR devices that would previously have cost many thousands of dollars are available for a small

fraction of that and enable exposure therapy simulations to be delivered even to a fully-contained personal device in an individual’s home environment. The use of VRET for spider phobia is a good example of the development and expected efficacy of the approach, with meta-analyses comprising over a thousand patients and treating a range of different phobias, showing little benefit to gold-standard in-vivo exposure therapy, and with similar levels of attrition. Systematic reviews, however, suggest that improvements to the quality of VRET studies are warranted as well as studies that are sufficiently powered to test non-inferiority with the gold-standard treatment.

63 Alliance

The therapist-patient relationship, and the capacity of the doctor to be a healing influence on their patient, their ideal qualities and style of interaction, has ancient roots in Western culture (Horvath, 2018a). Like the first charac- terizations of phobias, these stretch back as far as Hippocrates. The signifi- cance of the relationship, particularly in so far as it relates to psychotherapeu- tic practice, its measurement according to psychometric theory, and relationship to treatment outcome, is however, much more recent. Freud who founded psychodynamic therapy in the early 1900s, and is considered to have been the first modern psychologist to have conducted an in-depth exploration of the relationship as it pertains to psychotherapy, identified having a friendly and sympathetic attitude when interacting with a patient as critically important (Horvath & Luborsky, 1993). Since then, however, variably more or less focus has been placed on this relationship versus the specific techniques and tools used by therapists during treatment (Horvath, 2018a).

In psychodynamic treatments, the development of a healthy personal-emotional relationship between therapist and client is considered the heart of the treatment (Horvath, 2018b), providing a remediating bond to both elicit trans- ference and resolve unconscious conflicts from the patient’s past, the inevitable emotional upsets and strains weathered by means of an alliance (or therapeutic “pact”). Joseph Wolpe founder in the 1950s of Behavioral Therapy, on the other hand, is understood to have placed little emphasis on the therapeutic relationship. Instead, Wolpe suggested that the therapist’s job is to focus on selecting proper strategies, exercises and corrective methods to aid the patient (Horvath, 2018a). Another well-known psychologist and contemporary of Wolpe, Carl Rogers, emphasized strongly the therapeutic relationship in the practice of Client-Centered Therapy and particularly the personal qualities the therapist needed to facilitate the self-actualization of the patient. These were defined as “therapist offered facilitative conditions” such as empathy, unconditional positive regard and trustworthiness. More recently, psychologists such as Lars-Göran Öst, the founder of One-Session Treatment for specific phobia, has stated that a good therapeutic relationship is a necessity but not sufficient for treatment efficacy (Öst, 2012a). That is, the therapeutic relationship forms a foundation by which a patient can have sufficient trust in their

therapist to expose themselves to their most anxiety provoking situations but it is not the specific mechanism of change.

According to Horvath (2018a), due to the relative diversity of treatments that were successful in obtaining positive outcomes in patients, and lack of statistical difference between these various treatments according to meta-analytic studies conducted in the 1970s, an emphasis on discovering “common factors” as a theoretical construct as opposed to “specific factors” (tasks, tools, homework or techniques) unique to a particular treatment was initiated. It is during this period that common factors as a mechanism of treatment elevated the therapeutic relationship from what had been a discussion limited to psychodynamic circles to a broader dialogue in the research community. Between 1960 and 2010 some 31,600 studies have been published on one or more aspects of the therapy relationship (Horvath, 2018a). Specifically, it meant that the term “alliance” now became synonymous with a set of generic components ubiquitous to all treatments in which existed a helping relationship (Horvath, 2011), including traditional psychodynamic treatments but also technology- based cognitive-behavioral treatments conducted online and so on.

Since the 1970s, multiple theories have emerged as frameworks for understanding and measuring alliance, but the most popular is the Working Alliance theory proposed by (Bordin, 1979). In a recent meta-analysis of 306 studies with alliance-outcome data, although 39 unique measures were identified, approximately three-quarters of these used a scale based on the Working Alli- ance Inventory (Flückiger et al., 2018). The theory relies on a three-part model including a personal attachment or bond component, agreement on the goals of treatments, and the tasks used to accomplish those goals. Rather than alliance being a component of treatment along the lines of a set of skills (for example, Roger’s facilitative conditions), or synonymous with the relationship itself, alliance is understood to be a purposeful, joint endeavor and collaboration (Hatcher & Barends, 2006).

The flexibility of the Working Alliance model, often being referred to as being pan-theoretical, has no doubt contributed to its popularity and its use across many psychological schools (both psychodynamic and cognitive behavioral) as well as many helping professions (from medicine to physical therapy). It may also be related to Bordin placing little in the way of restriction on what alliance is not or rather the boundaries of the working alliance (Horvath 2018a). According to Horvath (2018a), the definition that Bordin formulated was narrative and not persuasive, that is it identified the process of alliance, what alliance does in therapy, how it comes about and functions, as well as the importance of alliance in repairing stresses and ruptures, but not a definition of what it is. As a result of the flexibility of Bordin’s theory and it being available to a relatively open interpretation, developers of new treatments

65 have had the opportunity to use and modify the alliance instrument in ways that satisfied the requirements for understanding their particular treatment area.

A good example of the extent to which alliance research has expanded beyond its original use as an assessment of the face-to-face patient-therapist relationship is in the evaluation of internet-based guided self-help (Berger, 2017) using autonomous downloadable mobile self-help apps with no therapist necessary (Berry, Salter, Morris, James, & Bucci, 2018), and VRET where a patient may have a therapist but they are occluded by the 3D HMD (Meyerbröker & Emmelkamp, 2010). The use and evaluation of internet- based treatments has been steadily increasing, totally some 300 controlled studies to date (Andersson et al., 2019) but so far, the evidence for therapeutic alliance in the context of these treatments has been slower to accumulate. A 2012 wide ranging systematic review of the relationship between e-therapies and outcome identified just 11 (1.3%) mostly ICBT treatments that investi- gated alliance (all text, non-video-based treatments; Sucala et al., 2012). A more recent systematic review using stricter search criteria identified 6 internet therapy studies with alliance data from 1658 relevant articles but these were all therapist-guided ICBT for depression and anxiety (Pihlaja et al., 2018). A total 18 studies were identified as part of the Flückinger et al. (2018) meta-analysis, but these also included telephone and videoconferencing-based treatments in addition to internet and email-based interventions. The total number of VR and augmented reality exposure therapy-based treatments including a measure of alliance is 11 (including from those in this thesis), ranging from treatments for fear of flying and acrophobia (Meyerbröker & Em- melkamp, 2010), to small-animal phobia (Wrzesien et al., 2013), adjustment disorder (Miragall et al., 2015), as well as interventions for social anxiety disorder (Anderson et al., 2013; Bouchard et al., 2017; Ngai et al., 2015).

In general, treatments that are conducted outside of the traditional psycho- therapists office, without face-to-face contact with a therapist, suggest a more complex relationship with alliance (Andersson, 2009b). In internet-based cognitive behavioral therapy, guided self-help treatments are conducted at geo- graphic distance, typically using self-help modules with or without asynchro- nous email or real-time texting with a guiding therapist (Berger, 2017). In the case of many mHealth apps, individuals using the app are likely to never have contact with a therapist at all. As a result of a VR headset obscuring the face of a participant during treatment, as mentioned previously, whether or not a therapist is present with a patient in the same room they will not be able to see each other while virtual stimuli or situations are being presented. In automated VRET or mobile self-help apps, the patient is likely never to interact with a therapist during the treatment (even verbally), however like many mental

health treatments, they may be present for initial and final assessments to conduct a diagnostic interview, review a suggested course of action, as well as confirm that treatment was completed successfully.

A narrative review on therapeutic alliance in internet interventions (Berger, 2017), suggests that while the perception may be that face-to-face contact is critical to alliance formation, common therapist behaviors such as warmth, empathy, and a sense of understanding, are also present even in internet-based treatments via text messages to patients. It is these common factors which may enable patients to maintain participation in these treatments without any human contact (Cavanagh & Millings, 2013). Barazzone, Cavanagh and Richards, (2012) used a qualitative approach to review common factors found in the automated features of three guided ICBT treatments for depression. Their research suggests that empathy, genuineness and unconditional acceptance are all engendered by emulating interactions thought conventionally to be tied solely to human relationships. For example, by using generalized empathetic statements, testimonials and fictional characters, the authors suggest that the programs were able to convey warmth and sincerity, as well as providing insights into the experience of having depression. While this research suggests that these programs may be good at establishing an alliance, it remains to be determined whether they will be equally good at further developing and maintaining the alliance, such as through possible ruptures.

In automated VRET that includes not just text-based interaction with a self- help application or contact with a therapist at a distance, but interaction with an embodied virtual therapist, an alliance relationship closer in kind to a human therapist connection may be possible. How alliance develops in relation to an embodied virtual avatar could parallel in some way how immersion and presence in VR occurs. Immersion is an objective measure of the degree to which sensorimotor contingencies are available to users in the virtual world that mimic those in the real-world (such as movement, touch, audio etc.). So too could sufficient capacities of the virtual therapist be provided, such as interactivity, visual presence, and compassionate response, that one could inter- relate with the virtual therapist as if it were a human therapist. Importantly, like presence in VR, which is the psychological experience of actually being there and that events are really happening (Slater, 2009), the distinction between treating the virtual therapist as real or as simply a well-designed program may not be binary. High presence, or in this case treating the virtual therapist as if it were a real therapist, may be possible up to and until a sensorimotor contingency available in the real-world was not possible in the virtual world, and even then “creative mental processing” may compensate for breaks in presence by other psychological mechanisms.

67 Such investigation is really the domain of social presence or co-presence research, defined as the subjective experience of being with a “another” in virtual reality, and one of its main contributions is the extent of immersion necessary to engender the feeling that the other person in the virtual environment is “real” (Oh et al., 2018). Multiple components have been studied that contribute to the experience of co-presence such as the visual representation of the virtual human, its behavioral realism, level of interactivity, whether it was displayed with depth cues, monoscopically, using high- or low-resolution displays, and if the simulation included haptic feedback or high-quality audio or not. According to the authors of the systematic review (Oh et al., 2018), those aspects that contributed positively toward co-presence included interactivity, depth cues, haptic feedback, and audio quality, as well as physical prox- imity and personality traits in the virtual human. Perhaps, surprisingly visual representation and the quality of visual display were a weaker predictor of co- presence. The authors suggest that while photographic and anthropomorphic realism can enhance the experience of social presence it must be matched by consistency with realistic behavior. It should be noted finally that more co- presence may not always be welcomed (Oh et al., 2018). This would be the case for example with individual who was uncomfortable with social commu- nication or was feeling more vulnerable, and therefore the use of less engaging virtual therapists may actually be warranted in some cases to maximize the positive outcome of the intervention study. The visual and behavioral capacities given to the virtual therapist, how this affects the level of co-presence and in turn the alliance with the virtual therapist, will need to be studied more extensively as this research unfolds.

In summary, the patient-therapist relationship has traditionally been considered important to healing, although it has been emphasized somewhat dif- ferently in cognitive-behavioral and psychodynamic theory and practice. A diversity of tools have been used to measure the construct but the most frequently cited are based on the pan-theoretical tripartite Working Alliance framework proposed by Bordin. Given that in technology-based ICBT and VRET the patient and therapist may not be visible to one another, and therefore data on alliance has been slower to accumulate, evidence suggest that therapist relational factors are present even then. In automated treatments there may not be any human therapist involved in the interaction at all, nevertheless positive common factors such as empathy and warmth may still be transmitted through text or the voice and presence of an embodied virtual therapist. Evidence from social presence research suggests how development of relational agents and alliance may continue to advance over time such as by using interactivity, improving audio quality, or emphasizing the personality traits of the virtual therapist.

Aims of the thesis

The overall aim of this thesis has been the study of an automated virtual reality exposure therapy treatment for spider phobia. Using inexpensive hardware and not requiring a human therapist to deliver, such treatments may help ensure that the benefits of exposure therapy can be disseminated widely. A rigorous methodology was designed to evaluate the efficacy and safety of the novel virtual reality treatment as compared to gold-standard in-vivo exposure therapy. Effort was made to provide sufficient justification necessary to allow for future effectiveness trials to be conducted in community settings outside the laboratory, such as in local clinics or a home environment. A process evaluation focusing on the virtual therapist component of the treatment and requiring the development of a new questionnaire was undertaken to better understand how the treatment imparted its therapeutic effects. Finally, quantitative and qualitative baseline data concerning what participants fear about spiders was analyzed to explore how future generations of exposure therapy treatment could be improved.

69 Overview of empirical studies

There are three studies included in this dissertation focused on the use of virtual reality to treat individuals with spider phobia using an automated treatment format. The first study (Study 1) was conducted to evaluate the efficacy of this novel treatment against gold-standard in-vivo exposure therapy (the reference treatment) using a non-inferiority design. The in-vivo behavioral approach test was the primary outcome measure and patients were followed over a 12-month period. The second study (Study 2) evaluated a working alliance tool constructed specifically for use with embodied virtual therapists in an automated treatment format. The instruments factor structure, convergent validity, and psychometric properties were measured and its relationship to treatment outcome assessed. The third study (Study 3) performed an evaluation of the characteristics of spiders that spider fearful individuals find most frightening, their frequency, strength, and association with self-reported fear.

Study I Automated virtual reality exposure therapy for spider phobia vs. in-vivo one-session treatment: A randomized non-inferiority trial

Aims

Evidence from efficacy studies conducted using virtual reality exposure therapy (VRET) since the 1990s has suggested that the treatment may be as effective as in-vivo exposure therapy, the gold standard treatment for specific phobias, while simultaneously being less aversive to patients, reducing treatment refusal and improving treatment seeking behavior (Garcia-Palacios et al., 2007). Until recently however, the technology has been expensive, requiring high-level maintenance and development skills, and as a result was only found at universities and other large institutions limiting accessibility (Slater & Sanchez-Vives, 2016). In 2014, the first modern, low-cost virtual reality headset was produced in large numbers by Oculus Rift as a development kit (DK2; the company later acquired by Facebook) and in that same year Sam- sung began including basic virtual reality functionality in its mobile smartphones (Samsung Gear VR), useable with an inexpensive headset and thereby further reducing the difficulty and price of accessing the technology. Along with improvements to software distribution, like online software stores and downloadable VRET applications, for the first time the opportunity to create a highly-scalable exposure therapy treatment for specific phobia was made possible (Lindner et al., 2017).

The Virtual Reality Immersive Method for Spider (Phobia) Exposure Ther- apy (VIMSE) trial for the treatment of patients with spider phobia was initiated. This trial provided the first direct test of VRET using inexpensive new versions of the technology (Samsung Gear VR) against gold-standard in-vivo exposure therapy. With the goal to improve the dissemination potential of the technology yet further, the treatment was intended to be delivered as an autonomous digital technology (Fairburn & Patel, 2017), requiring no human therapist to conduct the treatment, and instead relying on an embodied virtual therapist and gamification to ensure engagement and treatment adherence (Fleming et al., 2017; Provoost et al., 2017). Unlike previous VRET studies comparing VRET to an in-vivo treatment such as Michaliszyn et al., (2010), the study benefited from a non-inferiority design. Non-inferiority trials are the preferred method of comparing a new treatment to reference treatment when no significant difference is assumed (Piaggio et al., 2012). While it is not statistically possible to prove two treatments are identical, it is possible to say that a new treatment is not worse than a reference treatment by an acceptable margin (Mauri & D’Agostino, 2017). Modern, highly-effective OST (Zlomke

71 & Davis, 2008), a massed single-session in-vivo exposure therapy treatment, was selected as the reference to compare against the novel automated VRET. Unlike Michaliszyn et al. (2010), patients were not transferred from VRET to the in-vivo arm if they failed to respond to the virtual stimuli (3 participants), and for the first time in VRET spider phobia studies, participants were followed for a full year (as compared to just 3 months previously).

This study was intended to determine not just whether VRET could be referred to as a treatment “not inferior” to in-vivo exposure therapy according to a pre-specified margin (as stipulated by non-inferiority designs), but be ef- ficaciously delivered without a human providing the treatment. It was also intended to afford all the reasonable justification necessary for specific phobia VRET treatments to be tested in the future outside the laboratory, in local clinics or even patient’s own homes, with minimal adverse or negative effects.

Method

In this study, participants were recruited from the local community and included in the study if they met diagnostic criteria for spider phobia according to the DSM-5. Patients had to be available to visit Stockholm University on at minimum three occasions and meet other standardized inclusion and exclusion criteria. One exclusion criteria unique to virtual reality was a lack of stereo- scopic vision or balance problems that could interfere with viewing the virtual stimuli. The treatment was fully contained but as this was the first known clinical study to implement an automated VRET treatment, a human therapist was present at all times for ethical reasons and in case of severe emotional response. They were instructed to act as a “computer technician,” with primary responsibility of fitting of the headset and addressing any technical problems with the virtual reality equipment.

The primary outcome measure selected was the behavioral approach test (BAT; Öst, Salkovskis, & Hellström, 1991), composed of 13 approach steps and incorporating a standardized approximately 2-cm wide in-vivo spider na- tive to Sweden and which participants in both the VRET and in-vivo treatment arms would be asked to approach in the presence of a blinded assessor (at pre- , post- and 3- and 12-months after treatment). Similar to past meta-analyses evaluating the efficacy of VRET interventions, this would provide an opportunity to test the generalizability of VRET to real-world conditions (Morina et al., 2015).

This study was conducted as a randomized non-inferiority design. The non- inferiority margin used in the study was based on pre-existing data as the minimum improvement needed to achieve clinically significant change (set at 2

points; Andersson et al., 2009a) and published in a research protocol of the study (Miloff et al., 2016). A sample size of 100 participants was calculated as required, with a conservative standard deviation of 4 (actual SD 2.47 and 2.71), 80% power to detect the non-inferiority margin, and using a 1-sided 95% confidence interval (sealedenvelope.com). Secondary outcome measures included meeting diagnostic criteria for specific phobia according to The Structured Clinical Interview for DSM-IV (SCID) adapted to DSM-5 (First et al., 2002). It also included self-reported fear, using the Fear of Spiders Ques- tionnaire (FSQ; Szymanski & O’Donohue, 1995) and the Spider Phobia Ques- tionnaire (SPQ; Muris & Merckelbach, 1996), the former benefitting from improved time-scale discrimination, and the later included primarily as a comparison with previous published research.

Results and discussion

The primary purpose of this study was to test non-inferiority of a novel automated VRET for spider phobia as compared to traditional in-vivo exposure therapy, given that a series of past VRET and spider phobia specific studies (Michaliszyn et al., 2010) indicated no significant difference between the treatments. Participants with an average age of 34 (SD = 10.35) and mostly female (> 80%) were randomized 50 to each group. This was a relatively highly educated sample with more than half in each group (29/50 and 35/50 in OST and VRET, respectively) having a college or university education. There were no significant difference in the groups at baseline on either demographic variables or outcome measures. All participants met criteria for specific phobia prior to treatment using a structured clinical interview.

Non-inferiority was identified at 3- and 12-months follow-up, but not at post-assessment as was previously hypothesized, and VRET was significantly worse than the in-vivo treatment until 12-months (upper-bound did not cross null). The results on behavioral avoidance continued to narrow over follow- up, however, and this was likely as a result of a ceiling effect. Of those treated with in-vivo exposure therapy, approximately 50% achieved the maximum score (12 points) on the BAT at each of the time periods, as compared to 11 to 15% in the VRET group, and showing no deterioration (p = .376). Never- theless, at 3 and 12-months follow-up the mean difference on the BAT were just 1.34 and .90 on observed data, or 1.21 and .81 on intention-to-treat analysis data, respectively. These scores amounted to the difference between physically touching a spider and holding one’s hand close it. Given that the VRET participants never had the opportunity to practice handling a spider even under VRET conditions (e.g., using tactile feedback; Hoffman et al., 2003), the last two levels on the BAT (11 and 12) which required touching an in-vivo spider are likely to have been quite challenge. It will be interesting to test in future

73 studies whether VRET with haptic feedback might improve translation of virtual practice to behavioral approach outcomes in the physical world.

On self-reported fear (FSQ and SPQ), the secondary outcome measures, large effect size reductions were noted for both VRET and OST at post and follow-out. In the VRET group effect sizes were in the range of 1.2 to 1.5, and in the in-vivo group these results were in the range of 2.0 and 2.7 with significantly better improvement in the in-vivo treated relative to the VRET. Treat- ment results were maintained in the in-vivo group over follow-up on both secondary outcome measures but continued to improve in VRET group on the SPQ measure relative to the OST group. Despite the significantly greater improvement overall in regards to self-reported fear using in-vivo exposure therapy relative to VRET, there was no significant difference in individuals meeting clinically significant change, with 56% and 38% meeting cut-offs at post- treatment, respectively (p’s > .071). This is in contrast to the measurement of behavioral avoidance where only 28% of participants in the VRET group met clinically significant change cut-offs as compared to 72% in the OST group (p < .001).

Finally, it is important to note that there were no significant differences in mean scores on a negative effects questionnaire (p = .145). According to the three most commonly reported symptoms in each group, in-vivo participants reported more anxiety related negative effects (e.g., “I felt more worried”; 13/48 participants), whereas VRET participants tended towards reporting more dissatisfaction with the treatment itself (e.g., “I felt that my expectations for the treatment were not fulfilled”; 19/47 participants). Given the relative similarity between results on many outcome measures in this study, the difference in type of most commonly reported negative effect suggests a stepped care model using automated VRET first may be warranted prior to recom- mending in-vivo treatments. Nevertheless, considering this study is among the first automated exposure therapy treatments currently available, there should be ample room for improvement, and this improvement appears to be justified.

Study 2 Measuring Alliance Toward Embodied Virtual Therapists in the Era of Automated Treatments: The Virtual Therapist Alliance Scale (VTAS)

Aims

In order to deliver a truly automated exposure therapy application at the minimum a participant must be provided a number of explicit pieces of information. These might include psychoeducation underscoring how learning theory works (e.g., habituation), practical advice about how to perform the treatment (e.g., the need for step-by-step increases in intensity), and how to conduct maintenance following exposures (e.g., keeping a record of exposures). These are the specific skills and tasks of exposure treatments, however, starting in the 1970s a new perspective on the benefits of psychotherapy suggested the importance of non-specific factors as well (Horvath, 2018a; Zilcha-Mano et al., 2019). Alliance was the name given to measurement of one aspect of these non-specific factors, referred to as the therapeutic relationship, and working alliance (Bordin, 1979) the most commonly used conceptualization for how to do so, composed of a tripartite model based on agreement on the tasks and goals of treatment, and bond between patient and therapist.

With modern technologies such as VRET, it is possible to not just provide a self-help treatment manual or information video alongside phobia stimuli, but incorporate an embodied virtual therapist able to deliver all the general instructions discussed above but with a human voice form, instructions based on real-time patient interactions with the stimuli, as well as guidance and verbal reinforcement of the patient as treatment progresses. Given that many automated internet-based self-help treatments suffer from low-adherence or poor efficacy, the use of embodied conversational agents has been suggested as a possible means of “bridging the gap” between interventions that include a human therapist as part of the treatment regime, and those treatments that are unguided (Provoost et al., 2017).

Despite a number of recently published automated VRET clinical trials incorporating an embodied virtual therapist (Donker et al., 2019; Freeman et al., 2018; Miloff et al., 2019), to date none has measured working alliance with the avatar, a potentially important predictor of efficacy in their treatment. This may be partly as a result of the lack of purpose-built tools for measuring alliance towards embodied virtual therapists, given that the term virtual therapist is not simply interchangeable with human therapist in questionnaires , and recent research suggests that neither is referring to alliance with the “app” sufficient for measuring the subtleties of the patient therapeutic relationship

75 (Berry et al., 2018). This study aimed to develop a working alliance measure specific to embodied virtual therapists referred to as the Virtual Therapist Al- liance Scale (VTAS), evaluate its psychometric properties and factor structure, and measure its relationship to patient treatment outcome.

Methods

The VTAS scale (17 items) was based primarily on Working Alliance In- ventory items (WAI; Hatcher & Gillaspy, 2006; Horvath & Greenberg, 1989) from across the three WAI subscales (Task, Goal and Bond) deemed most applicable for use with a virtual therapist. Also, included were items used for measuring the level of empathy of a virtual therapist from a previous study (Pontier & Siddiqui, 2008). Finally, co-presence items from a previous scale (Bailenson et al., 2003), which can be referred as the experience of being with another in a virtual environment, were included. All items were re-written specifically for this new scale (in reference to the “virtual therapist”).

Participants used in the evaluation of this scale came from two separate but related studies. The first study was Study 1 referred to above (Miloff et al., 2019) and the second study (Lindner et al., 2020) a replication of Study 1 but conducted without a human therapist being present at all (instead the patient received a set of written instructions) and using a slightly updated VRET software in which the voice of the virtual therapist was English (instead of Swe- dish). A total of 70 participants completed VTAS questionnaires.

Exploratory factor analysis of the VTAS was evaluated using parallel analysis scree plot and factor loadings, extracted using maximum likelihood esti- mation, and an oblimin rotation. Internal consistency used Cronbach α and McDonald’s ω. Convergent validity was assessed using bivariate correlations between the co-presence factor and z-corrected presence scores from two separate scales, and an empathy-bond factor compared with system usability scores. Steiger’s tests compared correlations between each factor with one variable in common (VTAS total score). Multiple linear regression was used to predict VTAS total and factor scores based on all possible predictors at once.

Results and discussion

Data from this analysis suggests that the two VRET studies used to evaluate the VTAS provided an appropriate test of patient alliance with an embodied virtual therapist. Evaluation of demographic and baseline self-reported fear characteristics indicated no significant differences between participants from

both studies (p’s > .432). Scores on the VTAS items followed an approximately normal distribution based on visual and computational evaluation (i.e., Shapiro-Wilk test), and internal consistency on the VTAS was high in both studies (α’s > .915).

Exploratory factor analysis indicated a two-factor solution, with most task, goal and co-presence items in factor 1 (12 items) and most bond and empathy items in factor 2 (5 items). A task-goal and bond two-factor solution is consistent with factor analysis of the WAI in past samples with human therapists (e.g., Falkenström et al., 2015). Factor loadings were all above recent recommended cut-offs (i.e., > 0.35; Clark & Watson, 2019). Internal consistency was high on VTAS total scores (α = .921) and factor 1 (α = .931), and adequate in Factor 2 (α = .710). Although Clark and Watson (2019) have recently rec- ommend cut-offs of α = .80, high coefficient α is difficult to achieve with few items (just 5 items in factor 2).

Correlations between the VTAS (its factors) and self-reported fear indicated no significant correlation between pre- and post-treatment, but between post- and follow-up the VTAS significantly correlated with treatment outcome on both total (r = .310) and factor 1 scores (r = .333). There was no significant correlation still on factor 2 over follow-up. Correlations on the VTAS are consistent with recent evidence (over 30 000 patients) from a meta-analysis of alliance correlations with outcome, in the range of r = .275 to .278 for internet- based and traditionally delivered treatments, respectively. Convergent validity of factors suggested factor 1 correlated significantly more with presence z- scores (p < .001), as would be expected considering all co-presence items were found in that factor, and with only user friendliness predicting factor 2 scores (p < .001), also expected considering use friendliness has certain commonali- ties with bond or empathy such as providing a sense of accessibility and guidance (Barazzone et al., 2012).

Future studies will benefit from evaluating the strength of patient alliance towards virtual therapists with different forms of embodiment (voice-only versus physical), levels of interactivity, and types of personality and relational behavior, and exploring their association with outcomes. The research here was completed with studies using a single treatment session, and it will be worthwhile investigating whether strong alliance with a virtual therapist may help improve adherence in automated treatments as suggested by Provoost et al., (2017). The evaluation of how generalizable these findings are is also important, considering the relatively low mean age of participants (< 35 years) and high educational attainment and literacy skills. This is particularly the case considering that the utility of automated VRET treatments are likely to be higher in low and middle-income countries where access to evidence-based treatments are at their least.

77 Study 3 What is so frightening about spiders? Self-rated and self-disclosed impact of different characteristics and associations with phobia symptoms

Aims

Spider phobia is among the most common specific phobias (Oosterink et al., 2009) and yet except for an older study (Davey, 1991) little is known about what visual characteristics of spiders or other spider related cues are fear provoking for phobic individuals. The identification of these characteristics have a number of possible benefits. In the context of face-to-face in-vivo treatments, if clinicians are aware of the primary cues that patients are fearful of, treatments can be more focused, and behavioral experiments setup ahead of time to precisely explore emotional triggers. This is in keeping with Öst’s (2012) recommendations in his manual for one-session treatment which suggests specific exposures rather than broad non-specific exposures are more rapidly able to shift avoidance behavior, subjective anxiety and promote new learning. Knowledge of what cues are fear inducing to patients may also help ensure that fear hierarchies can be developed that maintain a high-level of arousal for patients throughout the treatment process as recommended to maximize inhibitory learning (Craske et al., 2014).

In the context of VRET treatments, past studies with spider phobic patients have sometimes moved patients to an in-vivo condition if they were not sufficiently aroused by the virtual stimuli (e.g., Michaliszyn et al., 2010). Some of the mostly highly reported “negative effects” in the clinical trial, commonly defined as adverse or unwanted events as a result of treatment (Rozental et al., 2016), included for example that “the treatment did not produce any results”. In the context of treatments conducted with a clinician, if their patient was not experiencing any results from working virtual stimuli, transitioning to the use of an in-vivo stimuli would not be particularly difficult (as long as the clinician was trained adequately). However, given recent advances in automated VRET that could include providing these evidence-based treatments at community health clinics, perhaps with the assistance of a nurse or psychosocial health workers, or as a truly automated downloadable intervention on a personal virtual reality device at-home, failure to be adequately aroused by the treatment could be quite a negative outcome for the patient. Ensuring that VRET is not rejected and patients are not demoralized resulting in excess attrition should be a priority for future treatments, particularly considering dropout rates are similar between in-vivo and VRET treatments and the most commonly reported reason for dropping out was failure to immerse in the VRET environment (Benbow & Anderson, 2019).

Methods

All participants who applied for participation in the VIMSE clinical trial were required to complete a baseline screening inventory for inclusion. In addition to a battery of self-report measures such as the FSQ (Szymanski & O’Donohue, 1995), and standardized demographic and contact questions, participants filled out quantitative ratings of different spider characteristics (7 items), their relative impact on situational fear (5-step rating; from no impact to great impact), as well as a free-text question on spider characteristics (“what is it about spiders that you find frightening”). The quantitative section included questions designed to discriminate between components of “legginess” referred to as important in Davey (1991) as well as other potentially clinically important visual and movement related characteristics of spiders (without being exhaustive). Nearly 200 participants completed this preliminary baseline screening (n = 194).

Analysis of the data included descriptive statistics for each item of the quantitative questionnaire, factor analysis (using parallel analysis) to create sub-scales of characteristics, and normalized numbers from these factors (di- vided by number of items) as loan predictors in single linear regression models and together as covariates in a multiple regression model with self-reported fear of spider self-report (FSQ) as the dependent variable. Qualitative analysis used concatenated Swedish raw data, cleared of punctuation and common stoppage words, spider-related nouns, emotion-related verbs, adverbs and related terms in different tenses manually stemmed. A processed frequency table of words were then translated into English equivalents and visualized using a word cloud.

Results and discussion

Factor analysis of the quantitative results indicated a two-factor solution based on factor loadings and visual inspection of a parallel analysis scree plot. These groupings corresponded to a movement-related and appearance-related factor of 3 and 4 items, respectively. The color of the spider obtained the lowest mean impact scores and how close it is, how large it is and movement related scores (how it moves and how much it moves) the highest score. Free form qualitative results indicating “legs” as the most common term, followed by “fast” or “move,” provide added support to the importance of movement relative characteristics as those most fear provoking. Including both factor scores as predictors in single regression models or together in a multiple re-

79 gression model did not, however, find a significant relationship with self-reported fear (p’s > .736), indicating that patient pre-treatment severity of fear was not associated with a particular spider characteristic.

These results suggest that the dynamic characteristics of spiders, their movements, distance from the individual and size, are the qualities that induce the greatest fear in spider fearful patients and are therefore the primary content that should be included in future VRET applications. When used in the context of in-vivo exposure therapy, having a healthy spider (rather than a weak or tired spider) that is able explore and interact with patients might be critical to achieving adequate arousal necessary for treatment efficacy. It will be important to confirm these results, however, in populations outside Sweden given that harmful spiders in other countries have familiar clear visual mark- ings such as the red spots and/or hourglass shape on widow spiders (Latrodec- tus), that may be particularly arousing to spider fearful individuals.

General Discussion

The overall purpose of this thesis has been the development and evaluation of an automated exposure therapy application for treating specific phobias, in this case spider phobia. In the first study we have shown that over follow-up automated VRET was not inferior to gold-standard in-vivo exposure therapy according to a pre-registered behavioral avoidance margin. In the second study we evaluated using a novel working alliance questionnaire an important component of automated treatments, namely the virtual therapist, which provided patient guidance through the program, psychoeducation and encouragement during exposures. In the third study, we explored what were to spider fearful patients the fear-provoking features of spiders, an aspect of exposure therapy treatments that may be valuable for creating improved future generations of VRET treatments.

The efficacy of automated VRET for spider phobia

The automated VRET used in the research studies included in this thesis and designed by multiple authors on these papers was central to the efficacy of the treatment patients received. It was composed of components based on traditional in-vivo exposure therapy but also components that had not yet to our knowledge been incorporated in such a treatment. Conventional aspects of the treatment included a progression system that used a fear hierarchy beginning with stimuli that were small or likely to be of relatively low arousal (e.g., a cartoon spider) and transitioning to larger, more realistic and fear inducing stimuli. The psychoeducation provided to patients was based on theory and practice developed over generations by researchers and psychologists, relying roughly on Öst’s OST treatment manual (Davis et al., 2012). Patients spent periods of time simply viewing a spider in different areas of a home environment to encourage habituation. Sometimes the spider was found walking on a table in a TV room, at other times in the corner of the living room or outside on a patio, but also walking directly towards them. The later “looming” exercise with a spider was a form of completion step before moving onto the next stage of the fear hierarchy and new more challenging stimuli. This

81 exercise would not be particularly different from a patient being able to man- age a spider walking up their arm before moving onto interacting with a next larger type of spider, first just looking at it, and then greater degrees of contact.

The more novel aspects of the treatment included turning the progression system into a game, requiring patients to solve simple puzzles, gain points, and generally interact with spiders in ways not possible in a conventional therapy session. Although, patients remained seated for the duration of the treatment (the Samsung Gear VR device allowed only rotational head tracking), interactivity with the application by means of games may have contributed to a sense of presence with the environment by encouraging interaction. It should be noted however, that we found no significant difference between those who experienced high and low presence and outcomes on behavioral avoidance or self-reported fear (p’s > .014). Further research is warranted, but as has been noted elsewhere (Diemer et al., 2015), it may be that a certain threshold of presence is required after which additional presence, that is feeling like one is in a virtual reality environment and that events are actually occurring, is merely supplementary.

Inhibitory learning methodologies were used, such as having patients interact with multiple varieties of spiders (deepened extinction learning) and in different areas of a suburban (Swedish) row house (i.e., multiple contexts). This would typically be quite difficult to accomplish with in-vivo spiders in a naturalistic setting in the real world. Spiders were also presented in such a way as to encourage patients to help and care for them (positive valence). This was based on evidence from previous clinical trials that those viewing a 7-min story about a kind and wise personified spider (e.g., Charlotte’s Web, a children’s story) were more successful during a later fear reinstatement exercise than those in a control group (i.e., lower self-reported fear and improved behavioral approach; Dour, Brown, & Craske, 2016). In-session VRET exercises often combined positive valence and gamification. For example, at different stages of the treatment, participants helped a spider cross a pond by moving floating lily pads, protected a spider with an umbrella from rain, and helped feed spiders. Patients learned about the life-cycle of spiders via a “spider ex- pert,” a female voice who described spider importance to the ecosystem and other generally positive messages about spiders during the course of the treatment. Finally, the virtual therapist was present with the patient from the first interactions with the VRET application, and was used to create a supportive, warm, empathic presence under circumstances that would otherwise be highly aversive for the individual.

Theoretical implications of VRET efficacy

We are now familiar with and aware of the benefits of exposure therapy and how habituation and inhibitory learning function to reduce avoidance and fear, however there are aspects of VRET that are quite different than traditional in-vivo exposure. For one, patients are well aware that the spiders they are seeing are entirely virtual, and therefore harmless, and unable to physically touch them or be in the same room as them, and yet treatment outcomes are similar and spider phobic patients seem to experience high emotional and physiological mobilization as a result of the VRET treatment. This can be at- tested by anyone present while a patient is treated with VRET or the high SUDS reported in some studies. This relates therefore to an important question regarding the mechanisms by which the treatment effects of VRET are made possible. If understood and evaluated, then future VRET may be able to take advantage of these mechanisms and utilize them for therapeutic effect.

One notable characteristic of virtual reality that the epigraph from Jaron Lanier in the opening pages of this thesis refers to is that in virtual reality there is a clear distinction between one’s inner life and what is happening outside of you. In the normal day-to-day we assume that what is happening inside of us (our cognitions and emotions) are because of what is going on outside. When we get scared of a spider in virtual reality, however; and recognize ob- jectively that what we’re seeing is virtual, this may help us to understand that the fear is not located in the stimuli itself but rather in our orientation towards or beliefs about the stimuli. The discrepancy between internal dialogue (what we think is going on) versus external reality (what is actually happening) is highlighted by Öst in his OST treatment manual. The final goal of OST he suggests is that the patient “…no longer believe in his or her catastrophic cognitions” (Davis et al., 2012, p. 70). Although, this recognition is certainly possible with conventional treatments and even likely, the use of virtual stimuli might increase the speed at which this discrepancy between our thoughts and reality can be noted faster and better than interacting with the in-vivo stimuli itself.

Related to this, VRET may assist us in observing this process by which the interpretive function of the mind “explains” what we’re seeing (i.e., interpretation bias). Evolutionarily speaking, the hominids most successful at inter- preting the waving of grasses at the edge of a camp as a stalking lion, would have been those humans that became our ancestors. As a result, perceptual capacities may not be geared towards deducing objective reality as we expect,

83 what is referred to as veridical perception, but rather re-constructing a reality that is in more keeping with evolutionary pressures to survive. This has been termed the Interface Theory of Perception by Hoffman at the University of California, Irvine (Hoffman, 2018). The author uses the example of a desktop interface on a computer, we see a file folder on the desktop as blue, rectangular and off to one corner, but must be aware that the reality of the file exists simply as lines of code in computer memory, the interface just helps us to get a job done in the real-world. The same may exist for us as we interact with the world around us, learning to recognize the grass moving as merely wind blowing, or a spider as just a common harmless insect wandering along. The interpretative capacity of the mind that sees fear where there is none is a biological “hack” that kept us safe, and using virtual reality to help note this discrepancy between the sensory input and our minds interpretative ability may provide beneficial opportunities for future VRET research.

There are other theoretical aspects of using virtual reality to conduct exposure therapy that while more basic conceptually may still prove useful to explore in the future. Unlike a computer screen or photograph, stimuli in virtual reality are seen in three-dimensions, requiring reconstruction of the visual space using binocular vision, motion parallax and other visual cues. Inhibitory learning theory suggests that the more cognitive (and perhaps) neurological resources that can be harnessed and devoted to an exposure task, the greater the learning that will be possible (Craske et al., 2014). Part of the benefit of this might be the memory related enhancements of attending to and recording a multiplicity of sensory cues (Weisman & Rodebaugh, 2018). Future VRET treatments could make use of more complex computer-generated environments and visual features, spider or otherwise, to increase cognitive and neurological processing.

In virtual reality the visual field is reactive to users behavior, such that a movement of a participant’s head in the VRET simulation used in this study (for example), could allow the participant to knock over a plastic cup that was hiding a spider, or move a pillow to protect a ball from falling on a spider. These actions are referred to as sensorimotor contingencies (if I do this, than that will happen) that mimic behaviors undertaken in the real world, are required for perception, and give rise to the feeling of immersion (Slater, 2009); if the virtual world responds like the real-world than it will be treated as such (Lanier, 2017). Unlike simply looking at a 2D picture, the reaction and coun- ter-reactions that virtual reality requires and stimulates in a user provide a much more engaging experience, likely improved learning, and should be enhanced wherever possible to create greater behavioral change in the real world.

Potential threats to validity of results and limitations

While we can say with reasonable certainty that the VRET treatment that was conducted here was highly efficacious, consistent with multiple prior studies of VRET with spider phobic patients, and that the treatment is non- inferior over follow-up as compared to in-vivo exposure therapy, future replications will be needed to confirm the later result. Replications will be needed in different populations and even countries as would be normal for assessing the efficacy of any new clinical treatment, but also in part because of design decisions in this clinical trial.

Behavioral approach tests using in-vivo spiders were incorporated as a primary outcome measure in order to evaluate the generalizability of treatment results to real-world conditions. This is typically preferable to treatment outcome measures being restricted to self-report or what the patient “thinks” is their current level of fear and avoidance of spiders (rather than demonstrated interaction with an in-vivo spider). However, as a result of BAT measurements before and after treatment (and 3- and 12-months follow-up) patients in both conditions did experience some actual exposure time to real-life spiders. This is of practical consequence, of course, primarily for patients randomized to the VRET condition who would otherwise have not had any contact with an in-vivo spiders during treatment (only virtual spiders). These were short exposures, likely in excess of no more than 2 or 3 min, and did not enable patients to practice any skills related to interacting with the spiders or habituation to the stimuli, however they were able to expose themselves to the full intensity of their emotional response and for that reason there may have been some benefit to treatment outcomes. Unfortunately, there is little research evaluating whether very short exposure to spiders (or other stimuli) has any efficacy in and of itself.

In terms of the overall efficacy of the treatment, patients were instructed following the single session of therapy to engage in treatment maintenance. Maintenance instructions typically revolve around recognizing anxiety as normal, responding to setbacks, and differentiating them from relapses but also how to conduct self-exposures. In order to retain consistency between VRET and in-vivo exposure therapy, and provide optimal benefits of treatment to both (in keeping with ethical considerations), participants were provided these recommendations as well as instructions and worksheets on how to do it. This is standard procedure in OST following a treatment and is conceived of as helping to further develop the skills learned by the patient with the therapist (like a driver practicing new driving skills; Davis et al., 2012), however some benefit of treatment over follow-up may be the result of self-exposure rather than the treatment itself. This has been hypothesized as a reason for continued

85 improvement over follow-up in specific phobia studies (e.g., Wolitzky-Taylor et al., 2008) but so far there has been a lack of data collected on the subject. We anticipated being able to collect self-exposure data but unfortunately due to logistical reasons were not able to do so consistently. Nevertheless, even if the primary mechanism of change in VRET studies were that it led to self- exposures the credibility of treatment results should not be discounted.

Another important limitation of the clinical trial was that a human therapist was present with patients throughout the duration of the VRET treatment. This was done primarily out of ethical concern for the patient. At the time of the planning and development stages of the clinical trial, to our knowledge no other exposure treatment had ever been carried out in an automated treatment format before. Self-help exposure treatments had been conducted in the past, via pictures and video presented to patients online (Andersson et al., 2009a), or a videotape being mailed to patients (Öst, Ferebee & Furmark, 1997), but we felt this was qualitatively different to patients having realistic spiders strapped to their faces via a virtual reality headset. As a result, and exercising a high-degree of caution for our patients, a human therapist was present with patients during the treatment but were specifically instructed to act as a “computer technician” and only intervene with supportive therapy if necessary. At the close of the clinical trial we inquired via survey with our VRET therapists whether they had provided any basic or more advanced emotional support to their patients and we did identify that a small percentage of all VRET patients received psychologist-level assistance (approx. 10%). It should be noted that other than supportive therapy, the most a clinical psychologist could have done in this context was offer cognitive restructuring and reminding patients of the treatment rationale. Given that some prior meta-analyses have suggested that non-exposure treatments are likely no more efficacious than placebo-controls (Wolitzky-Taylor et al., 2008), actual benefits are likely to be limited. Still future trials will want to explore whether patients can carry out these automated exposures entirely by themselves.

A final remaining question is the quality of the reference treatment (in-vivo exposure therapy) administered to the patients. This is critical given that the efficacy of the VRET treatment was compared only to in-vivo exposure therapy and like most non-inferiority designs the clinical trial did not contain a wait-list control group. The in-vivo exposure therapy used in this trial was OST which is considered a highly efficacious treatment, despite being conducted over only a single session (Zlomke & Davis, 2008). The therapists conducting this treatment were all Masters-level students in their last stages of clinical training as psychologists and thus had ongoing or completed video- taped CBT treatments under supervision; they were also trained particularly for this trial by a licensed psychologist and psychotherapist with specific ex- pertise conducting OST. Supervision was also conducted in groups with all

therapists by the licensed psychologist to discuss the administration of treatments and their own learning. Where this study may have differed from past OST trials is that protocol therapists were not tested on their ability to conduct OST in order to qualify for participation in the study, for example using an Öst developed rating scale for therapist competency in OST (Öst, 2012, p. 171), and treatment sessions were not recorded for adherence to methodology by an external reviewer.

Nevertheless, it is still possible to compare uncontrolled (within-group) effect sizes in this study to previous studies. A systematic review of OST (Zlomke & Davis, 2008) including spider phobia specific studies identified uncontrolled effect sizes for self-reported fear (SPQ) of between Cohen’s d = 3.05 and 2.84 in two studies by Öst using therapist directed treatments. In this study, uncontrolled effect sizes according to the SPQ are similar at post-treatment (d = 2.72) but perhaps slightly lower. It should be noted they are still higher than a more recent study (Andersson et al., 2009a), which obtained an effect size on the SPQ of Cohen’s d = 2.58 in their OST treatment for spider phobia at post-treatment. In regard to BAT as an outcome measure, effect sizes cannot be computed for many prior studies due to standard deviation not being included, however OST treated participants in Andersson et al. (2009a) obtained a Cohen’s d effect size = 2.73 on this measure. This is higher than the effect size on the BAT of OST participants in this study (d = 2.39).

Even if the quality of the implemented OST was slightly worse than previous clinical trials, given the supervisory oversight and additional training included here, it is likely that the OST received by patients was on par or superior to treatments that would be conducted in a community setting. It should be acknowledged that although the clinical practice of OST looks alluringly simple because of its short format, many treatments may be required before the therapist optimizes outcomes. In Ollendick et al., (2009), the efficacy of OST in their RCT with 196 phobic youth aged 7-16 was reported as good (approx. 50% were diagnosis free at post- and follow-up) but was still far lower than outcomes obtained in a prior pilot study (approx. 80% clinically significant change). The author suggested this was as a result of the extensive experience of therapists in the pilot study (mean 8.5 years post-CBT training), and whereas therapists in the RCT met all treatment adherence and competence measures indicating mastery of OST and were also trained and super- vised by the creator of the treatment itself, had previously limited experience conducting OST before participating in this trial.

Finally, selection of the non-inferiority margin is worth discussing. The non-inferiority margin is the maximum allowable difference between a novel and reference treatment required to be able to refer to the novel treatment as

87 not inferior to the comparator. The rationale of non-inferiority trials is therefore to show that a novel treatment is similar in treatment effects to a pre- existing often gold-standard treatment but with added benefits such as lower- cost, greater safety, or fewer sessions. As noted in Piaggio et al. (2012), the CONSORT extension statement for reporting non-inferiority and equivalence trials, there is no one way of specifying the non-inferiority margin, although defining the margin is frequently considered among the most difficult decisions that has to be taken (Wiens, 2002). In most cases, however, the smallest value of clinically important effect is the margin selected (Piaggio et al., 2012). This is in contrast to superiority studies in which finding no significant difference between treatments (i.e. p > .05) does not provide indication of clinical relevance; such a decision requires separate consideration (The Euro- pean Agency for the Evaluation of Medicinal Products, 2008). If the non-inferiority margin selected is too small than an unnecessary expenditure in terms of participants and resources will have been used to determine the difference between two treatments, and if the criteria too large than suggesting non-inferiority with the comparator is meaningless. In this study a non-inferiority margin of 2 points on the in-vivo BAT was selected, which had been used in past spider phobia studies (e.g., Andersson et al., 2009a and Garcia-Palacios et al., 2002) to determine minimum clinically significant change, and was based on an earlier analysis of existing data (Öst et al., 1998). Non-inferiority was then calculated using the confidence-interval approach recommended by Piaggio et al. (2012), with non-inferiority identified according to the lower-bound of the confidence interval difference score, and the upper-bound used to determine whether the difference score crosses null (i.e., was not worse). Piaggio et al. (2012) suggests that often a 1-sided 97.5% confidence interval (i.e., 2- sided 95% CI) is the one selected, however if a 1-sided 5% significant level is deemed acceptable, a 90% 2-sided confidence interval can also be used as was done in this study.

The challenges and future of automated VRET

Automated VRET is a promising means of scaling up exposure therapy, by providing access to a fully contained evidence-based treatment for any individual with a virtual reality device (Lindner et al., 2017). This is in contrast to the current situation where only highly trained individuals are qualified to conduct exposure therapy (normally clinical psychologists) and fewer still are in- terested or willing to provide this evidence-based treatment (Deacon et al., 2013; Olatunji et al., 2009). While the studies in this thesis have shown that individuals are able to derive benefit from this treatment, which was not inferior to gold-standard in-vivo exposure therapy over long-term follow-up, it

remains unclear whether an individual receiving this treatment at a local community healthcare center or in their own home would obtain similar positive results.

Other unknowns include the relative novelty of virtual reality, which at the start of the clinical trial was at a peak and still remains high (Lindner et al., 2017), however could lessen over time as it has done before. The credibility and expectancy of a treatment has been shown to relate positively to its efficacy (Devilly & Borkovec, 2000) and time will tell whether interest in and expectations about virtual reality and related technologies remains high. Nev- ertheless, VRET treatments as this thesis report have now over more than 20 years of testing proven efficacious even during periods when its novelty has lessened, therefore the continued ability of the technology to administer benefits to patients seems plausible. One aspect of the relationship between virtual reality and society that has not changed over 20 years, but soon might however, is once virtual reality proliferates to the point at which it becomes as common as using a computer or smartphone. Facebook and its founder Mark Zuckerberg particularly are famous for stating the goal of getting 1 billion people using virtual reality devices (Morris, 2017), and the prices and quality of these devices are improving. Will it be possible for individuals with specific phobia to be treated with VRET created by small independent developers once they become accustomed to the visuals and interactivity of the equivalent of triple-A computer game? Time will tell.

As suggested by Jaron Lanier in his book Dawn of the new everything: encounters with reality and virtual reality (Lanier, 2017) what virtual reality is will always be superior what it is currently capable of. As virtual reality improves, it trains individuals to become more discerning users of the technology. To someone in the 1990s the giant polygons and low-resolution screens of virtual reality were impressive, to someone now they might be nearly impossible to view for any length of time. Current cutting-edge technologies exist today such as ray-tracing, which simulates the physical properties of light as opposed to rasterized graphics which layer pixels on a mesh of polygons and triangles. They may take hours or days of processing time to compute a static image (although that is rapidly changing) but produce scenes that are largely indistinguishable from photographs of real objects (Scarfe and Glennester, 2019). However, close observation of even these new technologies may eventually be seen to be insufficient. Jaron Lanier suggests that what is actually occurring is that individuals are becoming more discriminating users of our own sensory systems capacity to perceive. Perhaps, this is similar to the still constant and rapid improvement of picture taking technology where despite nearly two hundred years of invention and progress we still continue to notice improvements.

89 Even if VRET applications never meet the same visual standards as video games, efficacy may still be more dependent on the quality of the psychological theory integrated into the software, than how highly detailed the stimuli are. If stimuli already return natural reactions from a patient than any additional increase in realism may provide only diminishing returns (Scarfe & Glennerster, 2019). Similar to how improvements in traditional in-vivo exposure therapy have unfolded over generations, what is quite likely is that rather than a one-size-fits-all generic VRET application for treating spider phobia, we will begin to see tailoring between different specialized components within applications based on individual unique fears or cognitive distortions. For example, some categories of spider phobic patients may be particularly sensitive to the movements of spiders whereas others high in disgust proneness may benefit from viewing other individuals touching a spider, triggering a fear of spreading contagion. As Öst has recommended in past guidance, focusing treatments on unique fears rather than a broad-based approach are likely to improve outcomes. Given the flexibility in coding virtual reality simulations and environments, evaluating the possible improved efficacy of targeting unique fears or cognitive distortions versus a generic approach, is a theoretical question that future research will be able to answer.

Ultimately, the most hopeful future for VRET may lie not in perfecting the treatment for individual unique idiosyncratic fears and information processing biases with ever greater precision, however warranted this is; but ensuring that the basic psychological tools that have been incorporated into this technology for decades may reach as many individuals as possible. There is the real possibility that in the future algorithmic-supported VRET (Bălan et al., 2020) that uses machine learning to automatically categorize patients and optimize their treatment, but are based on large datasets from users in Western countries, will create treatments perfectly suited to individual high-income patients but less appropriate for those in lower income brackets or in other regions of the world. Even the treatment for spider phobia included here may be considered opti- mized to a Western population given the use of an environment incorporating a classic Swedish row house and interior common to more affluent European countries. This may be resolved in the future using augmented reality which projects phobic stimuli directly onto whatever is in the individual’s current physical environment. These technologies are currently undergoing rapid advances in their basic technology (e.g., Microsoft Hololens), and have shown in past studies similar efficacy as compared to in-vivo exposure therapy for small animal phobias (Botella et al., 2016).

In high-income countries where individuals generally have access to high- quality psychological care, VRET may prove a worthwhile first step for treating their specific phobia and which can continue on with self-exposures or a visit to a clinical psychologist if necessary. On the other hand, in low and

middle-income countries automated VRET may be one of the few efficacious treatments for specific phobia that are available. Effort will need to be made to ensure that despite clinical research and information technologies ever-ex- panding ambitions for cutting edge evidence-based treatments, we will continue to develop treatments that are effective even in low-cost virtual reality devices and affordable by rural clinics, hospitals and one day the majority of individuals. In that way we will continue a tradition started by some of the great researchers in this field such as Lars-Görgan Öst and others who made strong and innovative efforts to ensure that effective evidence-based treatments could be disseminated as widely as possible.

91 Acknowledgements

This doctoral dissertation would not have been possible without a large and talented group of people supporting me and this research.

I want to begin by thanking my supervisors Per Carlbring and Lotta Reu- terskiöld who not only helped feed and sustain my desire for learning but helped structure and organize it - and ensured I finished on time! I really can’t thank you enough for all your assistance over these years. To Per particularly, you made a dream come true by inviting me to work with you, and reinforcing my application for the PhD. I will always be very thankful for the exciting opportunities you opened for me. To Philip Lindner, I can truly and humbly say that this work would not have been possible without your support. Thank you for your collaboration, friendship, and guidance. To William Hamilton, thank you for your ambition to make virtual reality treatments a reality. None of what we have done here would have been happened without your ingenuity and hard work. Thank you to the many senior academics and administrators who have made this degree possible. Fredrik Jönsson, you were not only an excellent organizer of the PhD program, but an excellent support throughout all this. Toron Lindholm, we met early on prior to my PhD studies, and I was always grateful for your encouragement. Ann-Charlotte Smedler, we’ve had many engaging interactions over the years. Thank you for your assistance including the excellent edits to this thesis. Laura Ferrer Wreder, one of the few other North Americans around, thank you for all the friendly conversations that made the Department feel more like home. Stefan Wiens, thank you for all our time spent working together and teaching me about biological measures and analysis. Tommy Olin, I’ve called on you countless times for help and its always been great interacting. Monika Karlsson, you’ve been a warm and friendly person to work with, and thanks for all your assistance. Håkan Fischer, thank you for all the positive and inspiring energy you brought to us at the Department. To all my PhD brothers and sisters in House 8: Lichen, Jonas, Katja, Ri- kard, and not to forget David and Alexander, thank you all for our time together; Lillian too. Jakob and Karin you arrived later but you’re no less part of this family. Lars and Robert, I owe you each many thanks for your help,

friendship, and great conversions over the years. To all my PhD brothers and sisters in House 14 and GEL, you’re also part of this extended academic family. There are many memories I’ll have forever, good times that made getting through the hard times of a PhD possible. Turning to my biological family, thank you Mom and Dad (Helen and Maury). Apparently, as a young child I bombarded you with “why” questions. Thank you for supporting my curiosity, helping me find answers through education, and bringing me on amazing adventures around the world. Thank you even more for all the love, compassion, openness, and endless support. To my grandparents Edna and Roy, Harold and Gloria, you have been fierce in your love of learning and knowledge. You also lived beautiful lives of truthfulness, family, and service to others. You’ll always be shining examples to me of how to lead a good life. To my siblings Hayley and Zach, there’s not a day that goes by that I’m not grateful I have you in my life. Thanks for always being there for me and for your love and support over the years, and for all the laughter and joy. Also, for adding my wonderful brother-in-law Liam to the family. Can’t wait for more adventures. To Charisse, you’ve been my closest friend for the past decade and none of this would have happened if it wasn’t for you. Thank you for listening to my advice to move to Sweden and bringing me along all these years later. You’ve made this period in Stockholm so much fun. To all my other Baháʼí brothers and sisters in Stockholm, Anil, Assrar, Zhena, Ina, Payam, Shahram, Ersida, Elham, Gloria, Adib, Neda, Nadi, Mahsa, Anna, Sara, Lua, Sana, Camelia, Ruha, Mahan, Samin and Abeer, Af- nan, Faezeh, Arfa, Patrik, Moa and Soheil. What wonderful times we’ve had over these years. Thanks also to the rest of my friends, family, cousins, aunts and uncles, who have helped me in ways big and small, and who I’m so grateful to have in my life. Particularly to my Uncle Michael, thank you for giving me so much of yourself, your curiosity, generosity, excitement, and interest in my life and work. Finally, I remember my Uncle Peter who recently passed away in the tragic crash of Ethiopian Airlines flight 302. Thank you for your lifelong passion for justice, for combatting poverty, and fighting for environmental sustainability. Your memory will always inspire my efforts.

“This was the first time that we actually escaped from the earth, and at that time I suddenly realized that everything in life is relative. When you’re in a room your world revolves around those walls, when you’re outside then your world revolves around what your eye can see, and suddenly when you’re in a spacecraft you think in terms of oceans, of islands”. James Lovell, Apollo Astronaut

References

Abramowitz, J. S., & Blakey, S. M. (Eds.). (2020). Clinical handbook of fear and anxiety: Maintenance processes and treatment mechanisms. American Psychological Association. https://doi.org/10.1037/0000150-000 American Psychiatric Association. (2013). Diagnostic and Statistical Manual of Mental Disorders: Diagnostic and Statistical Manual of Mental Disorders (Fifth). American Psychiatric Association. Anderson, P. L., Price, M., Edwards, S. M., Obasaju, M. A., Schmertz, S. K., Zimand, E., & Calamaras, M. R. (2013). Virtual reality exposure therapy for social anxiety disorder: A randomized controlled trial. Journal of Consulting and Clinical Psychology, 81(5), 751–760. https://doi.org/10.1037/a0033559 Andersson, G. (2009b). Using the Internet to provide cognitive behaviour therapy. Behaviour Research and Therapy, 47(3), 175–180. https://doi.org/10.1016/j.brat.2009.01.010 Andersson, G., Titov, N., Dear, B. F., Rozental, A., & Carlbring, P. (2019). Internet-delivered psychological treatments: From innovation to implementation. World Psychiatry, 18(1), 20–28. https://doi.org/10.1002/wps.20610 Andersson, G., Waara, J., Jonsson, U., Malmaeus, F., Carlbring, P., & Öst, L. (2009a). Internet‐Based Self‐Help Versus One‐Session Exposure in the Treatment of Spider Phobia: A Randomized Controlled Trial. Cognitive Behaviour Therapy, 38(2), 114–120. https://doi.org/10.1080/16506070902931326 Awaad, R., & Ali, S. (2016). A modern conceptualization of phobia in al- Balkhi’s 9th century treatise: Sustenance of the Body and Soul. Jour- nal of Anxiety Disorders, 37, 89–93. https://doi.org/10.1016/j.janxdis.2015.11.003 Azriel, O., & Bar-Haim, Y. (2020). Attention bias. In J. S. Abramowitz & S. M. Blakey (Eds.), Clinical handbook of fear and anxiety: Mainte- nance processes and treatment mechanisms (pp. 203–218). American Psychological Association. https://doi.org/10.1037/0000150-000 Bailenson, J. N., Blascovich, J., Beall, A. C., & Loomis, J. M. (2003). Inter- personal Distance in Immersive Virtual Environments. Personality and Social Psychology Bulletin, 29(7), 819–833. https://doi.org/10.1177/0146167203029007002

95 Bălan, O., Moise, G., Moldoveanu, A., Leordeanu, M., & Moldoveanu, F. (2020). An Investigation of Various Machine and Deep Learning Techniques Applied in Automatic Fear Level Detection and Acro- phobia Virtual Therapy. Sensors (Basel, Switzerland), 20(2). https://doi.org/10.3390/s20020496 Barazzone, N., Cavanagh, K., & Richards, D. A. (2012). Computerized cognitive behavioural therapy and the therapeutic alliance: A qualitative enquiry. British Journal of Clinical Psychology, 51(4), 396–417. https://doi.org/10.1111/j.2044-8260.2012.02035.x Barger, M. S., & White, W. B. (2000). The Daguerreotype: Nineteenth-Cen- tury Technology and Modern Science. JHU Press. Barlow, D. H. (2002). Anxiety and its disorders: The nature and treatment of anxiety and panic (2nd ed). Guilford Press. Beesdo-Baum, K., & Knappe, S. (2012). Developmental Epidemiology of Anxiety Disorders. Child and Adolescent Psychiatric Clinics of North America, 21(3), 457–478. https://doi.org/10.1016/j.chc.2012.05.001 Benbow, A. A., & Anderson, P. L. (2019). A meta-analytic examination of attrition in virtual reality exposure therapy for anxiety disorders. Jour- nal of Anxiety Disorders, 61, 18–26. https://doi.org/10.1016/j.janxdis.2018.06.006 Berger, T. (2017). The therapeutic alliance in internet interventions: A narrative review and suggestions for future research. Psychotherapy Re- search, 27(5), 511–524. https://doi.org/10.1080/10503307.2015.1119908 Berry, K., Salter, A., Morris, R., James, S., & Bucci, S. (2018). Assessing Therapeutic Alliance in the Context of mHealth Interventions for Mental Health Problems: Development of the Mobile Agnew Rela- tionship Measure (mARM) Questionnaire. Journal of Medical Inter- net Research, 20(4), e90. https://doi.org/10.2196/jmir.8252 Bigelow, K. M., & Morris, E. K. (2001). John B. Watson’s advice on child rearing: Some historical context. Behavioral Development Bulletin, 10(1), 26–30. http://dx.doi.org/10.1037/h0100479 Boettcher, H., & Barlow, D. H. (2019). The unique and conditional effects of interoceptive exposure in the treatment of anxiety: A functional analysis. Behaviour Research and Therapy, 117, 65–78. https://doi.org/10.1016/j.brat.2018.12.002 Boettcher, H., Brake, C. A., & Barlow, D. H. (2016). Origins and outlook of interoceptive exposure. Journal of Behavior Therapy and Experi- mental Psychiatry, 53, 41–51. https://doi.org/10.1016/j.jbtep.2015.10.009 Bordin, E. S. (1979). The generalizability of the psychoanalytic concept of the working alliance. Psychotherapy: Theory, Research & Practice, 16(3), 252–260. https://doi.org/10.1037/h0085885

Botella, C., Baños, R. M., Perpiñá, C., Villa, H., Alcañiz, M., & Rey, A. (1998). Virtual reality treatment of claustrophobia: A case report. Be- haviour Research and Therapy, 36(2), 239–246. https://doi.org/10.1016/S0005-7967(97)10006-7 Botella, Cristina, Pérez-Ara, M. Á., Bretón-López, J., Quero, S., García-Pala- cios, A., & Baños, R. M. (2016). In Vivo versus Augmented Reality Exposure in the Treatment of Small Animal Phobia: A Randomized Controlled Trial. PLOS ONE, 11(2), e0148237. https://doi.org/10.1371/journal.pone.0148237 Bouchard, S., Côté, S., St-Jacques, J., Robillard, G., & Renaud, P. (2006). Effectiveness of virtual reality exposure in the treatment of arachnophobia using 3D games. Technology and Health Care: Official Jour- nal of the European Society for Engineering and Medicine, 14(1), 19– 27. Bouchard, S., Dumoulin, S., Robillard, G., Guitard, T., Klinger, É., Forget, H., Loranger, C., & Roucaut, F. X. (2017). Virtual reality compared with in vivo exposure in the treatment of social anxiety disorder: A three-arm randomised controlled trial. British Journal of Psychiatry, 210(4), 276–283. https://doi.org/10.1192/bjp.bp.116.184234 Breslau, J., Miller, E., Jin, R., Sampson, N. A., Alonso, J., Andrade, L. H., Bromet, E. J., de Girolamo, G., Demyttenaere, K., Fayyad, J., Fukao, A., Gălăon, M., Gureje, O., He, Y., Hinkov, H. R., Hu, C., Kovess- Masfety, V., Matschinger, H., Medina-Mora, M. E., … Kessler, R. C. (2011). A multinational study of mental disorders, marriage, and divorce. Acta Psychiatrica Scandinavica, 124(6), 474–486. https://doi.org/10.1111/j.1600-0447.2011.01712.x Carl, E., Stein, A. T., Levihn-Coon, A., Pogue, J. R., Rothbaum, B., Em- melkamp, P., Asmundson, G. J. G., Carlbring, P., & Powers, M. B. (2019). Virtual reality exposure therapy for anxiety and related disorders: A meta-analysis of randomized controlled trials. Journal of Anx- iety Disorders, 61, 27–36. https://doi.org/10.1016/j.janxdis.2018.08.003 Carlin, A. S., Hoffman, H. G., & Weghorst, S. (1997). Virtual reality and tactile augmentation in the treatment of spider phobia: A case report. Be- haviour Research and Therapy, 35(2), 153–158. https://doi.org/10.1016/S0005-7967(96)00085-X Cavanagh, K., & Millings, A. (2013). (Inter)personal Computing: The Role of the Therapeutic Relationship in E-mental Health. Journal of Contem- porary Psychotherapy, 43(4), 197–206. https://doi.org/10.1007/s10879-013-9242-z Chan, E. K. F., Timmermann, A., Baldi, B. F., Moore, A. E., Lyons, R. J., Lee, S.-S., Kalsbeek, A. M. F., Petersen, D. C., Rautenbach, H., Förtsch, H. E. A., Bornman, M. S. R., & Hayes, V. M. (2019). Human origins in a southern African palaeo-wetland and first migrations. Nature, 575(7781), 185–189. https://doi.org/10.1038/s41586-019-1714-1

97 Choy, Y., Fyer, A. J., & Lipsitz, J. D. (2007). Treatment of specific phobia in adults. Clinical Psychology Review, 27(3), 266–286. https://doi.org/10.1016/j.cpr.2006.10.002 Clark, L. A., & Watson, D. (2019). Constructing validity: New developments in creating objective measuring instruments. Psychological Assess- ment. https://doi.org/10.1037/pas0000626 Computerhistory.org. (2019). Intel “x86” Family and the Microprocessor Wars—CHM Revolution. https://www.computerhistory.org/revolution/digital-logic/12/330 Côté, S., & Bouchard, S. (2005). Documenting the Efficacy of Virtual Reality Exposure with Psychophysiological and Information Processing Measures. Applied Psychophysiology and Biofeedback, 30(3), 217– 232. https://doi.org/10.1007/s10484-005-6379-x Craske, M. G., Kircanski, K., Zelikowsky, M., Mystkowski, J., Chowdhury, N., & Baker, A. (2008). Optimizing inhibitory learning during exposure therapy. Behaviour Research and Therapy, 46(1), 5–27. https://doi.org/10.1016/j.brat.2007.10.003 Craske, M. G., Rauch, S. L., Ursano, R., Prenoveau, J., Pine, D. S., & Zinbarg, R. E. (2009). What is an anxiety disorder? Depression and Anxiety, 26(12), 1066–1085. https://doi.org/10.1002/da.20633 Craske, M. G., & Stein, M. B. (2016). Anxiety. The Lancet, 388(10063), 3048–3059. https://doi.org/10.1016/S0140-6736(16)30381-6 Craske, M. G., Treanor, M., Conway, C. C., Zbozinek, T., & Vervliet, B. (2014). Maximizing exposure therapy: An inhibitory learning approach. Behaviour Research and Therapy, 58, 10–23. https://doi.org/10.1016/j.brat.2014.04.006 Crocq, M.-A. (2015). A history of anxiety: From Hippocrates to DSM. Dia- logues in Clinical Neuroscience, 17(3), 319–325. Cuijpers, P., Veen, S. C. van, Sijbrandij, M., Yoder, W., & Cristea, I. A. (2020). Eye movement desensitization and reprocessing for mental health problems: A systematic review and meta-analysis. Cognitive Behaviour Therapy, 0(0), 1–16. https://doi.org/10.1080/16506073.2019.1703801 Davey, G. C. L. (1991). Characteristics of individuals with fear of spiders. Anxiety Research, 4(4), 299–314. https://doi.org/10.1080/08917779208248798 Davis, M., Ressler, K., Rothbaum, B. O., & Richardson, R. (2006). Effects of D-Cycloserine on Extinction: Translation From Preclinical to Clinical Work. Biological Psychiatry, 60(4), 369–375. https://doi.org/10.1016/j.biopsych.2006.03.084 Davis, T. E., Ollendick, T. H., & Öst, L.-G. (Eds.). (2012). Intensive one-session treatment of specific phobias. Springer-Verlag.

Davis, T. E., Ollendick, T. H., & Öst, L.-G. (2019). One-Session Treatment of Specific Phobias in Children: Recent Developments and a System- atic Review. Annual Review of Clinical Psychology, 15(1), 233–256. https://doi.org/10.1146/annurev-clinpsy-050718-095608 De Jongh, A., Ten Broeke, E., & Renssen, M. R. (1999). Treatment of Specific Phobias with Eye Movement Desensitization and Reprocessing (EMDR): Protocol, Empirical Status, and Conceptual Issues. Journal of Anxiety Disorders, 13(1), 69–85. https://doi.org/10.1016/S0887- 6185(98)00040-1 de Quervain, D. J.-F., Bentz, D., Michael, T., Bolt, O. C., Wiederhold, B. K., Margraf, J., & Wilhelm, F. H. (2011). Glucocorticoids enhance extinction-based psychotherapy. Proceedings of the National Academy of Sciences of the United States of America, 108(16), 6621–6625. https://doi.org/10.1073/pnas.1018214108 Deacon, B. J., Farrell, N. R., Kemp, J. J., Dixon, L. J., Sy, J. T., Zhang, A. R., & McGrath, P. B. (2013). Assessing therapist reservations about exposure therapy for anxiety disorders: The Therapist Beliefs about Ex- posure Scale. Journal of Anxiety Disorders, 27(8), 772–780. https://doi.org/10.1016/j.janxdis.2013.04.006 Devilly, G. J., & Borkovec, T. D. (2000). Psychometric properties of the credibility/expectancy questionnaire. Journal of Behavior Therapy and Experimental Psychiatry, 31(2), 73–86. https://doi.org/10.1016/s0005-7916(00)00012-4 Diemer, J., Alpers, G. W., Peperkorn, H. M., Shiban, Y., & Mühlberger, A. (2015). The impact of perception and presence on emotional reactions: A review of research in virtual reality. Frontiers in Psychology, 6. https://doi.org/10.3389/fpsyg.2015.00026 Donker, T., Cornelisz, I., van Klaveren, C., van Straten, A., Carlbring, P., Cuijpers, P., & van Gelder, J.-L. (2019). Effectiveness of Self-guided App-Based Virtual Reality Cognitive Behavior Therapy for Acro- phobia: A Randomized Clinical Trial. JAMA Psychiatry, 76(7), 682. https://doi.org/10.1001/jamapsychiatry.2019.0219 Dour, H. J., Brown, L. A., & Craske, M. G. (2016). Positive valence reduces susceptibility to return of fear and enhances approach behavior. Jour- nal of Behavior Therapy and Experimental Psychiatry, 50, 277–282. https://doi.org/10.1016/j.jbtep.2015.09.010 Eaton, W. W., Kalaydjian, A., Scharfstein, D. O., Mezuk, B., & Ding, Y. (2007). Prevalence and incidence of depressive disorder: The Balti- more ECA follow-up, 1981-2004. Acta Psychiatrica Scandinavica, 116(3), 182–188. https://doi.org/10.1111/j.1600-0447.2007.01017.x Eaton, William W, Bienvenu, O. J., & Miloyan, B. (2018). Specific phobias. The Lancet Psychiatry, 5(8), 678–686. https://doi.org/10.1016/S2215-0366(18)30169-X Eaton, William W., Kessler, R. C., Mortensen, P. B., Rebok, G. W., & Roth, K. (2019). The Population Dynamics of Mental Disorders. Oxford

99 University Press. https://www-oxfordscholarship- com.ezp.sub.su.se/view/10.1093/oso/9780190916602.001.0001/oso- 9780190916602-chapter-7 Etkin, A., & Wager, T. D. (2007). Functional Neuroimaging of Anxiety: A Meta-Analysis of Emotional Processing in PTSD, Social Anxiety Dis- order, and Specific Phobia. American Journal of Psychiatry, 164(10), 1476–1488. https://doi.org/10.1176/appi.ajp.2007.07030504 Fairburn, C. G., & Patel, V. (2017). The impact of digital technology on psychological treatments and their dissemination. Behaviour Research and Therapy, 88, 19–25. https://doi.org/10.1016/j.brat.2016.08.012 Falkenström, F., Hatcher, R. L., & Holmqvist, R. (2015). Confirmatory Factor Analysis of the Patient Version of the Working Alliance Inventory– Short Form Revised. Assessment, 22(5), 581–593. https://doi.org/10.1177/1073191114552472 Fernández-Álvarez, J., Rozental, A., Carlbring, P., Colombo, D., Riva, G., Anderson, P. L., Baños, R. M., Benbow, A. A., Bouchard, S., Bretón- López, J. M., Cárdenas, G., Difede, J., Emmelkamp, P., García-Pala- cios, A., Guillén, V., Hoffman, H., Kampann, I., Moldovan, R., Mühl- berger, A., … Botella, C. (2019). Deterioration rates in Virtual Reality Therapy: An individual patient data level meta-analysis. Journal of Anxiety Disorders, 61, 3–17. https://doi.org/10.1016/j.janxdis.2018.06.005 First, M. B., Robert, L., Gibbon, M., & Williams, J. B. W. (2002). Structure clinical interview for DSM-IV-TR Axis I, research version. Biometric Research. Fleming, T. M., Bavin, L., Stasiak, K., Hermansson-Webb, E., Merry, S. N., Cheek, C., Lucassen, M., Lau, H. M., Pollmuller, B., & Hetrick, S. (2017). Serious Games and Gamification for Mental Health: Current Status and Promising Directions. Frontiers in Psychiatry, 7. https://doi.org/10.3389/fpsyt.2016.00215 Flückiger, C., Del Re, A. C., Wampold, B. E., & Horvath, A. O. (2018). The alliance in adult psychotherapy: A meta-analytic synthesis. Psycho- therapy, 55(4), 316–340. https://doi.org/10.1037/pst0000172 Foa, E. B., & Kozak, M. J. (1986). Emotional processing of fear: Exposure to corrective information. Psychological Bulletin, 99(1), 20–35. https://doi.org/10.1037/0033-2909.99.1.20 Fodor, L. A., Coteț, C. D., Cuijpers, P., Szamoskozi, Ștefan, David, D., & Cristea, I. A. (2018). The effectiveness of virtual reality based interventions for symptoms of anxiety and depression: A meta-analysis. Scientific Reports, 8(1), 1–13. https://doi.org/10.1038/s41598-018- 28113-6 Freeman, D., Reeve, S., Robinson, A., Ehlers, A., Clark, D., Spanlang, B., & Slater, M. (2017). Virtual reality in the assessment, understanding, and treatment of mental health disorders. Psychological Medicine, 47(14), 2393–2400. https://doi.org/10.1017/S003329171700040X

100

Freeman, D., Haselton, P., Freeman, J., Spanlang, B., Kishore, S., Albery, E., Denne, M., Brown, P., Slater, M., & Nickless, A. (2018). Automated psychological therapy using immersive virtual reality for treatment of fear of heights: A single-blind, parallel-group, randomised controlled trial. The Lancet Psychiatry, 5(8), 625–632. https://doi.org/10.1016/S2215-0366(18)30226-8 Garb, J. E., González, A., & Gillespie, R. G. (2004). The black widow spider genus Latrodectus (Araneae: Theridiidae): phylogeny, biogeography, and invasion history. Molecular Phylogenetics and Evolution, 31(3), 1127–1142. https://doi.org/10.1016/j.ympev.2003.10.012 Garcia-Palacios, A., Botella, C., Hoffman, H., & Fabregat, S. (2007). Com- paring Acceptance and Refusal Rates of Virtual Reality Exposure vs. In Vivo Exposure by Patients with Specific Phobias. CyberPsychol- ogy & Behavior, 10(5), 722–724. https://doi.org/10.1089/cpb.2007.9962 Garcia-Palacios, A., Hoffman, H., Carlin, A., Furness, T. A., & Botella, C. (2002). Virtual reality in the treatment of spider phobia: A controlled study. Behaviour Research and Therapy, 40(9), 983–993. https://doi.org/10.1016/S0005-7967(01)00068-7 Garcia-Palacios, A., Hoffman, H. G., Kwong See, S., Tsai, A., & Botella, C. (2001). Redefining Therapeutic Success with Virtual Reality Expo- sure Therapy. CyberPsychology & Behavior, 4(3), 341–348. https://doi.org/10.1089/109493101300210231 Glantz, K., Durlach, N. I., Barnett, R. C., & Aviles, W. A. (1996). Virtual reality (VR) for psychotherapy: From the physical to the social environment. Psychotherapy, 33(3), 464–473. http://dx.doi.org/10.1037/0033-3204.33.3.464 Hair, N. L., Hanson, J. L., Wolfe, B. L., & Pollak, S. D. (2015). Association of Child Poverty, Brain Development, and Academic Achievement. JAMA Pediatrics, 169(9), 822–829. https://doi.org/10.1001/jamapediatrics.2015.1475 Harris, B. (2002). Whatever happened to little Albert? American Psychologi- cal Association. https://doi.org/10.1037/10421-012 Hatcher, R. L., & Barends, A. W. (2006). How a return to theory could help alliance research. Psychotherapy, 43(3), 292–299. http://dx.doi.org/10.1037/0033-3204.43.3.292 Hatcher, R. L., & Gillaspy, J. A. (2006). Development and validation of a revised short version of the working alliance inventory. Psychother- apy Research, 16(1), 12–25. https://doi.org/10.1080/10503300500352500 Hayes, S. C., Strosahl, K. D., & Wilson, K. G. (2012). Acceptance and commitment therapy: The process and practice of mindful change, 2nd ed. Guilford Press. Hayes-Skelton, S. A., & Eustis, E. H. (2020). Experiential avoidance. In J. S. Abramowitz & S. M. Blakey (Eds.), Clinical handbook of fear and

101 anxiety: Maintenance processes and treatment mechanisms (pp. 115– 131). American Psychological Association. https://doi.org/10.1037/0000150-000 Hellström, K., & Öst, L.-G. (1995). One-session therapist directed exposure vs two forms of manual directed self-exposure in the treatment of spider phobia. Behaviour Research and Therapy, 33(8), 959–965. https://doi.org/10.1016/0005-7967(95)00028-V HITLab. (n.d.). HITLab Projects: VR Therapy for Spider Phobia. Retrieved December 28, 2019, from http://www.hitl.washington.edu/projects/exposure/ Hodges, L. F., Rothbaum, B. O., Kooper, R., Opdyke, D., Meyer, T., Graaff, J. J. de, Williford, J. S., & North, M. M. (1994). Presence as The De- fining Factor in a VR Application: Virtual Reality Graded Exposure in the Treatment of Acrophobia. Hoffman, D. D. (2018). The Interface Theory of Perception. In J. T. Wixted & J. Serences (Eds.), Stevens’ Handbook of Experimental Psychology and Cognitive Neuroscience, Sensation, Perception, and Attention (Vol. 2). John Wiley & Sons, Incorporated. http://ebookcen- tral.proquest.com/lib/sub/detail.action?docID=5295000 Hoffman, H. G., Garcia-Palacios, A., Carlin, A., Furness III, T. A., & Botella- Arbona, C. (2003). Interfaces That Heal: Coupling Real and Virtual Objects to Treat Spider Phobia. International Journal of Human- Computer Interaction, 16(2), 283–300. https://doi.org/10.1207/S15327590IJHC1602_08 Hoffman, Y. S. G., Pitcho-Prelorentzos, S., Ring, L., & Ben-Ezra, M. (2019). “Spidey Can”: Preliminary Evidence Showing Arachnophobia Symp- tom Reduction Due to Superhero Movie Exposure. Frontiers in Psy- chiatry, 10. https://doi.org/10.3389/fpsyt.2019.00354 Horvath, A. O. (2011). Alliance in Common Factor Land: A view through the research lens | Research in Psychotherapy: Psychopathology, Process and Outcome. Research in Psychotherapy, 14(1), 121–135. https://doi.org/10.4081/ripppo.2011.45 Horvath, A. O. (2018a). The psychotherapy relationship: Where does the alliance fit? In O. Tishby (Ed.), Developing the therapeutic relationship: Integrating case studies, research, and practice (pp. 15–28, Chapter xiii, 362 Pages). American Psychological Association (Washington, DC, US). https://search.proquest.com/docview/2061384560/abstract/7120E7F251D74C34PQ/2 Horvath, A. O. (2018b). Research on the alliance: Knowledge in search of a theory. Psychotherapy Research, 28(4), 499–516. https://doi.org/10.1080/10503307.2017.1373204 Horvath, A. O., & Greenberg, L. S. (1989). Development and validation of the Working Alliance Inventory. Journal of Counseling Psychology, 36(2), 223–233. http://dx.doi.org/10.1037/0022-0167.36.2.223

102

Horvath, A. O., & Luborsky, L. (1993). The role of the therapeutic alliance in psychotherapy. Journal of Consulting and Clinical Psychology, 61(4), 561–573. https://doi.org/10.1037/0022-006X.61.4.561 Jacquart, J., Roquet, R. F., Papini, S., Powers, M. B., Rosenfield, D., Smits, J. A. J., & Monfils, M.-H. (2017). Effects of acute exercise on fear extinction in rats and exposure therapy in humans: Null findings from five experiments. Journal of Anxiety Disorders, 50, 76–86. https://doi.org/10.1016/j.janxdis.2017.05.010 Kessler, R. C., Amminger, G. P., Aguilar‐Gaxiola, S., Alonso, J., Lee, S., & Ustun, T. B. (2007). Age of onset of mental disorders: A review of recent literature. Current Opinion in Psychiatry, 20(4), 359–364. https://doi.org/10.1097/YCO.0b013e32816ebc8c Kessler, R. C., Petukhova, M., Sampson, N. A., Zaslavsky, A. M., & Wittchen, H.-U. (2012). Twelve-month and lifetime prevalence and lifetime morbid risk of anxiety and mood disorders in the United States: Anx- iety and mood disorders in the United States. International Journal of Methods in Psychiatric Research, 21(3), 169–184. https://doi.org/10.1002/mpr.1359 Kleim, B., Wilhelm, F. H., Temp, L., Margraf, J., Wiederhold, B. K., & Rasch, B. (2014). Sleep enhances exposure therapy. Psychological Medicine, 44(7), 1511–1519. https://doi.org/10.1017/S0033291713001748 Knopf, K., & Pössel, P. (2009). Individual response differences in spider phobia: Comparing phobic and non-phobic women of different reactivity levels. Anxiety, Stress, & Coping, 22(1), 39–55. https://doi.org/10.1080/10615800802169358 Knowles, K. A., Cox, R. C., Armstrong, T., & Olatunji, B. O. (2019). Cogni- tive mechanisms of disgust in the development and maintenance of psychopathology: A qualitative review and synthesis. Clinical Psy- chology Review, 69, 30–50. https://doi.org/10.1016/j.cpr.2018.06.002 Lanier, J. (2017). Dawn of the New Everything: Encounters with Reality and Virtual Reality. Henry Holt and Company. Lavallee, K. L., & Schneider, S. (2019). Chapter 8—Separation Anxiety Dis- order. In S. N. Compton, M. A. Villabø, & H. Kristensen (Eds.), Pe- diatric Anxiety Disorders (pp. 151–176). Academic Press. https://doi.org/10.1016/B978-0-12-813004-9.00008-6 LDV capital. (2017, August). LDV Capital. https://www.ldv.co/insights/2017 Lewis, L. E., & Drewett, R. F. (2006). Psychophysiological correlates of anxiety: A single-case study. Journal of Anxiety Disorders, 20(6), 829– 835. https://doi.org/10.1016/j.janxdis.2005.08.005 Lieb, R., Miché, M., Gloster, A. T., Beesdo-Baum, K., Meyer, A. H., & Wittchen, H.-U. (2016). Impact of Specific Phobia on the Risk of On- set Of Mental Disorders: A 10-Year Prospective-Longitudinal Com- munity Study of Adolescents and Young Adults. Depression and Anx- iety, 33(7), 667–675. https://doi.org/10.1002/da.22487

103 Lindner, P. (2019). Attitudes Toward and Familiarity With Virtual Reality Therapy Among Practicing Cognitive Behavior Therapists: A Cross- Sectional Survey Study in the Era of Consumer VR Platforms. Fron- tiers in Psychology, 10, 12. Lindner, P., Miloff, A., Bergman, C., Andersson, G., Hamilton, W., & Carl- bring, P. (2020). Gamified, Automated Virtual Reality Exposure Therapy for Fear of Spiders: A Single-Subject Trial Under Simulated Real-World Conditions. Frontiers in Psychiatry, 11. https://doi.org/10.3389/fpsyt.2020.00116 Lindner, P., Miloff, A., Hamilton, W., Reuterskiöld, L., Andersson, G., Pow- ers, M. B., & Carlbring, P. (2017). Creating state of the art, next-generation Virtual Reality exposure therapies for anxiety disorders using consumer hardware platforms: Design considerations and future directions. Cognitive Behaviour Therapy, 46(5), 404–420. https://doi.org/10.1080/16506073.2017.1280843 Lipp, O. V., Waters, A. M., Luck, C. C., Ryan, K. M., & Craske, M. G. (2019). Novel approaches for strengthening human fear extinction: The roles of novelty, additional USs, and additional GSs. Behaviour Research and Therapy, 103529. https://doi.org/10.1016/j.brat.2019.103529 Lipsitz, J. D., Mannuzza, S., Klein, D. F., Ross, D. C., & Fyer, A. J. (1999). Specific phobia 10–16 years after treatment. Depression and Anxiety, 10(3), 105–111. https://doi.org/10.1002/(SICI)1520- 6394(1999)10:3<105::AID-DA3>3.0.CO;2-X LoBue, V., & Adolph, K. E. (2019). Fear in infancy: Lessons from snakes, spiders, heights, and strangers. Developmental Psychology, 55(9), 1889–1907. http://dx.doi.org.ezp.sub.su.se/10.1037/dev0000675 Magee, W. J., Eaton, W. W., Wittchen, H.-U., McGonagle, K. A., & Kessler, R. C. (1996). Agoraphobia, Simple Phobia, and Social Phobia in the National Comorbidity Survey. Archives of General Psychiatry, 53(2), 159–168. https://doi.org/10.1001/archpsyc.1996.01830020077009 Maples-Keller, J. L., Bunnell, B. E., Kim, S.-J., & Rothbaum, B. O. (2017). The Use of Virtual Reality Technology in the Treatment of Anxiety and Other Psychiatric Disorders: Harvard Review of Psychiatry, 25(3), 103–113. https://doi.org/10.1097/HRP.0000000000000138 Maples-Keller, J. L., & Rauch, S. A. M. (2020). Habituation. In J. S. Abramowitz & S. M. Blakey (Eds.), Clinical handbook of fear and anxiety: Maintenance processes and treatment mechanisms (pp. 249– 263). American Psychological Association. https://doi.org/10.1037/0000150-000 Maples-Keller, J. L., Yasinski, C., Manjin, N., & Rothbaum, B. O. (2017). Virtual Reality-Enhanced Extinction of Phobias and Post-Traumatic Stress. Neurotherapeutics, 14(3), 554–563. https://doi.org/10.1007/s13311-017-0534-y Marks, I. M. (1978). Living with Fear. McGraw-Hill.

104

Mataix-Cols, D., Cruz, L. F. de la, Monzani, B., Rosenfield, D., Andersson, E., Pérez-Vigil, A., Frumento, P., Kleine, R. A. de, Difede, J., Dunlop, B. W., Farrell, L. J., Geller, D., Gerardi, M., Guastella, A. J., Hof- mann, S. G., Hendriks, G.-J., Kushner, M. G., Lee, F. S., Lenze, E. J., … Thuras, P. (2017). D-Cycloserine Augmentation of Exposure- Based Cognitive Behavior Therapy for Anxiety, Obsessive-Compul- sive, and Posttraumatic Stress Disorders: A Systematic Review and Meta-analysis of Individual Participant Data. JAMA Psychiatry, 74(5), 501–510. https://doi.org/10.1001/jamapsychiatry.2016.3955 Mauri, L., & D’Agostino, R. B. (2017). Challenges in the Design and Inter- pretation of Noninferiority Trials. New England Journal of Medicine, 377(14), 1357–1367. https://doi.org/10.1056/NEJMra1510063 McCann, R. A., Armstrong, C. M., Skopp, N. A., Edwards-Stewart, A., Smo- lenski, D. J., June, J. D., Metzger-Abamukong, M., & Reger, G. M. (2014). Virtual reality exposure therapy for the treatment of anxiety disorders: An evaluation of research quality. Journal of Anxiety Dis- orders, 28(6), 625–631. https://doi.org/10.1016/j.janxdis.2014.05.010 McEwen, B. S. (2007). Physiology and Neurobiology of Stress and Adapta- tion: Central Role of the Brain. Physiological Reviews, 87(3), 873– 904. https://doi.org/10.1152/physrev.00041.2006 McMillan, D., & Lee, R. (2010). A systematic review of behavioral experiments vs. exposure alone in the treatment of anxiety disorders: A case of exposure while wearing the emperor’s new clothes? Clinical Psy- chology Review, 30(5), 467–478. https://doi.org/10.1016/j.cpr.2010.01.003 Meyerbröker, K., & Emmelkamp, P. M. G. (2010). Virtual reality exposure therapy in anxiety disorders: A systematic review of process-and-outcome studies. Depression and Anxiety, 27(10), 933–944. https://doi.org/10.1002/da.20734 Michaliszyn, D., Marchand, A., Bouchard, S., Martel, M.-O., & Poirier-Bis- son, J. (2010). A Randomized, Controlled Clinical Trial of In Virtuo and In Vivo Exposure for Spider Phobia. Cyberpsychology, Behavior, and Social Networking, 13(6), 689–695. https://doi.org/10.1089/cyber.2009.0277 Milad, M. R., Rosenbaum, B. L., & Simon, N. M. (2014). Neuroscience of fear extinction: Implications for assessment and treatment of fear- based and anxiety related disorders. Behaviour Research and Ther- apy, 62, 17–23. https://doi.org/10.1016/j.brat.2014.08.006 Miloff, A., Lindner, P., Dafgård, P., Deak, S., Garke, M., Hamilton, W., Hein- soo, J., Kristoffersson, G., Rafi, J., Sindemark, K., Sjölund, J., Zenger, M., Reuterskiöld, L., Andersson, G., & Carlbring, P. (2019). Auto- mated virtual reality exposure therapy for spider phobia vs. in-vivo one-session treatment: A randomized non-inferiority trial. Behaviour

105 Research and Therapy, 118, 130–140. https://doi.org/10.1016/j.brat.2019.04.004 Miloff, A., Lindner, P., Hamilton, W., Reuterskiöld, L., Andersson, G., & Carlbring, P. (2016). Single-session gamified virtual reality exposure therapy for spider phobia vs. traditional exposure therapy: Study protocol for a randomized controlled non-inferiority trial. Trials, 17(1), 60. https://doi.org/10.1186/s13063-016-1171-1 Minns, S., Levihn-Coon, A., Carl, E., Smits, J. A. J., Miller, W., Howard, D., Papini, S., Quiroz, S., Lee-Furman, E., Telch, M., Carlbring, P., Xan- thopoulos, D., & Powers, M. B. (2018). Immersive 3D exposure- based treatment for spider fear: A randomized controlled trial. Jour- nal of Anxiety Disorders, 58, 1–7. https://doi.org/10.1016/j.janxdis.2018.05.006 Miragall, M., Baños, R. M., Cebolla, A., & Botella, C. (2015). Working alliance inventory applied to virtual and augmented reality (WAI-VAR): Psychometrics and therapeutic outcomes. Frontiers in Psychology, 6. https://doi.org/10.3389/fpsyg.2015.01531 Morina, N., Ijntema, H., Meyerbröker, K., & Emmelkamp, P. M. G. (2015). Can virtual reality exposure therapy gains be generalized to real-life? A meta-analysis of studies applying behavioral assessments. Behav- iour Research and Therapy, 74, 18–24. https://doi.org/10.1016/j.brat.2015.08.010 Morris, B. (2017, October 11). Facebook Sets Goal of a Billion Virtual-Real- ity Users, Unveils New Headset. Wall Street Journal. https://www.wsj.com/articles/facebook-sets-goal-of-a-billion-virtual-reality-users-unveils-new-headset-1507764852 Muris, P., & Merckelbach, H. (1996). A comparison of two spider fear questionnaires. Journal of Behavior Therapy and Experimental Psychia- try, 27(3), 241–244. https://doi.org/10.1016/S0005-7916(96)00022-5 Muris, P., Schmidt, H., & Merckelbach, H. (1999). The structure of specific phobia symptoms among children and adolescents. Behaviour Re- search and Therapy, 37(9), 863–868. https://doi.org/10.1016/S0005- 7967(98)00201-0 Nakataki, M., Soravia, L. M., Schwab, S., Horn, H., Dierks, T., Strik, W., Wiest, R., Heinrichs, M., Quervain, D. J.-F. de, Federspiel, A., & Morishima, Y. (2017). Glucocorticoid Administration Improves Ab- errant Fear-Processing Networks in Spider Phobia. Neuropsycho- pharmacology, 42(2), 485–494. https://doi.org/10.1038/npp.2016.207 New, J. J., & German, T. C. (2015). Spiders at the cocktail party: An ancestral threat that surmounts inattentional blindness. Evolution and Human Behavior, 36(3), 165–173. https://doi.org/10.1016/j.evolhumbe- hav.2014.08.004

106

Ngai, I., Tully, E. C., & Anderson, P. L. (2015). The Course of the Working Alliance during Virtual Reality and Exposure Group Therapy for So- cial Anxiety Disorder. Behavioural and Cognitive Psychotherapy, 43(2), 167–181. https://doi.org/10.1017/S135246581300088X Norton, P. J., & Price, E. C. (2007). A Meta-Analytic Review of Adult Cog- nitive-Behavioral Treatment Outcome Across the Anxiety Disorders: The Journal of Nervous and Mental Disease, 195(6), 521–531. https://doi.org/10.1097/01.nmd.0000253843.70149.9a Oh, C. S., Bailenson, J. N., & Welch, G. (2018). A Systematic Review of So- cial Presence: Definition, Antecedents, and Implications. Front. Ro- botics and AI, 2018. https://doi.org/10.3389/frobt.2018.00114 Öhman, A., Flykt, A., & Esteves, F. (2001). Emotion drives attention: Detect- ing the snake in the grass. Journal of Experimental Psychology: Gen- eral, 130(3), 466–478. https://doi.org/10.1037/0096-3445.130.3.466 Öhman, A., & Mineka, S. (2001). Fears, phobias, and preparedness: Toward an evolved module of fear and fear learning. Psychological Review, 108(3), 483–522. http://dx.doi.org/10.1037/0033-295X.108.3.483 Olatunji, B. O., Armstrong, T., & Elwood, L. (2017). Is Disgust Proneness Associated With Anxiety and Related Disorders? A Qualitative Re- view and Meta-Analysis of Group Comparison and Correlational Studies: Perspectives on Psychological Science. https://doi.org/10.1177/1745691616688879 Olatunji, B. O., Deacon, B. J., & Abramowitz, J. S. (2009). The Cruelest Cure? Ethical Issues in the Implementation of Exposure-Based Treatments. Cognitive and Behavioral Practice, 16(2), 172–180. https://doi.org/10.1016/j.cbpra.2008.07.003 Ollendick, T. H., Lewis, K. M., Cowart, M. J. W., & Davis, T. (2012). Predic- tion of Child Performance on a Parent–Child Behavioral Approach Test With Animal Phobic Children. Behavior Modification, 36(4), 509–524. https://doi.org/10.1177/0145445512448191 Ollendick, T. H., Öst, L.-G., Reuterskiöld, L., Costa, N., Cederlund, R., Sirbu, C., Davis, T. E., & Jarrett, M. A. (2009). One-session treatment of specific phobias in youth: A randomized clinical trial in the United States and Sweden. Journal of Consulting and Clinical Psychology, 77(3), 504–516. https://doi.org/10.1037/a0015158 Oosterink, F. M. D., Jongh, A. D., & Hoogstraten, J. (2009). Prevalence of dental fear and phobia relative to other fear and phobia subtypes. Eu- ropean Journal of Oral Sciences, 117(2), 135–143. https://doi.org/10.1111/j.1600-0722.2008.00602.x Opriş, D., Pintea, S., García-Palacios, A., Botella, C., Szamosközi, Ş., & Da- vid, D. (2012). Virtual reality exposure therapy in anxiety disorders: A quantitative meta-analysis: Virtual Reality Exposure Therapy. De- pression and Anxiety, 29(2), 85–93. https://doi.org/10.1002/da.20910

107 Öst, L.-G. (1989). One-session treatment for specific phobias. Behaviour Re- search and Therapy, 27(1), 1–7. https://doi.org/10.1016/0005- 7967(89)90113-7 Öst, L.-G. (1992). Blood and injection phobia: Background and cognitive, physiological, and behavioral variables. Journal of Abnormal Psy- chology, 101(1), 68–74. https://doi.org/10.1037/0021-843X.101.1.68 Öst, L.-G. (1996). One-session group treatment of spider phobia. Behaviour Research and Therapy, 34(9), 707–715. https://doi.org/10.1016/0005-7967(96)00022-8 Öst, L.-G. (2012a). One-Session Treatment: Principles and Procedures with Adults. In T. E. Davis, T. H. Ollendick, & L.-G. Öst (Eds.), Intensive one-session treatment of specific phobias (pp. 59–96). Springer. Öst, L.-G. (2012b). Training Therapists in One-Session Treatment and As- sessing Their Competence. In T. E. Davis III, T. H. Ollendick, & L.- G. Öst (Eds.), Intensive One-Session Treatment of Specific Phobias (pp. 161–175). Springer. https://doi.org/10.1007/978-1-4614-3253- 1_8 Öst, L.-G., Ferebee, I., & Furmark, T. (1997). One-session group therapy of spider phobia: Direct versus indirect treatments. Behaviour Research and Therapy, 35(8), 721–732. https://doi.org/10.1016/S0005- 7967(97)00028-4 Öst, L.-G., Salkovskis, P. M., & Hellström, K. (1991). One-session therapist- directed exposure vs. Self-exposure in the treatment of spider phobia. Behavior Therapy, 22(3), 407–422. https://doi.org/10.1016/S0005- 7894(05)80374-0 Öst, L.-G., Stridh, B.-M., & Wolf, M. (1998). A clinical study of spider phobia: Prediction of outcome after self-help and therapist-directed treatments. Behaviour Research and Therapy, 36(1), 17–35. https://doi.org/10.1016/S0005-7967(97)10018-3 Page, S., & Coxon, M. (2016). Virtual Reality Exposure Therapy for Anxiety Disorders: Small Samples and No Controls? Frontiers in Psychology, 7. https://doi.org/10.3389/fpsyg.2016.00326 Parsons, T. D., & Rizzo, A. A. (2008). Affective outcomes of virtual reality exposure therapy for anxiety and specific phobias: A meta-analysis. Journal of Behavior Therapy and Experimental Psychiatry, 39(3), 250–261. https://doi.org/10.1016/j.jbtep.2007.07.007 Piaggio, G., Elbourne, D. R., Pocock, S. J., Evans, S. J. W., Altman, D. G., & CONSORT Group. (2012). Reporting of Noninferiority and Equiva- lence Randomized Trials: Extension of the CONSORT 2010 State- ment. JAMA, 308(24), 2594–2604. https://doi.org/10.1001/jama.2012.87802 Pihlaja, S., Stenberg, J.-H., Joutsenniemi, K., Mehik, H., Ritola, V., & Joffe, G. (2018). Therapeutic alliance in guided internet therapy programs for depression and anxiety disorders – A systematic review. Internet Interventions, 11, 1–10. https://doi.org/10.1016/j.invent.2017.11.005

108

Pontier, M., & Siddiqui, G. F. (2008). A Virtual Therapist That Responds Em- pathically to Your Answers. In H. Prendinger, J. Lester, & M. Ishi- zuka (Eds.), Intelligent Virtual Agents (Vol. 5208, pp. 417–425). Springer Berlin Heidelberg. https://doi.org/10.1007/978-3-540- 85483-8_42 Powers, M. B., Asmundson, G. J. G., & Smits, J. A. J. (2015). Exercise for Mood and Anxiety Disorders: The State-of-the Science. Cognitive Be- haviour Therapy, 44(4), 237–239. https://doi.org/10.1080/16506073.2015.1047286 Powers, M. B., & Emmelkamp, P. M. G. (2008). Virtual reality exposure therapy for anxiety disorders: A meta-analysis. Journal of Anxiety Disor- ders, 22(3), 561–569. https://doi.org/10.1016/j.janxdis.2007.04.006 Powers, M. B., Medina, J. L., Burns, S., Kauffman, B. Y., Monfils, M., As- mundson, G. J. G., Diamond, A., McIntyre, C., & Smits, J. A. J. (2015). Exercise Augmentation of Exposure Therapy for PTSD: Ra- tionale and Pilot Efficacy Data. Cognitive Behaviour Therapy, 44(4), 314–327. https://doi.org/10.1080/16506073.2015.1012740 Powers, M. B., Smits, J. A. J., & Telch, M. J. (2004). Disentangling the Effects of Safety-Behavior Utilization and Safety-Behavior Availability Dur- ing Exposure-Based Treatment: A Placebo-Controlled Trial. Journal of Consulting and Clinical Psychology, 72(3), 448–454. https://doi.org/10.1037/0022-006X.72.3.448 Provoost, S., Lau, H. M., Ruwaard, J., & Riper, H. (2017). Embodied Conver- sational Agents in Clinical Psychology: A Scoping Review. Journal of Medical Internet Research, 19(5), e151. https://doi.org/10.2196/jmir.6553 Rachman, S., Radomsky, A. S., & Shafran, R. (2008). Safety behaviour: A reconsideration. Behaviour Research and Therapy, 46(2), 163–173. https://doi.org/10.1016/j.brat.2007.11.008 Rapee, R. M., Sanderson, W. C., McCauley, P. A., & Di Nardo, P. A. (1992). Differences in reported symptom profile between panic disorder and other DSM-III-R anxiety disorders. Behaviour Research and Ther- apy, 30(1), 45–52. https://doi.org/10.1016/0005-7967(92)90095-X Reed, G. M., First, M. B., Kogan, C. S., Hyman, S. E., Gureje, O., Gaebel, W., Maj, M., Stein, D. J., Maercker, A., Tyrer, P., Claudino, A., Gar- ralda, E., Salvador‐Carulla, L., Ray, R., Saunders, J. B., Dua, T., Poz- nyak, V., Medina‐Mora, M. E., Pike, K. M., … Saxena, S. (2019). Innovations and changes in the ICD‐11 classification of mental, behavioural and neurodevelopmental disorders. World Psychiatry, 18(1), 3–19. https://doi.org/10.1002/wps.20611 Reid, A. M., Bolshakova, M. I., Guzick, A. G., Fernandez, A. G., Striley, C. W., Geffken, G. R., & McNamara, J. P. (2017). Common Barriers to the Dissemination of Exposure Therapy for Youth with Anxiety Dis- orders. Community Mental Health Journal, 53(4), 432–437. https://doi.org/10.1007/s10597-017-0108-9

109 Reuterskiöld, L., & Öst, L.-G. (2012). Real World Applications of One-Ses- sion Treatment. In T. E. Davis III, T. H. Ollendick, & L.-G. Öst (Eds.), Intensive One-Session Treatment of Specific Phobias (pp. 127–141). Springer. https://doi.org/10.1007/978-1-4614-3253-1_6 Romano, M., Ma, R., Moscovitch, M., & Moscovitch, D. A. (2020). Autobi- ographical memory bias. In J. S. Abramowitz & S. M. Blakey (Eds.), Clinical handbook of fear and anxiety: Maintenance processes and treatment mechanisms (pp. 183–202). American Psychological Asso- ciation. https://doi.org/10.1037/0000150-000 Rothbaum, B. O., Hodges, L. F., Kooper, R., Opdyke, D., Williford, J. S., & North, M. (1995). Effectiveness of computer-generated (virtual reality) graded exposure in the treatment of acrophobia. The American Journal of Psychiatry, 152(4), 626–628. https://doi.org/10.1176/ajp.152.4.626 Rothbaum, B. O., Hodges, L. F., Kooper, R., Opdyke, D., Williford, J. S., & North, M. (1995). Virtual reality graded exposure in the treatment of acrophobia: A case report. Behavior Therapy, 26(3), 547–554. https://doi.org/10.1016/S0005-7894(05)80100-5 Rothbaum, B.O., Hodges, L., Watson, B. A., Kessler, G. D., & Opdyke, D. (1996). Virtual reality exposure therapy in the treatment of fear of flying: A case report. Behaviour Research and Therapy, 34(5), 477–481. https://doi.org/10.1016/0005-7967(96)00007-1 Rozental, A., Kottorp, A., Boettcher, J., Andersson, G., & Carlbring, P. (2016). Negative Effects of Psychological Treatments: An Explora- tory Factor Analysis of the Negative Effects Questionnaire for Moni- toring and Reporting Adverse and Unwanted Events. PLOS ONE, 11(6), e0157503. https://doi.org/10.1371/journal.pone.0157503 Sarlo, M., Palomba, D., Angrilli, A., & Stegagno, L. (2002). Blood phobia and spider phobia: Two specific phobias with different autonomic cardiac modulations. Biological Psychology, 60(2), 91–108. https://doi.org/10.1016/S0301-0511(02)00030-3 Scarfe, P., & Glennerster, A. (2019). The Science Behind Virtual Reality Dis- plays. Annual Review of Vision Science, 5(1), 529–547. https://doi.org/10.1146/annurev-vision-091718-014942 Sewart, A. R., & Craske, M. G. (2020). Inhibitory learning. In J. S. Abramowitz & S. M. Blakey (Eds.), Clinical handbook of fear and anxiety: Maintenance processes and treatment mechanisms (pp. 265– 285). American Psychological Association. https://doi.org/10.1037/0000150-000 Shiban, Y., Pauli, P., & Mühlberger, A. (2013). Effect of multiple context exposure on renewal in spider phobia. Behaviour Research and Ther- apy, 51(2), 68–74. https://doi.org/10.1016/j.brat.2012.10.007

110

Shiban, Y., Schelhorn, I., Pauli, P., & Mühlberger, A. (2015). Effect of combined multiple contexts and multiple stimuli exposure in spider phobia: A randomized clinical trial in virtual reality. Behaviour Research and Therapy, 71, 45–53. https://doi.org/10.1016/j.brat.2015.05.014 Shimada‐Sugimoto, M., Otowa, T., & Hettema, J. M. (2015). Genetics of anxiety disorders: Genetic epidemiological and molecular studies in humans. Psychiatry and Clinical Neurosciences, 69(7), 388–401. https://doi.org/10.1111/pcn.12291 SIGGRAPH. (2018, August 14). VR @ 50: Ivan Sutherland’s 1968 Head- Mounted 3D Display System. ACM SIGGRAPH Blog. https://blog.siggraph.org/2018/08/vr-at-50-celebrating-ivan-sutherland.html/ Silverman, D. (1992, June 7). From the archives: NASA explores virtual reality in 1992. HoustonChronicle.Com. https://www.houstonchronicle.com/techburger/article/From-the-archives-Virtual-Reality- 12709811.php Slater, M. (2009). Place illusion and plausibility can lead to realistic behaviour in immersive virtual environments. Philosophical Transactions of the Royal Society B: Biological Sciences, 364(1535), 3549–3557. https://doi.org/10.1098/rstb.2009.0138 Slater, M. (2018). Immersion and the illusion of presence in virtual reality. British Journal of Psychology, 109(3), 431–433. https://doi.org/10.1111/bjop.12305 Slater, M., & Sanchez-Vives, M. V. (2016). Enhancing Our Lives with Im- mersive Virtual Reality. Frontiers in Robotics and AI, 3. https://doi.org/10.3389/frobt.2016.00074 Soares, S. C., Lindström, B., Esteves, F., & Öhman, A. (2014). The Hidden Snake in the Grass: Superior Detection of Snakes in Challenging At- tentional Conditions. PLOS ONE, 9(12), e114724. https://doi.org/10.1371/journal.pone.0114724 Soravia, L. M., Heinrichs, M., Winzeler, L., Fisler, M., Schmitt, W., Horn, H., Dierks, T., Strik, W., Hofmann, S. G., & de Quervain, D. J.-F. (2014). Glucocorticoids enhance in vivo exposure-based therapy of spider phobia. Depression and Anxiety, 31(5), 429–435. https://doi.org/10.1002/da.22219 Statistica. (2019, November 11). Number of smartphone users worldwide 2014-2020. Statista. https://www.statista.com/statistics/330695/number-of-smartphone-users-worldwide/ Stein, M. B., Simmons, A. N., Feinstein, J. S., & Paulus, M. P. (2007). In- creased Amygdala and Insula Activation During Emotion Processing in Anxiety-Prone Subjects. American Journal of Psychiatry, 164(2), 318–327. https://doi.org/10.1176/ajp.2007.164.2.318 Stinson, F. S., Dawson, D. A., Patricia Chou, S., Smith, S., Goldstein, R. B., June Ruan, W., & Grant, B. F. (2007). The epidemiology of DSM-IV specific phobia in the USA: Results from the National Epidemiologic

111 Survey on Alcohol and Related Conditions. Psychological Medicine, 37(7), 1047–1059. https://doi.org/10.1017/S0033291707000086 Sucala, M., Schnur, J. B., Constantino, M. J., Miller, S. J., Brackman, E. H., & Montgomery, G. H. (2012). The Therapeutic Relationship in E- Therapy for Mental Health: A Systematic Review. Journal of Medical Internet Research, 14(4), e110. https://doi.org/10.2196/jmir.2084 Sutherland, I. E. (1968). A Head-mounted Three Dimensional Display. Pro- ceedings of the December 9-11, 1968, Fall Joint Computer Confer- ence, Part I, 757–764. https://doi.org/10.1145/1476589.1476686 Szymanski, J., & O’Donohue, W. (1995a). Fear of spiders questionnaire. Journal of Behavior Therapy and Experimental Psychiatry, 26(1), 31–34. https://doi.org/10.1016/0005-7916(94)00072-T Szymanski, J., & O’Donohue, W. (1995b). The potential role of state-dependent learning in cognitive therapy with spider phobics. Journal of Ra- tional-Emotive and Cognitive-Behavior Therapy, 13(2), 131–150. https://doi.org/10.1007/BF02354458 Tanner, M. K., Hake, H. S., Bouchet, C. A., & Greenwood, B. N. (2018). Running from fear: Exercise modulation of fear extinction. Neurobi- ology of Learning and Memory, 151, 28–34. https://doi.org/10.1016/j.nlm.2018.03.021 The European Agency for the Evaluation of Medicinal Products. (2008). Points to consider on switching between superiority and non-inferiority: Points to consider on switching between superiority and non-inferiority. British Journal of Clinical Pharmacology, 52(3), 223–228. https://doi.org/10.1046/j.1365-2125.2001.01397-3.x Thorpe, S. J., & Salkovskis, P. M. (1999). Animals phobias. In G. C. L. Davey (Ed.), Phobias: A Handbook of Theory, Research and Treatment (p. 26). Wiley. Thyer, B. A., & Himle, J. (1987). Phobic anxiety and panic anxiety: How do they differ? Journal of Anxiety Disorders, 1(1), 59–67. https://doi.org/10.1016/0887-6185(87)90023-5 Vachon, D. D., Krueger, R. F., Rogosch, F. A., & Cicchetti, D. (2015). As- sessment of the Harmful Psychiatric and Behavioral Effects of Differ- ent Forms of Child Maltreatment. JAMA Psychiatry, 72(11), 1135– 1142. https://doi.org/10.1001/jamapsychiatry.2015.1792 Valmaggia, L. R., Latif, L., Kempton, M. J., & Rus-Calafell, M. (2016). Vir- tual reality in the psychological treatment for mental health problems: An systematic review of recent evidence. Psychiatry Research, 236, 189–195. https://doi.org/10.1016/j.psychres.2016.01.015 Van Houtem, C. M. H. H., Laine, M. L., Boomsma, D. I., Ligthart, L., van Wijk, A. J., & De Jongh, A. (2013). A review and meta-analysis of the heritability of specific phobia subtypes and corresponding fears. Journal of Anxiety Disorders, 27(4), 379–388. https://doi.org/10.1016/j.janxdis.2013.04.007

112

Ventis, W. L., Higbee, G., & Murdock, S. A. (2001). Using Humor in Sys- tematic Desensitization to Reduce Fear. The Journal of General Psy- chology, 128(2), 241–253. https://doi.org/10.1080/00221300109598911 Vervliet, B., Craske, M. G., & Hermans, D. (2013). Fear Extinction and Re- lapse: State of the Art. Annual Review of Clinical Psychology, 9(1), 215–248. https://doi.org/10.1146/annurev-clinpsy-050212-185542 Wardenaar, K. J., Lim, C. C. W., Al-Hamzawi, A. O., Alonso, J., Andrade, L. H., Benjet, C., Bunting, B., de Girolamo, G., Demyttenaere, K., Flo- rescu, S. E., Gureje, O., Hisateru, T., Hu, C., Huang, Y., Karam, E., Kiejna, A., Lepine, J. P., Navarro-Mateu, F., Oakley Browne, M., … de Jonge, P. (2017). The cross-national epidemiology of specific phobia in the World Mental Health Surveys. Psychological Medicine, 47(10), 1744–1760. https://doi.org/10.1017/S0033291717000174 Watson, D. R., Garfinkel, S. N., Gould van Praag, C., Willmott, D., Wong, K., Meeten, F., & Critchley, H. D. (2019). Computerized Exposure Ther- apy for Spider Phobia: Effects of Cardiac Timing and Interoceptive Ability on Subjective and Behavioral Outcomes. Psychosomatic Med- icine, 81(1), 90. https://doi.org/10.1097/PSY.0000000000000646 Watson, J. B. (1913). Psychology as the behaviorist views it. Psychological Review, 20(2), 158–177. https://doi.org/10.1037/h0074428 Wechsler, T. F., Kümpers, F., & Mühlberger, A. (2019). Inferiority or Even Superiority of Virtual Reality Exposure Therapy in Phobias?—A Sys- tematic Review and Quantitative Meta-Analysis on Randomized Con- trolled Trials Specifically Comparing the Efficacy of Virtual Reality Exposure to Gold Standard in vivo Exposure in Agoraphobia, Specific Phobia, and Social Phobia. Frontiers in Psychology, 10. https://doi.org/10.3389/fpsyg.2019.01758 Weghorst, S. J., Sieburg, H. B., & Morgan, K. S. (1996). Health Care in the Information Age, Medicine Meets Virtual Reality: 4. Conference Edi- tion. IOS Press / Omsha. Weisman, J. S., & Rodebaugh, T. L. (2018). Exposure therapy augmentation: A review and extension of techniques informed by an inhibitory learning approach. Clinical Psychology Review, 59, 41–51. https://doi.org/10.1016/j.cpr.2017.10.010 Wiens, B. L. (2002). Choosing an equivalence limit for noninferiority or equivalence studies. Controlled Clinical Trials, 23(1), 2–14. https://doi.org/10.1016/S0197-2456(01)00196-9 Williams, N. L., Connolly, K. M., Cisler, J. M., Elwood, L. S., Willems, J. L., & Lohr, J. M. (2009). Disgust: A cognitive approach. In Disgust and its disorders: Theory, assessment, and treatment implications. (pp. 57–73, Chapter xvii, 324 Pages). American Psychological Associa- tion (Washington, DC, US). https://search.proquest.com/docview/614169851/abstract/48E5171838B74163PQ/1

113 Wittchen, H.-U., Nelson, C. B., & Lachner, G. (1998). Prevalence of mental disorders and psychosocial impairments in adolescents and young adults. Psychological Medicine, 28(1), 109–126. https://doi.org/10.1017/S0033291797005928 Wolitzky-Taylor, K. B., Horowitz, J. D., Powers, M. B., & Telch, M. J. (2008). Psychological approaches in the treatment of specific phobias: A meta-analysis. Clinical Psychology Review, 28(6), 1021–1037. https://doi.org/10.1016/j.cpr.2008.02.007 Wolpe, J., & Plaud, J. J. (1997). Pavlov’s contributions to behavior therapy: The obvious and the not so obvious. American Psychologist, 52(9), 966–972. http://dx.doi.org/10.1037/0003-066X.52.9.966 World Health Organization. (2018). International statistical classification of diseases and related health problems (11th Revision). https://icd.who.int/browse11/l-m/en Wrzesien, M., Bretón-López, J., Botella, C., Burkhardt, J.-M., Alcañiz, M., Pérez-Ara, M. Á., & del Amo, A. R. (2013). How Technology Influ- ences the Therapeutic Process: Evaluation of the Patient-Therapist Relationship in Augmented Reality Exposure Therapy and In Vivo Exposure Therapy. Behavioural and Cognitive Psychotherapy, 41(4), 505–509. https://doi.org/10.1017/S1352465813000088 Yoshikawa, H., Aber, J. L., & Beardslee, W. R. (2012). The effects of poverty on the mental, emotional, and behavioral health of children and youth: Implications for prevention. The American Psychologist, 67(4), 272– 284. https://doi.org/10.1037/a0028015 Zilcha-Mano, S., Roose, S. P., Brown, P. J., & Rutherford, B. R. (2019). Not Just Nonspecific Factors: The Roles of Alliance and Expectancy in Treatment, and Their Neurobiological Underpinnings. Frontiers in Behavioral Neuroscience, 12. https://doi.org/10.3389/fnbeh.2018.00293 Zlomke, K., & Davis, T. E. (2008). One-Session Treatment of Specific Pho- bias: A Detailed Description and Review of Treatment Efficacy. Be- havior Therapy, 39(3), 207–223. https://doi.org/10.1016/j.beth.2007.07.003

i This article was published in Behaviour Research and Therapy, Miloff et al., Auto- mated virtual reality exposure therapy for spider phobia vs. in-vivo one-session treatment: A randomized non-inferiority trial, 118, 130-140, Copyright Elsevier (2019). ii This article was published in Scandinavian Journal of Psychology, Lindner et al., What is so frightening about spiders? Self-rated and self-disclosed impact of different characteristics and associations with phobia symptoms, 60, 1-6, Copyright Wiley (2019).

114