Investigation of extinction learning in multi-context paradigm

Размещено на http://www.allbest.ru

Размещено на http://www.allbest.ru

Introduction

Many problems of our society connected with the fact that human brain has evolved in conditions which were vastly different from our modern reality. We have a tendency to overeat because food always had been strongly limited; we have a tendency to believe in pseudoscience because the brain is predisposed to make conclusions based on very limited information, which was the best possible way before the invention of the scientific method; we also have a tendency to quickly acquire specific phobias because in a dangerous prehistorical world, strong fear of spiders or snakes was very beneficial for surviving.

We carry this dark heritage of the biological evolution throughout our comfortable and secure existence. Lifetime prevalence of specific phobias is about 4.9%, although precise numbers vary across studies (Somers, Goldner, Waraich & Hsu, 2006). One of the most common fears is arachnophobia. Even in Sweden, where all local spiders are completely non-dangerous, 1,2% of males and 5,6% of females are suffering from spider phobia (Fredrikson, Annas, Fisher & Wik, 1995). It is suggested that humans can be predisposed to quickly learn the fear of spiders. In adult participants, images of spiders work as more effective conditioned stimuli in fear acquisition studies with resepct to neutral pictures (Ohman & Mineka, 2010). Moreover, even 3-year-old children detect spiders faster than neutral stimuli in visual search tasks (LoBue, 2010).

Specific phobias have relatively small but significant negative effect on the quality of life (Cramer, Torgersen & Kringlen, 2005), and it is important to develop therapeutic approaches which can help people to attenuate their fears. The most widely used paradigm to treat acquired fears is the fear extinction learning. This method suggests that animal or human interacts with the fearful stimuli without any negative reinforcement, and this process gradually leads to the attenuation of the previously learned fear (for a review, see Myers & Davis, 2007). The clinical implication of the fear extinction paradigm is called exposure therapy. In a meta-analysis of 33 randomized studies (Wolitzky-Taylor, Horowitz, Powers & Telch, 2008), it has been demonstrated that exposure therapy provides significantly better results than non-exposure approaches in both post-treatment and follow-up estimations of fear (Cohen's d were 0,44 and 0,35 respectively).

Normally, exposure therapy of specific phobia suggests that participant interacts with the real fearful object, such as a living spider, or experience realistic stimuli in a virtual reality environment (Choy, Fyer & Lipsitz, 2007) which lead to more robust fear extinction than movies or pictorial stimuli. However, real or realistic stimuli also have some important disadvantages. First, interaction with fearful stimuli is stressful, so that some people may refuse such treatment even if they know that it can help. Second, it is not always possible to provide a realistic environment for patients with rare and unusual phobias. Third, such stimuli do not always fit the scientific research purposes (e.g. fear extinction) because they are intrinsically less standardized than videotapes or pictures of fearful objects. Hence, it is important to develop controlled scientific methods which can increase the efficacy of movies and pictures in exposure therapy in cases where the disadvantages of in vivo exposure and virtual reality are more important than their benefits.

One of the most remarkable problems with the exposure therapy, especially with videotapes and images, is the lack of generalization. Fear extinction learning is context-dependent (Bouton, 2004), which means that participants can overcome their fear related to particular stimuli but still be afraid of any other similar stimuli when they are shown in different contexts. The promising way to overcome that problem is a multi-context paradigm. It has been already demonstrated by Vansteenwegen et al. (2007) that presentation of videotaped spiders in several contexts lead to more effective fear extinction than presentation in a single context. In the current work we proposed to test a similar hypothesis by using only pictures.

The core part of the current project is the comparison of two different groups of spider-phobic participants: a multiple contexts group has been exposed to a set of pictures with spiders located in different backgrounds, while a single context group has been exposed to a set of pictures with spiders located in a single background. After the exposure, all participants went through the renewal stage, during which they were observing a completely new set of pictures with spiders in different contexts. The hypothesis suggests that the decrease of fear should be higher in a multiple contexts group. The fear has been estimated by the battery of tests, including questionnaires, behavioral approach test and registration of skin galvanic response. We have not obtained statistically significant results for behavioral measurements, but the difference in electrodermal activity between groups has demonstrated a consistency with the initial hypothesis. This distinction between overt and covert fear measurements can be seminal for the planned next stage of the research project, where it is suggested to apply transcranial magnetic stimulation to explore the role of ventromedial prefrontal cortex in fear extinction learning.

1. Literature review

phobia psychological exposure

1.1 Fear acquisition and extinction

General principles of learning.

Any kind of fear processing, whether it pertains to acquisition or extinction, is possible because the brain is generally predisposed to acquire and update knowledge about cause-to-effect relationships in the surrounding world, which is crucial for the survival. That is why the literature review starts from a brief summary of main scientific breakthroughs connected with the general understanding of neural principles which underlie any form of learning.

The first systematic study of the ability to form associations between different external events had been conducted by Ivan Pavlov. In his famous experiments with salivating dogs, Pavlov demonstrated that conditioned stimulus (e.g. sound) which precedes the unconditioned stimulus (e.g. food) for several times allows an animal to form an association, or conditioned reflex, between the sound and the food, which is manifested as increased salivation (Pavlov, 1927). Moreover, if the conditioned stimulus repeats but does not accompanied by unconditioned stimulus anymore, the level of salivation gradually decreases. Pavlov called this phenomenon “extinction” (or, in Russian, “затухание”), and this term is widely used to the present day.

In the middle of the XX century, neuroscientists concentrated their efforts on the understanding of neural mechanisms which underlie the process of making associations. In 1949, Donald Hebb published his influential book “The organization of behavior”. Based on very limited experimental data available to that moment, he suggests that learning needs both transient reverberatory traces and permanent structural changes which are potentially able to stabilize the former, and proposes a hypothetical mechanism of connection between these two stages: “When an axon of cell A is near enough to excite a cell B and repeatedly or persistently takes part in firing it, some growth process or metabolic change takes place in one or both cells such that A's efficiency, as one of the cells firing B, is increased” (Hebb, 1949).

This assumption was amazingly coherent with later discoveries. The most important contribution to the research of neural mechanisms of memory and learning was made by Eric Kandel. History of his discoveries can be briefly summarized in a sentence “How the lucky choice of a model object leads to breakthroughs findings and a Nobel prize”. Kandel and his colleagues concentrated their efforts on studying of the gill-withdrawal reflex in aplysia. It is an inborn behavior, but the intensity of the reflex can be modified by the individual experience both for a short or a long time. Moreover, the gill-withdrawal reflex is maintained by a limited number of particular neurons, which can be identified and studied separately both in the whole aplysia or in a cell culture. This model allowed to describe with a high precision many biochemical mechanisms underlying short- and long-term memory formation (for a review, see Kandel, Dudai & Mayford, 2014). The key fundamental discoveries were the following:

a) Short-term memory is based on the selective facilitation of synapses involving biochemical and functional changes in both presynaptic and postsynaptic neurons (for an example of the experimental study on aplysia, see Hawkins, Abrams, Carew & Kandel, 1983);

b) Long-term memory formation requires protein synthesis and connected with the growth of the new synaptic connections between neurons involved to particular behavioral reaction (for an example of the experimental study on aplysia, see Schacher, Castellucci & Kandel, 1988).

These core principles can be applied both to invertebrates and vertebrates (Kandel et al, 2014) and shed light on any type of learning, including fear extinction. In general, a formation of the stable memory trace requires several repetitions of the combination of stimuli and takes time needed for the synthesis of the new proteins and establishing of the newly facilitated neuronal networks. The specific manifestations of these requirements in the application to fear extinction learning and the particular molecular pathways crucial for the fear extinction will be considered in the next chapters.

The one molecular complex which is especially important for Hebb-type synaptic plasticity - and hence, for the associative learning - is the N-methyl-D-aspartate (NMDA) receptor. The unique character of NMDA receptor connected with the fact that it has two essential requirements for the activation: first, the presence of excitatory amino acids (primarily glutamate) released from a presynaptic neuron, and, second, the depolarization of the postsynaptic membrane (Cotman, Monaghan & Ganong, 1988). Only if both conditions are accomplished, the ion channel opens and transmit Ca2+ into the cell. The latter works as an intracellular messenger and initiate molecular cascades which leads to long-term potentiation of the synapse and in some cases to the formation of new synapses. Thus, NMDA receptor plays a key role in the coincident in pre- and postsynaptic activity. Importantly, postsynaptic depolarization could be connected not only with the signals from the same presynaptic neuron which provides glutamate but also from activation of some additional nearby neurons. So, the properties of NMDA receptor explain why the nervous system is constructionally predisposed to making associations between concurrent events, which is crucially important for the fear acquisition but also may be useful during the fear extinction.

The concept of coincidence might be applied to the description of memory at different levels. In a recent review, Kukushkin and Carew (2017) suggest that each particular homeostatic disturbance in the neural cell can be described in terms of time, and the memory can be considered as a nested hierarchical interaction between time windows of different lifespans describing electrophysiological, biochemical and even cytoarchitectonic changes in the neural circuits. Hence, on the level of particular cell the external experience (coded as temporal pattern of signals received from the neighborhood cells) initiates homeostatic disturbances (from short-lasting, such as the opening of ion channels, to long-lasting, such as changes in gene expression) which, in turn, influence the processing of next signals. And, according to authors, this combination of temporally overlapping homeostatic disturbances is not something that just leads to the memory - it is the memory itself. This point of view is seminal for the considering the problem of fear extinction because it underlines that the memory is a fluid and dynamic state of the nervous system which can be drastically influenced by the processing of additional external signals.

Fear extinction as a specific form of learning.

Although many human studies, including the presented project, are concentrating on the treatment of pre-existed fear (due to the clinical significance of that problem), the most rigorous empirical data obtained from studies of the extinction of purposefully conditioned fear because this paradigm provides better standardization and allows more accurate comparison of fear and its extinction between subjects. Typically, fear acquisition training implies that animal or human participants are exposed to a neutral conditioned stimulus (CS), such as visual cue or auditory tone, which is paired with an aversive unconditioned stimulus (US), such as electric shock or highly unpleasant sound. After several repetitions, participants start to demonstrate behavioral, vegetative and hormonal fear reactions while the CS alone is presented. On the next stage, which is called extinction training, several presentations of the CS in absence of aversive US leads to attenuation of a fear response. This basic procedure allows many modifications (for a methodological review, see Lonsdorf et al, 2017), and this kind of experiments provides a reliable base for theoretical conclusions about the nature of fear extinction learning.

The central question is whether the fear extinction should be considered as a simple forgetting of the fear, or it is a specific form of learning. Both points of view coexisted in the literature in the middle of the XX century (for a historical perspective, see Delamater & Westbrook, 2014), but as far as experimental protocols become more sophisticated and analysis of results becomes more scrutinous, researchers are inclined to the second interpretation, commonly called “inhibitory learning model” (Craske, Treanor, Conway, Zbozinek & Vervliet, 2014) which is predominant nowadays. Main arguments in support of the statement that fear extinction is a kind of learning have been summarized by Myers and Davis (2007). First, without the purposeful extinction training, fear memories can be very stable and persist for months or years. Second, extinction is cue specific: if the aversive US has been associated with several neutral CS, then extinction learning with the particular CS leads to the attenuation of the fear associated with this CS, but there is no generalization to other learned stimuli. Third, extinction is usually not permanent. The association between CS and US can be recovered after the presentation of US alone (fear reinstatement), during the test in a new context (fear renewal), or just spontaneously return with the passage of time.

If extinction represents a specific form of learning, then the next question arises: what exactly should be learned? Delamater and Westbrook (2014) discuss two possible models. First, the extinction memory trace can be represented in the brain as an inhibitory association between CS and US. Second, it can be an excitatory connection between CS and “no US” memory. In XX century, these explanations were often perceived as competitive, which was compounded by the fact that it is practically impossible to prove one of them and reject another based purely on the behavioral data. Today, it is suggested that both explanations can be simultaneously true. Herry et al (2008) used single-unit recordings to analyze the activity of basal nuclei of the amygdala and have identified two distinct neural circuits participating in the fear extinction in mice. The first circuit, “fear neurons” were activated by CS during fear acquisition but demonstrated CS-evoked inhibition after the extinction. The second circuit, “extinction neurons”, did not demonstrate any specific activity during fear conditioning, but were activated by CS after the extinction training.

Myers and Davis (2007) also underline that fear extinction may be based on both associative and non-associative mechanisms. The examples of the latter are a modulation of the level of CS representation (which, in turn, connected with lower attention to the CS which is not informative anymore), and response fatigue which accumulates during the extinction. Non-associative mechanisms itself definitely do not play a key role in extinction learning (for example, they do not explain why processes of fear acquisition and extinction are largely context specific), but what is clear about fear extinction is that the brain is a complex system and many different learning (and forgetting) mechanisms can be simultaneously involved in a development of the same behavioral reaction.

There are several interconnected psychological theories devoted to the explanation of fear extinction and especially its clinical application, exposure therapy, in terms of cognitive and emotional processing of the newly acquired information (Telch, Cobb & Lancaster, 2014). For example, self-efficacy theory suggests that during exposure a participant gradually develops a feeling that he or she has acquired effective coping strategies and that shift in one's self-perception is the reason why they become successful in interaction with fearful stimuli. Expectancy theories concentrated on the fact that human fear is actually concentrated not on the stimulus itself but on imaginary consequences from interaction with this stimulus, so the main goal of exposure therapy is just proving that there are no such consequences in a real world. Emotional processing theories provide a bit clearer link to the biology because consider phobic events as a false triggering of avoidance response. The problem of these purely psychological doctrines is that it is difficult to confirm one of them and reject another based on experimental studies even when researchers specifically concentrated on this goal (Rupp, Doeber, Ehring & Vossbeck-Elsebush, 2017). That is why authors of experimental papers generally do not mention any specific psychological theoretical foundations, they just study fear extinction and brain processes underlying this phenomenon.

Neuroscience of fear acquisition and extinction.

Neuroscience of fear acquisition is relatively better characterized than neuroscience of fear extinction, although both fields are extensively studied in rodents and humans (Graham & Milad, 2011; Milad & Quirk, 2012). It is widely accepted that the main neural structure for the integration of information about the conditioned and unconditioned fearful stimuli is the basolateral complex of the amygdala. Disruption of its function through lesions or other experimental treatments jeopardizes the ability to acquire conditioned fear in rodents. The second key region for the fear acquisition in rodents is the prelimbic (PL) cortex. Level of activation of this area positively correlates with conditioned fearful behavioral responses. PL is considered as a part of the medial prefrontal cortex in rodents, but the commonly accepted functional analogue of PL in humans is the dorsal anterior cingulate cortex (dACC) which activation is increased during fear conditioning. There are also some data about the positive correlation between dACC thickness and skin conductance response during fear conditioning (Milad et al, 2007), although they had not been reproduced in a later study (Hartley, Fischl & Phelps, 2011). There are some other regions involved in fear response in humans, such as the insula, thalamus and periaqueductal gray matter of the midbrain.

Fear extinction also involves the basolateral complex of the amygdala. In rodents, two other important areas are the infralimbic (IL) region of the medial prefrontal cortex and the hippocampus (Milad, Rosenbaum & Simon, 2014). It is assumed that hippocampus integrates the information about an extinguished cue and a context in which the cue is presented. In case when the context is familiar (the same with the context of the extinction training) the hippocampus activates the infralimbic cortex, which, in turn, communicates with the inhibitory interneurons in the basolateral amygdala, which leads to suppression of the fear response. On the contrary, the presence of the cue in the unfamiliar context does not lead to the activation of this circuit, and fear response arises again.

Rodents research allow using a wide spectrum of modern experimental approaches to explore neural correlates of extinction learning. Researchers widely use microinfusions of pharmacological substances in the regions of interest, which is especially important to stimulate or impair processes connected with neuroplasticity. For example, it has been demonstrated that microinfusion of anisomycin, which inhibits the protein synthesis, into the medial prefrontal cortex before extinction training does not disrupt the fear extinction itself but leads to an inability to recall extinction during the following-day testing (Santini, Ge, Ren, Peсa de Ortiz & Quirk, 2004). The important role of neuroplasticity in extinction learning is also confirmed by experiments dedicated to influencing of the brain-derived neurotrophic factor (BDNF). It has been demonstrated that infusion of BDNF in infralimbic cortex induces fear extinction; administering of anti-BDNF antibodies into IL leads to slower fear extinction; immunocytochemistry demonstrates that extinction learning leads to increase of BDNF in ventral hippocampus; and infusion of BDNF into ventral hippocampus decreased freezing behavior in fear-conditioned rats (Rosas-Vidal, Do-Monte, Sotres-Bayon & Quirk, 2014).

A massive amount of detailed information about microcircuits involved in fear extinction in rodents is obtained by using implanted electrodes or optogenetics approaches (for the review, see Tovote, Fadok & Lьthi, 2015). As it has been already mentioned, microelectrode technique allowed to describe in basal amygdala two groups of neurons, associated with fear and with fear extinction, and analyze the patterns of their activation during switching behavior from fear extinguished to fear renewal state (Herry et al, 2008). The optogenetic approach helped to demonstrate that vmPFC-amygdala circuit in mice facilitates long-term extinction memory (Bukalo et al, 2015). Senn et al. (2014) used a set of optogenetic, immunohystochemical and behavioral methods for the detailed characterization of two distinct populations of neurons in basal amygdala which send projections to IL and PL subdivisions of a medial prefrontal cortex and demonstrated that the balance of their activity regulates fear extinction.

Researchers who work with humans are much more limited in the spectrum of available approaches. However, neuroimaging methods allow confirming that neural circuits involved in fear extinction in humans show high similarity with those ones of rodents. Ventromedial prefrontal cortex (vmPFC), the human homolog of the IL, activates during extinction recall, and the level of activation positively correlates with the inhibition of the reaction to conditioned stimuli. Moreover, there is a correlation between extinction recall and the thickness of the ventromedial prefrontal cortex (Milad & Quirk, 2012).

In addition to the basolateral complex of the amygdala, hippocampus, and vmPFC, several other regions are involved in fear extinction in humans (Greco & Liberzon, 2016). When a subject expects the aversive stimulus which actually does not occurred after the conditioned stimulus, he or she experienced a prediction error. This phenomenon is associated with the activation of the striatum which is involved in the estimation of probabilities of possible negative events. The absence of expected unconditioned stimulus is also associated with the activation in the anterior cingulate cortex, parietal cortex, and insula. It has been also demonstrated that fear extinction can be associated with the activation of dorsolateral prefrontal cortex, but only for those participants who had been previously trained to expect detention between conditioned and unconditioned stimuli, so presumably it happens because of the involving of short-term memory to estimate trace intervals during which participants are still expecting to receive the unconditioned stimulus after the presentation of the conditioned one (Ewald H. et al, 2014).

The problem of fear recovery.

As it has been mentioned in previous paragraphs, fear extinction should be considered as a specific form of learning, and, according to the experiments with rodents, even after completion of fear extinction training brain still maintains neural circuits associated with the initial fear. Under some conditions, they can be reactivated. Indeed, relapse of the fear is common in fear extinction studies. Bouton (2004) describes four different types of fear return: the renewal effect, spontaneous recovery, rapid reacquisition, and reinstatement. The renewal effect is observing when a subject (non-human animal or human being) gone through extinction learning in one context and then meet conditioned stimulus in another context. Spontaneous recovery means that fear might return in the same physical context but after some length of time after the extinction. Rapid reacquisition is initiated by the new presentation of a combination of conditioned and unconditioned stimulus. Reinstatement is observed if the subject exposed to the unconditioned stimulus alone.

There is a widely used terminology which describes the sequence of contexts by different characters with regard to the renewal effect. “ABA renewal” means that fear acquisition happened in a context A, then fear extinction learned in context B, and then, after returning to a context A, participant again will demonstrate fear reactions. “ABC renewal” assumes acquisition in the first context, extinction in the second, and testing in the third new context, which still leads to observed fear. The third version, “AAB renewal”, means that fear acquisition and extinction performed in the same context, and unfamiliar context B provokes fear response.

The renewal effect is ubiquitous and nearly inevitable. It arises in the huge number of different paradigms and even after very intense extinction training. Interestingly, the word “context” can characterize not only an environment but also internal state of the animal or human being. An astonishing study performed by Mystkowski, Mineka, Vernon and Zinbarg (2003) demonstrated that fear renewal can be influenced even by shifts in caffeine state. In this study, spider-fearful participants received either caffeine or placebo during exposure therapy and follow-up behavioral approach test where the final task was to touch the tarantula. Those who received the same drink (with or without caffeine) during both stages of the experiment demonstrated less fear renewal than those who received different drinks during exposure and follow-up behavioral approach test. The phenomenon of the fear renewal clearly emphasizes the huge role of context in the extinction learning which has important clinical implications.

The renewal effect has been demonstrated in humans. Vansteenwegen et al (2005) used two pictorial faces as conditioned stimuli and aversive noise as an unconditioned stimulus. Participants have been presented to both pictures, but only one was accompanied with the noise. When participants learn to associate one of the pictures with the aversive stimulus, they went through the extinction stage where both pictures were demonstrated without the negative reinforcement. It was the same room and generally same conditions with only one detail: half of the participants receive acquisition and extinction learning while the central light in the room was switched off, another half participate in extinction when the light was turned on. It was enough to find a difference between groups during the test, which happens without the central light. ABA-renewal group demonstrated clear conditioned responding, while AAA-renewal group does not.

Recent research shed some light on the neuroscience of fear renewal in humans. Hermann, Stark, Milad, and Merz (2016) performed the whole procedure of fear acquisition, extinction and renewal inside the MRI scanner. Conditioned stimuli were set as photographs of different rooms while electrical stimulation serves as an unconditioned stimulus. The context was modified through the different color of lamps, presented on the photographs. Extended ABC renewal paradigm was used: acquisition happens in a context A, extinction in context B, and second-day renewal in both previous contexts and in the new context C. Indeed, during the testing day participants demonstrate the lowest level of fear (measured predominantly through skin conductance response) in the extinction context and higher fear in acquisition and novel contexts. Research confirmed the key role of the hippocampus in the estimation of familiarity of context. Notably, there was an asymmetry: extinction context B was associated with the stronger activation in the left hippocampus, while novel context C was accompanied by stronger activation in the right hippocampus. Context C also leads to enhanced activation in insula, vmPFC, dACC and the right occipital cortex. In acquisition context A, there were notable differences between high renewal and low renewal groups: the first one demonstrated higher activation of the right hippocampus and right amygdala. These findings provide evidence for the hypothesis concerning the existence of hippocampal hemispheric specialization, which supposed that preferentially the left hippocampus involved in the maintenance of memory about safe contexts, while the right hippocampus more connected with the fear acquisition contexts.

It can be clearly assumed that the issue of the context is critically important for the fear renewal and for the prevention of fear renewal. The context-related behavioral experiments, as well as their importance for improving the efficacy of exposure therapy for the clinical purposes, will be considered in the chapter, specifically dedicated to this problem. But before this, it is important to clarify some methodological issues connected with the operationalization of fear.

Methodological issues: how to measure fear

Experimental psychology constantly has to deal with the problem of the accurate operationalization of its concepts. It is generally impossible to estimate the internal state directly - it is inferred from the set of indirect measurements, which, as it is suggested, are appropriate to characterize the required state itself because they vary in accordance with it (see Goodwin, 2010). In case of the fear, this problem is especially important because it is commonly studied both in animals (primarily rodents) and humans, and it is important to define whether they share the same reaction which can be studied by similar methods with a cross-special applicability of the main conclusions. Most authors incline to the positive answer to this question, considering fear as a central response to the perceived threat, which is mediated by neural circuits universal for all mammals and lead to the relatively same physiological and behavioral responses (Adolphs, 2013; Silva, Gross & Graff, 2016). However, LeDoux (2014) argues that the word “fear” is not completely suitable to describe that complex of defensive responses which is analyzed in fear conditioning studies. On the one hand, these responses can be elicited even in the absence of the conscious feeling of the fear (for example, Hamm et al (2003) described a fear conditioning for a visual cue in a patient with cortical blindness). On the other hand, LeDoux presumes that the fear as a result of cognitive processing is not obligatorily linked to the excitation if these survival circuits in all cases. Hence, he suggests reserving the word “fear” only for the conscious experience in humans and introducing another term, such as “defense conditioning” or “threat conditioning” for classical experimental paradigm where an animal learns to perceive the previously neutral stimulus as a dangerous one.

Although the proposed terminology does not become common (since 2014, there are 207 results in Google Scholar for “threat conditioning” and 17 400 for “fear conditioning”), indeed there is a difference in fear measurements between animal and human studies: humans are able to report their fear verbally. Psychologists invest a lot of efforts to develop reliable and valid questionnaires for each particular kind of fear (for example, for spider phobia: Szymanski & O'Donohue, 1995), and it is important. But even in human studies researchers normally combine self-reports with some non-verbal measurements, and, although particular methods are often different for rodents and humans, but it is explained rather by the objective differences between species (such as physical size, or stronger ethical requirements in human studies) then by implied difference in the nature of their fear. In both cases, researchers analyze excitation of fear-related brain circuits, estimate activation of the sympathetic nervous system via physiological responses, and measure behavioral outcomes. This chapter is concentrated on methods which are used in human studies, because the research project itself had been performed on human participants, and because the most important opportunities provided by rodent research had already been briefly described in the previous chapter.

In contemporary research, labor-consuming estimation of neural activity is generally applied to explore specific mechanisms of fear conditioning, not just for the fear measuring per se. Functional magnetic resonance imaging (fMRI) is needed to identify the whole spectrum of brain regions involved in fear conditioning (for a review, see Fullana et al, 2015). Magnetoencephalography has better temporal resolution than fMRI and useful for characterization of rapid neural responses during fear conditioning (Balderston, Schultz, Baillet & Helmstetter, 2014). Positron Emission Tomography (PET) allows characterizing biochemistry of fear conditioning in humans. For example, Ahs, Frick, Furmark, and Fredrikson (2015) demonstrated a negative correlation between serotonin transporter availability and fear acquisition (the functional role of serotonin transporter is to remove extracellular serotonin, so these data highlight its importance for fear-related learning).

In the vast majority of studies, fear is estimated via psychophysiological reactions. It is the golden mean of measuring: on the one hand, recording of vegetative responses is substantially easier and cheaper than neuroimaging studies, and, on the other hand, they are free from self-report biases and reflect biologically based processing of the fear which allows the comparison with animal data. The key psychophysiological measures are skin galvanic response (SGR), heart rate, fear-potentiated startle (eye blink), and pupillary response (Lonsdorf et al, 2017). Conceptually, all of them reflects the acute activation of the sympathetic nervous system which is evolutionarily linked with emotional arousal because its functional role is to prepare the body to fight-or-flight response (Purves et al, 2012; Bradley, Miccoli, Escrig & Lang, 2008; Bach, Friston & Dolan, 2010). This response requires sweating to prevent the overheat of muscles, and this sweating can be measured as changes in skin conductance. It requires increased heart rate to supply muscles with blood. Eye motor control is a bit more complicated: intense unexpected stimulus leads to reflectory eye blink (which potentially helps to prevent the eye from damage) while less intensive and more expected stimulation leads to pupil dilation which allows to receiving more light to appraise the situation.

Electrodermal activity was firstly recorded in XIX century and widely used at the present day. It is considered as one of the best methods to estimate the vegetative component of emotions, including fear response (Christopoulos, Uy, Yap, 2016). There are many possible algorithms of data recording (Boucsein et al, 2012), but the most typical is called exosomatic measurement with direct current. It suggests the application of a small voltage (like 0.5 V) to two electrodes placed on the intact skin. The scheme also includes a resistor in series with the skin, but its resistance is neglectable in comparison with skin resistance, so the current flow can be calculated as I = E/R (in accordance with the Ohm's law), where E is the applied voltage and R is the resistance of the skin. When emotional reaction arises, participant starts sweating, the skin conductance drops and hence the current increases. As an outcome, researchers obtain the graph of changes in current flow, which is simply called “skin galvanic response”. It includes two key compounds: tonic and phasic activity. The first reflects relatively long-term states while the second is short-lasting and can be elicited by a stimulus. Different experimental paradigms may be more concentrated on analysis tonic or phasic response, but in fear-related studies, tonic components are usually subtracted and the analysis is concentrated on stimuli-related events (Lonsdorf et al, 2017). Each particular stimulus presentation might lead to unclear (e.g. delayed, or very weak) response, but average SGR amplitude of many stimuli presentation usually demonstrate a reliable picture of skin conductance changing which allows estimating the emotional arousal elicited by the stimuli.

Behavioral observations are relatively more widely used in animal studies, because a clearly distinguishable behavioral reaction generally needs a strong threat which is connected with ethical and methodological issues in case of human participants (Lonsdorf et al, 2017). However, in studies of specific phobias there is an established method which allows to estimate human reactions: a behavioral approach test (BAT). It is a highly formalized sequence of tasks which suggests the interaction between a phobic participant and a fearful stimulus. The outcome measure is the stage on which the participant refuses to continue the interaction (for example, Soeter & Kindt, 2015) or, in the simplest variant of BAT, the distance between the participant and the phobic object (for example, Cote & Bouchard, 2005). Application of BAT constitutes the gold standard for demonstrating treatment effects in phobias because it solves the problem of transferability of results into the real-world setting (Shiban, Pauli & Muhlberger, 2012).

In the present study, we have combined three measures of fear: formalized self-report (fear of spiders questionnaire, FSQ), skin galvanic response and behavioral approach test with a real spider. The reasons of choice this particular combination of measures and detailed description of the procedure will be provided in chapter 2.

Exposure therapy: a clinical application of fear extinction.

Fear extinction procedure can be effective not only for the attenuation of the recently learned associations between conditioned and unconditioned stimulus but also in decreasing preexisted fears and phobias (Craske et al, 2014). Participants who are suffered from some fear may overcome their negative emotions after exposure to frightening stimulus in the secure laboratory context. This approach, called exposure therapy, traditionally considered as a part of cognitive behavioral therapy, in a frame of which it can be combined with other forms of treatment such as cognitive restructuring or relaxation training.

Different modifications of exposure therapy have proven their efficacy for the treatment of a wide range of mental health problems, including panic disorder, health-related anxiety, generalized anxiety disorder, childhood anxiety disorder and social phobia (Abramowitz, Deacon & Whiteside, 2011), but the most common clinical applications of exposure therapy are the treatment of posttraumatic stress disorder (PTSD), obsessive-compulsive disorder (OCD), and specific phobias.

Before considering these disorders, it is important to clarify some terminological unambiguity. Many relevant sources about the exposure therapy use an umbrella term “anxiety disorders” to address all of the above-mentioned conditions. Indeed, such disorders as OSD and PTSD were included in this group in the previous version of the Diagnostic and Statistical Manual of Mental Disorders (DSM-IV). However, in the newest version of the manual (DSM-V) phobias are still belong to anxiety disorders, whereas posttraumatic stress disorder and obsessive-compulsive disorder are now considered as separate mental health conditions (American Psychiatric Association, 2013).

Rauch, Eftekhari, and Rusec (2012) called exposure therapy “A gold standard for the PTSD treatment” and emphasize that it is not only leading to a decrease of symptoms but also reduces comorbid issues, including depression and negative health perception. In the case of PTSD, treatment usually takes several weeks to be successful and hence is called “prolonged exposure therapy”. As described by Taylor et al. (2003), it includes four sessions (1,5 hours each) of imaginal exposure, during which participants have to describe the traumatic events with a high amount of details and with focusing on the most harmful components of their experience, followed by four sessions of in vivo exposure to stimuli related with the trauma. Then participants receive recordings of their speech and are asked to listen to that for every day of the four following weeks. Moreover, they receive homework tasks related to life exposure (for example, visit the place where the traumatic event had occurred). This method had been compared with two other PTSD treatments, the relaxation training, and the eye movement desensitization and reprocessing, in a randomized design. It has been demonstrated that exposure therapy significantly outperforms two other methods in magnitude and speed of reducing avoidance behavior and in the number of participants who no longer met PTSD criteria after treatment.

Modification of exposure therapy which is commonly used to treat OCD is called “exposure and response prevention”. This approach is considered as a first-line treatment for the OCD since its introduction in the sixties (Meyer, 1966), although onward randomized clinical trials demonstrate that cognitive behavioral therapy techniques such as cognitive restructuring are equally effective for the treatment of OCD (McLean et al, 2001; Whittal, Thordarson & McLean, 2005; Rosa-Alcбzar, Sбnchez-Meca, Gуmez-Conesa & Marнn-Martнnez, 2008). Exposure and response prevention approach assumes that the therapist put the patient into the situation which generally provokes obsessive-compulsive behavior and does not let him/her perform the desired actions. For example, the first patient described by Meyer (1966) has a fear of objects presumably “contaminated by dirt”, could touch them only through the piece of paper and continuously cleaning and washing everything around her. Therapist created additional difficulties for cleaning behavior (such as turning off water taps in the patient's room) and performed treatment sessions dedicated to activities which were difficult for the patient, such as touching door knobs, handling dust bins, using public transport. During several sessions, the external restrictions were gradually softened, but the patient already has the experience of managing the compulsive behavior by herself. This therapeutic approach is still widely used nowadays, and researchers also mention that treating OSD patients demonstrates significant improvements in comorbid depression (Foa&McLean, 2016).

Exposure therapy is especially closely related to the classical fear extinction in case of specific phobias. Patients interact with fearful stimuli, which leads to attenuation of negative emotions. As well as experimental fear extinction, exposure therapy for specific phobias is a notably short-term treatment. Ost (1989) demonstrated that 18 out of 20 patients who suffer from specific phobia connected with injections, spiders, rats, cats, birds or dogs attend clinically significant and stable improvement throughout the only one session which combines in vivo exposure and modeling (the technique in which therapist demonstrate the interaction with fearful object and encourage patient to follow in tracks). A meta-analyses of 33 exposure therapy studies which included 1193 participants (Wolitzky-Taylor et al., 2008), however, demonstrates that many patients still need more than one visit to therapist (average number is 3.04 sessions), but confirms efficacy of exposure therapy for the wide range of specific phobias, including fear of flying, claustrophobia, dental phobia, acrophobia and blood phobia.

The most widely used method in clinical practice is the "in vivo exposure" (in cases when it is possible, predominantly for animal phobias). During the last years, the number of research dedicated to virtual reality (VR) applications in exposure therapy is rapidly growing. According to meta-analysis of 13 studies comparing VR exposure therapy with other approaches (Powers & Emmelkamp, 2008), its efficacy slightly outperforms the efficacy of the life exposure (Coneh's d = 0.35), and at the same time VR is more comfortable for patients (Garcia-Palacios, Botella, Hoffman & Fabregat, 2007). Imaginary exposure approach is relatively rarely used and seems to be less effective than VR approach (Wiederhold et al., 2002). Sometimes specialists also use videotaped stimuli or photographs of fearful objects (for example, Vansteenwegen et al., 2007; Pace-Schott et al., 2013), but these approaches are more common in the scientific studies of extinction of the preexisted or conditioned fear than in clinical practice, in other words, they are mostly used in the situations where standardization of stimuli set is more important than the therapeutic outcome for the particular patient.

In clinical trials, exposure therapy for specific phobias notably outperforms other approaches from the perspective of response rate, although, as any other treatment, does not care every patient. Choy et al (2007) reviewed studies in which success of the treatment has been estimated by the results of behavioral approach test. In case of animal phobias, the final task is to touch or handle the fearful creature, and 92% of participants were able to do it after several hours of exposure therapy. 87% of participants were able to complete the BAT in case of height or driving phobia, and 79% demonstrated clinically significant improvement in case of claustrophobia.

Exposure therapy has proven benefits for a wide spectrum of mental health issues. Clinical guidelines consider it as the most effective (or one of the best) treatment for PTSD (American Psychiatric Association, 2017), OCD (American Psychiatric Association, 2007) and specific phobias (Wolitzky-Taylor et al, 2008; Hood & Antony, 2012; Morina, Ijntema, Meyerbroker & Emmelkamp, 2015). However, it would be imprudent to declare that this approach is completely free from any disadvantages.

Paradoxically, one of the key problems with the exposure therapy is “public relations”. Although it is one of the best scientifically-proven methods to deal with a wide spectrum of anxiety-related disorders, it is often considered both by the wide audience and practicing psychotherapists as unethical and ineffective (Olatunji, Deacon & Abramowitz, 2009). About the half of American psychologists report that they familiar with the concept of exposure therapy (and particularly with imaginal exposure which is a common part of PTSD treatment), but only 17% use it in the treatment of PTSD. In large extent, it is connected with personal attitudes against exposure therapy, first of all with an apprehension to jeopardize patient's current state (Becker, Zayfert & Anderson, 2004). For the further investigation of therapists' personal beliefs and biases which lead them to needlessly exclude patients from the exposure treatment, Meyer, Farrell, Kemp, Blakey, and Deacon (2014) developed a questionnaire ironically called “Broken Leg Exception Scale” (BLES). The metaphor assumes that if any individual obviously cannot go to the cinema because of the broken leg, this anecdotal evidence leads some therapists to the belief that the cinema is also inappropriate for the people with healthy legs and should be closed at all. To estimate that fallacy, authors developed a list of patients' characteristics, such as emotional fragileness or low level of intelligence, and asked 182 practitioners to decide whether they would exclude a patient with such condition from the exposure therapy group. Importantly, there are no scientific data which suggests that low level of intelligence or any other characteristic from the list is incompatible with this form of treatment. It has been demonstrated that higher BLES scores (i.e. tendency to exclude patients) correlate with higher age of responders, lower level of their education, higher anxiety sensitivity of the therapist. It appears that these irrelevant characteristics of the therapist prevent patients from the opportunity to receive scientifically-based treatment.

Of course, attitudes against exposure therapy among psychologists are not completely groundless. Indeed, it is still unclear whether researchers have access to reliable and comprehensive information about the potential harm for the patients. Jonsson, Alaie, Parling and Arnberg (2014) reviewed clinical trials dedicated to psychological interventions for mental and behavioral disorders and mentioned that only 28 out of 132 trials included any information about the monitoring of harm, and 4 out of these 28 provided full description of the negative events as well as about the methods which used for estimation of them. It is possible to suggest that vast majority of researchers do not observe any negative consequences and correspondingly do not report them, but another plausible explanation is the report bias. Notably, authors found a positive correlation between the presence of the monitoring of harm in the article and the impact factor of the journal where the article is published. They also mentioned that the therapists who treat PTSD are more attentive to potential negative implications of the therapy than specialists who work with other disorders, probably because of the obviously distressing nature of this particular treatment. Indeed, up to a quarter of patients who have gone through imaginal exposure therapy against PTSD demonstrates exacerbation of symptoms during the early stage of treatment. However, there are no significant correlations between this symptom, probability of drop the therapy and clinical outcomes of the treatment (Foa, Zoellner, Feeny, Hembree & Alvarez-Conrad, 2002).