Intranasal Oxytocin as Enhancer of Psychotherapy Outcomes in Severe Mental Illness

1. Introduction 1.1 Oxytocin (OT) OT is a nine-amino acid neuropeptide synthesized in the magnocellular neurons of the paraventricular (PVN) and supraoptic nuclei (SON) of the hypothalamus, and released by the pituitary gland (Meyer-Lindenberg, Domes, Kirsch, \& Heinrichs, 2011). OT has both peripheral and central effects, and serves as a neurotransmitter, a neuromodulator and a hormone (see reviews in de Bono, 2003; Skuse \& Gallagher, 2009). Peripherally, it is released from the posterior pituitary into the bloodstream where it functions as a hormone and influences bodily functions, such as regulating uterine contractions during labor and lactation (Keverne \& Kendrick, 1992). Centrally, through direct axonal release from the PVN, OT acts as a neurotransmitter projecting to various critical brain structures rich with OT receptors such as the hippocampus, amygdala, striatum, and nucleus accumbens (Landgraf \& Neumann, 2004;Macdonald \& Macdonald, 2010; Ross \& Young, 2009). As a neuromodulator, OT creates general volume diffusion, as it is released from all parts of the neuronal membrane into the extracellular space, affecting many regions of the brain (Landgraf \& Neumann, 2004; Leng, Meddle, \& Douglas, 2008; Ludwig \& Leng, 2006).

1.2. Oxytocin and Social Behavior Perhaps one of the most well-known associations of OT is with social behavior. Often referred to as the "social bonding" hormone, OT has been consistently found to be involved in the formation and management of social bonds and in the promotion of affiliative prosocial behavior (Theodoridou, Rowe, Penton-Voak, \& Rogers, 2009). Several studies have shown that OT administration increases cooperation, generosity and trust in others. For example, Mikolajczak et al. (2010) investigated the role of OT in interpersonal trust, by evaluating participants' willingness to take social risks compared to nonsocial risks. They found that participants who received IN-OT showed higher levels of trust, and expressed willingness to take risks during social (but not general) interactions, as compared to the placebo group. Similarly, Zak et al. (2007) infused participants with IN-OT or placebo before engaging in a blinded money gifting generosity game, in which there had to make a one-shot decision on how to split a sum of money with a stranger. Their findings indicated that participants in the IN-OT group were 80% more generous than those given a placebo (Zak et al., 2007). In an attempt to assess the neural correlatives of the OT-trust association, Baumgartner, Heinrichs, Vonlanthen, Fischbacher, \& Fehr, (2008) assessed participants willingness to trust others even in light of a possible betrayal. They found that participants who had received IN-OT continued to show trusting behavior, even though their trust had been betrayed, as opposed to participants in the placebo group. During this task, they found reduced neural activation in the amygdala and caudate nucleus among participants receiving IN-OT, brain areas which are known to be involved in fear processing (amygdala) and behavioral adaptation (caudate nucleus) in situations with unknown outcome contingencies (Baumgartner, Heinrichs, Vonlanthen, Fischbacher, \& Fehr, 2008).

Several studies have put forth the hypothesis that OT is responsible primarily to pro-social behavior. For example, Guastella, et al., (2008) presented participants with happy, angry, or neutral human faces after receiving IN-OT\placebo, and then tested their memory of these faces. They found that participants administered with OT were more likely to remember previously seen happy faces compared with angry and neutral human faces. These findings have indicated that OT has specific influence aimed encoding and recalling positive social information (Guastella, et al., 2008). In another double-blind, placebo-controlled study, Domes et al., (2007) evaluated male volunteers' ability to infer the affective mental state of others using the Reading the Mind in the Eyes Test (RMET) after IN-OT\placebo. They found that IN-OT significantly improved performance on the RMET compared with placebo, suggesting that OT facilitate the interpretation of subtle social cues from eye regions (Domes, Heinrichs, Michel, Berger, \& Herpertz, 2007). These, as well the previous studies on the effect of OT on trust and generosity, have set grounds for the hypothesis suggesting OT is responsible for pro-social behavior.

Although many studies have demonstrated an association between OT and pro-social behavior, other studies have found some contradicting findings which called the pro-social hypothesis into question. These studies have seemed to demonstrate that OT, under specific contexts, might actually induce a negative effect on social behavior. For example, Shamay-Tsoory et al., (2009) found that OT administration increased feelings of envy in a simulated money game when the other player gained more money, and feelings of gloating during the experience of success. Eckstein et al., (2014) exposed participants to social stress following IN-OT\placebo administration, and found that participants in the OT group reported increased perceived social stress, as well as increased activity in the precuneus and cingulate cortex, areas which are known to be sensitive to signals of social stress. Striepens et al., (2012) found that IN-OT enhanced the impact of negative social stimuli on the induction of startle response, and induced the subsequent memory toward negative rather than neutral items. Using fMRI analysis, the authors also demonstrated that despite reducing amygdala activity, IN-OT have increased the impact of aversive social information by increasing neural responses in the insular cortex - which play a role in emotional modulation of memory.

A proposed theoretical framework that has been suggested to reconcile these seemingly conflicting studies is "The social salience hypothesis" (Shamay-Tsoory \& Abu-Akel, 2016). According to this hypothesis, OT may increase the salience of safety signals, attenuate stress and promote social approach in positive supportive contexts, yet increase the salience of threat signals, diminish social approach and increase anxiety in unpredictable threatening situations. In other words, this hypothesis suggests that OT affects human behavior in an adaptive and context-dependent manner (Andari, Hurlemann, \& Young, 2017). It has been suggested that the underlying mechanism which associates OT with modulation of salience throughout different contexts is through the role of OT in attentional processes. OT receptors are found in visual associative areas in humans (Freeman and Young 2016) and in regions involved in attention such as the diagonal band of Broca (Freeman et al. 2014). Moreover, IN-OT has been found to increases visual fixation towards communicative areas of the face such as the eye regions (Guastella et al. 2008), and has also been found to improve the capacity to read subtle non-verbal cues from eye regions, through which individuals understand the intentions and behaviors of social partners (Domes, Heinrichs, Michel, Berger, \& Herpertz, 2007). These findings have put forth the idea that OT operates through the induction of attentional shifts towards important social information, and thereby increase the salience of socially relevant cues.

A possible neurobiological trajectory to explain the effect of OT on social salience is through its regulatory functions in the dopaminergic system (Grace, 1991), a system which was previously offered to be involved in alerting level of salience to potentially important sensory cues (Bromberg-Martin et al., 2010). In this trajectory, alerting signals are sent to salience-coding dopamine neurons in the VTA, and from the VTA to mesolimbic structures (e.g. the nucleus accumbens). These structures are equally responsive to aversive and rewarding stimuli, and are responsible for the assessment of their value and valence (Bromberg-Martin et al., 2010). It has been suggested that these dopaminergic-mediated effects are modulated by the oxytocinergic system, as OT has also been found to increase activity in the VTA in response to social cues (Groppe et al., 2013), irrespective of their valence. The dopamine-OT interactive effect on salience and attention-reorienting has been suggested to also take place in the amygdala, given its established role in attention reorienting and the assignment of salience to social and positive and negative emotional stimuli (Rosenfeld, Lieberman, \& Jarskog, 2011). Accordingly, fMRI studies found that IN-OT administration enhances connectivity between amygdala and rostral medial frontal areas, regions that are crucial for social information processing and emotion regulation (Sripada et al., 2013).

In addition to contextual variables, the "social salience hypothesis" also stresses inter-individual factors such as gender, and personality traits. Studies have shown that the salience of the social environment, are also influenced by differential interpretations due to inter-individual factors. For example, Kubzansky et al., (2012) examined whether OT enhances responses to social stress, comparing effects across men and women. They found gender differences in response to IN-OT, where men reported less negative affect after the social stress task, while women reported more anger but better math performance. Similarly, Fischer-Shofty et al's (2013) found that OT improved the accuracy of perception of kinship in women, but not in men, while men's performance was improved only for competition recognition. Another inter-imdividual factor which was previously suggested to moderate the effect of OT on social salience is baseline social capabilities. Bartz, et al. (2010) examined whether the normal variance in social proficiency moderates the effects of OT on social-cognitive performance. After receiving IN-OT\placebo participants performed an empathic accuracy task that naturalistically measures social-cognitive abilities (Zaki, Bolger, \& Ochsner, 2008). They found that OT improved empathic accuracy only for less socially proficient individuals (e.g., Shamay-Tsoory \& Abu-Akel, 2016). Similarly, Feeser et al., (2015) examined whether OT improves metalizing abilities in individuals with impaired empathy. They found that OT (but not placebo) enhanced metalizing accuracy in participants with lower empathy scores. This supports the hypothesis that OT effects are strongly depends on inter-individual factors, such as social-cognitive skills like empathy. Other inter-individual differences in OT's effect on social behavior, such as attachment styles, will be discussed later in this paper.

1.3 Oxytocin and Psychotherapy Wampold (2001) defines psychotherapy as a "primarily interpersonal treatment that is based on psychological principles and involves a trained therapist and a client who has a mental disorder, problem or complaint". Several theorists have emphasized the importance of interpersonal connectedness in a therapeutic process. For example, Sigmund Freud, the founder of psychoanalysis, stated that a positive attachment between patient and therapist is crucial for the patient to stay committed to psychotherapy (Freud, 1912, 1913) (see Horvath and Luborsky, 1993). He felt that the positive, reality-based component of the relationship provided the basis for a therapeutic partnership, against the common foe, the client's neurosis (Freud, 1958). Carl Rogers, the founder of client-centered therapy, believed that the effectiveness of every form of psychotherapy is ultimately due to the therapist's capacity to form an authentic, accepting, and empathic relationship with the patient (Rogers, 1957). Accordingly, several studies have shown that successful psychotherapy relies on communication and understanding between the patient and therapist, and on the patient's perception of different aspects of the therapist's character, for example his\her trustworthiness (Ackermann \& Hilsenroth 2003). Given the weight of interpersonal aspects to psychotherapy, it is not surprising that OT, which has a well-established role in social bonding, has been suggested to play a part in different aspects of psychotherapy.

Several studies indicate that OT affects a number of interpersonal processes which are central to a successful therapy. Kéri \& Kiss (2011) explored OT's role in trust, by measuring OT levels after a task requiring intimate trust, sharing personal secrets with a stranger, in 60 healthy participants. They found that OT was elevated in the trust-related condition, (i.e. after sharing secrets), relative to a neutral baseline, (i.e. sharing neutral information), and also relative to a trust-unrelated condition (i.e. a mental arithmetic task), using repeated measures. They also found that stronger habituation of autonomic arousal (using skin conductance response) was associated with higher trust-related OT levels (Kéri \& Kiss, 2011). In another study, Lane et al (2013) examined OT's effect on people's willingness to share their emotions. After receiving IN-OT\placebo, 60 healthy men were then instructed to retrieve a painful memory. Next, participants' willingness to disclose to another person event-related facts vs. event-related emotions was evaluated. Interestingly, while the two groups were equally willing to disclose event-related facts, OT was found to specifically increase the willingness to share event-related emotions, which is known to have both calming and bonding effects (Lane et al., 2013). In another double-blind randomized placebo controlled study, OT's effect on Compassion Focused Imagery (CFI) was explored, that is, imagining another "mind" being deeply compassionate to oneself, while also evaluating the interaction of these effects with self-criticism and feeling socially safe with others. After completion of self-report measures (of self-criticism, attachment style, and social safeness), and receiving OT\placebo, participants attended two imagery sessions. Positive affect was measured before and after each session. They found that OT increased the ease of imagining compassionate qualities, however they found important individual differences in how CFI was experienced, indicating that the effects of oxytocin on affiliation may depend on attachment and self-evaluative styles (as participants higher in self-criticism, lower in self-reassurance, social safeness, and attachment security had less positive experiences of CFI under OT than placebo).

The idea that OT may be instrumental as a facilitator of relationship-focused processes such as psychotherapy has been suggested before (Olff, Langeland, Witteveen \& Denys, 2010), yet it has not been sufficiently studied. For example, in a study conducted among patients with schizophrenia, IN-OT given daily over 6 weeks, improved empathic accuracy when paired with social cognitive training (Davis et al. 2014). In another study related to the therapeutic alliance, IN-OT, but not placebo, improved the level of hypnotic responding in low hypnotizable individuals. It was suggested that this augmentation was made possible through neural networks related to the connection\rapport between the patient and the therapist (Bryant et al., 2012). In Guastella et al.'s (2009) study, 24 IU of IN-OT was given adjacent to exposure therapy for social anxiety disorder (in a period of 5 weeks). The findings showed improved mental representations of self\positive evaluations of appearance and speech performance as exposure treatment sessions progressed, as opposed to placebo, but overall treatment outcome from exposure therapy was not improved (Guastella et al., 2009).

A relatively methodologically sound research design that assessed the impact of IN-OT on facilitation of psychotherapy, can be found In MacDonald et al.'s (2013) study. In this study, 18 men\psychiatric outpatients, diagnosed with major depressive disorder, were given 40IU IN-OT or placebo, prior to a videotaped 20-muinute single session with a therapist. Contrary to their initial hypothesis, they found an increase of anxiety over the course of the therapy session under the influence of IN-OT relative to a placebo. Although these findings seemingly contradict the notion of IN-OT as facilitator of psychotherapy, the authors noted that several restrictions, aimed at improving the internal validity, might have hampered the results. For example, patients were instructed to prepare "3 personal event prompts…. personal incidents that would evoke distress between 4 and 6 on 10 point subjective units of distress scale" and were told that "one of these event prompts would be randomly chosen immediately before a session in which they would discuss this event with a therapist". This clearly does not resemble a first session of psychotherapy. Secondly, the therapists were instructed to "consult a script they held in their lap that contained a limited set of relatively neutral verbal interventions (i.e. "can you tell me more about that"); the script also reminded them to minimize verbal interaction, expressed warmth and overt support", which also do not resemble most kinds of psychotherapy. In line with the salience theory, it can be speculated that the IN-OT increased the salience of anxiety-provoking cues in this situation, as the participants later reported experiencing a lack of warmth from their therapist. As Cardoso (2013) expressed in his letter to the editor reviewing Macdonald's study, the results of this study "remind us that the effects of intranasal oxytocin on behavior vary across context and between persons". Thus, the question of the effect of IN-OT on outcome in standard, interpersonally-focused, psychotherapy, remains to be answered.

One of the suggested mechanisms which might underlie the association between OT and increased outcomes of psychotherapy is through the therapeutic alliance. A trans-theoretical definition of the therapeutic alliance has been defined by Bordin (1979), as consisting of (1) agreement of goals; (2) assignment of tasks; and (3) the development of a bond between patient and therapist. According to this, the optimal therapeutic alliance is achieved when patient and therapist share beliefs with regard to the goals of the treatment and view the methods used to achieve these as efficacious and relevant. Yet, these two components of the alliance can only develop if there is a personal relationship of confidence and regard, since any agreement on goals and tasks requires the patient to believe in the therapist's ability to help them, and the therapist in turn must be confident in the patient's resources. Substantial empirical literature has identified the therapeutic alliance as the most consistent predictor of psychotherapy outcome. Some theorists have referred to the alliance as the "quintessential integrative variable" (Wolfe \& Goldfried, 1988, p. 449) across treatment modalities (Horvath \& Symonds 1991; Martin, Garske \& Davis, 2000). A Recent meta-analysis that examined the impact of the therapeutic alliance found that the alliance was a robust, albeit moderate (r = 0.275) predictor of therapeutic outcome, accounting for about eight percent of variability in outcomes (Horvath et al., 2011). Therefore, one possible trajectory to associate OT to psychotherapy outcome is as an agent which facilitates the enactment of the therapeutic alliance. Accordingly, it has been proposed, that if given adjacent to psychotherapy, IN-OT may strengthen the engagement in the relational component, and act as "reinforcement\augmentation agent" of the therapeutic alliance (Olff et al 2010).

Two recent studies provide some support to the hypothesis that IN-OT is related to changes in the working alliance, which might in turn facilitate psychotherapy outcome. Zilcha-Mano, Porat, Dolev and Shamay-Tsoory (2017) recently conducted a correlational study aimed at assessing OT levels during psychotherapy sessions. 22 patients with major depressive disorder were randomized to one of two 16-session manualized short-term psychodynamic treatments, supportive-expressive or supportive, as part of an on-going randomized control trial. A total of 172 saliva samples were collected, before and after treatment sessions 4, 8, 12, and 16. Findings showed that the levels of OT increased significantly as sessions progressed, and were associated with more instances of conflict and rupture in the alliance, as assessed by (a) drops in the strength of the patient-rated alliance and (b) external behavioral coding of alliance ruptures, as were also seen by stronger efforts by therapists to resolve these rupture. In another recent study, Jobst et al (2018) evaluated whether OT plasma levels predict the clinical outcome of chronic depressive patients after Cognitive Behavioral Analysis System of Psychotherapy (CBASP). In their study, 16 patients with chronic depression, participated in this 10-week inpatient program. OT levels were measured before and after participants played a virtual ball-tossing game that mimics social exclusion. Clinical outcome after CBASP was evaluated with the Beck Depression Inventory-II (BDI-II) and the Hamilton Depression Rating Scale (HAMD-24). Using a repeated measures design, they found that lower OT levels at baseline predicted worse therapeutic outcome, as seen by smaller changes in self-rated depression scores, while higher levels of OT predicted superior outcome. This correlation was not found regarding the observer-rated scores.

To sum, it seems that studies assessing the effect of OT as a possible outcome enhancer in psychotherapy for clinical populations are scarce, and their findings are largely inconsistent. Moreover, these studies have listed many limitations, stemming from the complexity of recruitment in clinical\hospitalized populations; the provision of single-dose OT, which tends to cause a lower and insufficient effect (Goldman, Gomes, Carter, \& Lee, 2011); and short-term studies which do not allow to understand the long term effects OT has on interpersonal processes in therapy. Methodological constraints from past studies include no control group (Scantamburlo, Hansenne, Geenen, Legros, \& Ansseau, 2015), or insufficient manipulation of interpersonal aspects (MacDonald et al., 2013). Therefore, although some have suggested that administration of OT before psychotherapy may strengthen engagement in the therapeutic alliance (Macdonald \& Macdonald 2010), the question regarding this neuropeptide's ability to promote therapeutic processes remains unanswered.

1. The current study In this study we intend to overcome past constraints by evaluating the impact of intranasal administration of OT in patients suffering from acute stages of anxiety and depression disorders and undergoing intensive, relationship-focused psychotherapy during psychiatric hospitalization.
2. Research hypotheses
3. Patients receiving IN-OT will show a larger decreases in anxiety and depression symptoms as compared to a placebo controlled group.
4. Patients receiving IN-OT will exhibit faster increases in WA during psychotherapy as compared to a placebo controlled group.
5. This slope will be differential across patients with different attachment pattern (exploratory).
6. Findings will sustain after controlling for demographic, clinical and personality characteristics.
7. Methods and Procedure One-hundred-and-twenty admitted patients with anxiety and depression disorders will be randomized and double-blindly allocated to two groups: (a) psychotherapy + OT (n=60), and (b) psychotherapy + placebo (n=60). Patients will be followed for four weeks beginning at the start of their hospitalization, after signing a consent form. After completing baseline self-report measurements, they will be assessed for the severity of their anxiety and depression symptoms; their working alliance with their therapist; and their treatment outcome after each session. Psychotherapy will be delivered twice a week. Intranasal OT will be administered twice a day (at 08:00 a.m. and at 17:00 p.m.). The experimental group will receive - 32IU (16IU*2) of OT, Sorbitol, Benzyl, alcohol glycerol, distilled water. The control group will receive - 32IU (16IU*2) of Sorbitol, Benzyl, alcohol glycerol, meaning all ingredients except for the OT.

The substance for both study groups will be prepared in the hospital pharmacy, and will then be delivered to the wards according to names (in identical bottles), after randomization that will be conducted by the pharmacist. OT\PLACEBO will be inhaled in two sprays, one in each nostril. The dosage and form of delivery were determined according to standard guidelines in IN-OT studies. The nursing staff will administer the IN-OT\PLACEBO after proper training by the chief hospital pharmacist and principle investigator. As mentioned, the staff and researchers will be blind to the substance administered to each patient (OT\PLC). A month post intervention, patients will complete self-report measurements as part of a follow-up evaluation.

4.1 Measurements " Socio-demographic questionnaire " The Scale for Suicidal Ideation (SSI) (Beck Steer \& Ranieri, 1988) " The self-report measurement of Adult Attachment (Brennan, Clark, \& Shaver, 1998).

" State-Trait Anxiety Inventory (STAI) (Spielberger, 2010). " Hamilton Rating Scale for Depression (HAM-D) (Hamilton, 1980). " The Big Five Inventory - 5a (John, Donzhue \& Kentle, 1991). " OQ-45 Outcome Questionnaire (Lambert et al., 2012). " Side-effects questionnaire (Maoz, 2018) " Suicidal evaluation will be performed by a senior staff psychiatrist " Hopkins symptoms checklist -short form (HSCL-11) (Lutz, Tholen, Schurch, \& Berking, 2006) " The 6 item Working Alliance Inventory (WAI-6) (Falkenström, Hatcher, Skjulsvik, Larsson, \& Holmqvist, 2014)

1. Safety Past studies have shown that the IN-OT side-effects potential is very low. In a meta-analysis examining IN-OT safety, it was found that the side effect profile of IN-OT is identical to that of Placebo (MacDonald et al., 2011). Minor side effects which were reported by 5-6% of the patients include an increase in calmness\euphoric sensation, light headedness, and headache. Moreover, 3% of the patients reported a sense of stimulation in the nostrils as a result of the spray and dryness in mouth. Previous studies have found that a combination of IN-OT and standard anti-depressant medication does not harm their effectiveness and these pharmaceutical drugs and has even been found as beneficial (Scantamburlo, et al., 2015). Nowithstanding these results, and in order to insure optimal safety, side-effects profile will be evaluated by a one of the study psychiatrists in the entire intervention period, and if needed participation cessation of the patient will be cosidered.
2. Research Participation Cessation All participants will be notified in the informed consent form as well as verbally that they may cease their participation in the study at any time, without any consequences to their overall treatment in the ward, and without effecting their psychotherapy sessions, which they receive as part of their hospitalization. Moreover, if patients will report abnormal side-effects or increased suicidality (SSI>12) - their participation in the study will be terminated. In cases in which the patient will be discharged from hospitalization during the intervention period, for safety reasons, their participation in the study will be terminated.
3. Information safety The study questionnaires will be stored in a locked storage closet in the research lab in Shalvata for 7 years, and then will be shredded. Information collected for statistical analysis will not include participant names or any identifying information, just the participants study serial number, which will be listed in the study LOG and locked as well in the lab unit closet.
4. Statistical analyses

To assess the effect of OT on overall levels of anxiety and depression:

" A One way ANCOVA of t5 levels will be conducted of anxiety and depression, controlling for age, gender, t1 level of anxiety and depression, attachment and personality.

To assess the effect of OT administration on the psychotherapy process (WAI-6):

" Repeated measures ANCOVA will be conducted with demographic, personality and attachment entered as covariates and assessed for their interaction with WAI slope.

" A significant interaction will signify differential slopes of the working alliance across the OT and placebo groups, attachment pattern or personality variables.

" To assess the mediating role of the working alliance on OT-therapy outcome, a mediation analysis (Preacher \& Hayes, 2004) will be performed.

1. Limitations " No control for psychotherapy characteristics. " No control for therapists characteristics. " Produces noise but will not affect the validity of significant findings. " Clinical diagnoses is not confirmed using standardized empirical measures " Mostly self-report measures " The context is not measured