Adolescence is a period of psychobiological and social changes necessary to achieve psychological maturation and develop autonomy [1,2]. However, because of their history and temperament, some adolescents will have difficulty regulating their emotional reactivity, which could lead to emotional dysregulation and mental suffering when emotions are too intense or violent. This emotional dysregulation can be associated with psychopathology, such as depression [3], eating disorders [4], substance abuse [5], or antisocial and delinquent behavior [6]. In recent years, the impact of attachment bonds on adolescent emotional (dys)regulation has been seriously questioned.

Attachment has been defined as an innate psychobiological system that motivates people to seek proximity in times of distress [7-9]. The quality of this system is determined by early caregiving experiences and results in individual differences in attachment security. Harmonious attachment experiences can result in secure attachment, while disordered experiences can lead to insecure (ie, avoidant or anxious) or disorganized attachment [10]. Attachment patterns become part of the general interpersonal style that will influence strategies of closeness-distance regulation toward the attachment figure, as well as strategies of emotion regulation [11,12].

Despite the change in parent-child interaction during the process of adolescence, where more distance and more conflict linked with autonomy occur, parents remain a secure base for the adolescent in times of distress. It is expected that parents will continue to offer a protective base to help adolescents to regulate their emotions and to sustain a “goal-directed partnership” with them. This relationship context, in which adolescents might experience disagreement situations and develop a variety of emotional responses to deal with potentially elicited distress, is linked to the attachment style of the adolescent [13-16].

Social attachment interaction has health benefits, and its absence can be associated with both physical and mental illness, with broad consequences throughout the lifespan [17]. Researchers have shown that a securely attached adolescent reports less conflict with parents and fewer psychopathological symptoms [18-20]. During conflictual exchange, secure adolescents display collaborative communication, while also expressing anger and hostility. In this context, secure adolescents feel safe to communicate differences of opinions and vulnerabilities with trust and openness [14]. In contrast, insecurely attached, avoidant adolescents display deflecting and/or minimizing attachment strategies during interaction with their parents. They tend to manifest distracted and disengaged behavior and to elaborate on their disagreement, maintaining emotional distance, instead of reciprocity. Adolescents with ambivalent attachment may present entangled and/or oscillatory interactions. Here, the adolescent may manifest distressed or frustrated behavior when facing awkward reciprocity and parental inconstancy [21,22]. These two insecure attachment patterns could become risk factors for emotional dysregulation linked to a broad spectrum of psychopathology (eg, anxiety, depression, conduct disorders). Finally, disorganized parent-adolescent dyadic interaction is characterized by the inability to collaborate, marked by an unbalanced relationship, characterized by different profiles, such as role-confusion or hostile/passive patterns [23].

Recent findings suggest that a common neurobiological system appears to underlie attachment via release of various neuropeptides, the main one being oxytocin (OT) [24]. OT, synthesized in the magnocellular neurons of the supraoptic and paraventricular nuclei of the hypothalamus, projects to the posterior pituitary [25], where it is released into the bloodstream causing a wide range of bodily effects [26-28]. Within the brain, OT can act as a neurotransmitter and/or neuromodulator in various limbic, midbrain, and hindbrain structures [29]. Its central neuromodulatory role in sexual behavior [30], lactation, and childbirth [31] is increasingly recognized. Due to its role in parent-child bonding [32], empathy [24,33,34], trust [35,36], and the promotion of social behavior [37], OT has earned a strong popular and scientific reputation as the “hormone of love” or “hormone of attachment” [38].

These discoveries have led researchers to investigate the effects of intranasal oxytocin (IN-OT) within the healthy [39-45] and clinical [46-49] population. However, it appears that IN-OT can be more beneficial for some individuals than for others, and that the effects of IN-OT seem to be moderated by contextual factors (eg, presence of stranger versus friend [50]), and individual factors (eg, differences in attachment [39,43,44,51-57]). Individual attachment differences have proved to be a strong modulator of OT action [49,55,56], and mediate stress-buffering effects in times of stress [58]. Interestingly, the effects of OT appear to be moderated by attachment anxiety level [40]. In highly preoccupied anxious individuals, OT appears to exacerbate interpersonal insecurity and to affect how the quality of maternal interaction is remembered. After IN-OT administration, highly preoccupied individuals remembered their mothers as less caring and less close, while less anxiously attached individuals remembered their mothers as more caring and more close [40]. The level of attachment avoidance also seems to influence IN-OT effects. In highly avoidant individuals, IN-OT has been shown to increase constructive interaction [52]. For example, under the placebo (PB) condition, avoidant males trusted an unknown protagonist less, feared betrayal more, and decided less often to approach their protagonist cooperatively. However, these effects of attachment avoidance disappeared when males received IN-OT prior to decision-making. Participants high in attachment avoidance show more trust, lower betrayal aversion, and they cooperated more when given IN-OT rather than PB. Other findings [44] have shown that individual differences in attachment act as an endophenotype that moderates the effects of the OT system on social behaviors and cognition following social exclusion. OT appears to contribute to ongoing cooperation with a rejecting but initially cooperative partner, but only for those with low attachment avoidance [44]. Recent research [39] has shown that IN-OT increases communion skills in avoidantly attached individuals, who were especially likely to perceive themselves as more kind, warm, and gentle after receiving IN-OT than after receiving the PB. There was also a major effect of IN-OT on agency [59] in anxiously attached individuals, who showed a selective decrease in independent, self-confident behavior following IN-OT administration. This variability in IN-OT effects raises many questions and controversies regarding the effects of this hormone on social interaction, and especially on the influence of moderating factors. This lack of consensus together with many inconsistent results have led researchers [60-62] to openly question the true impact either of IN-OT in general or of its potential therapeutic effect [63]. Thus, to better understand the effects of OT and the impact of moderating factors on the mechanisms of socio-emotional interaction, we propose an original study protocol that provides new avenues of research into OT from both social and emotional perspectives.

This study protocol seeks to examine the effects of IN-OT administration on emotion regulation strategies in insecure adolescents. A reliable protocol is crucially needed in the field of attachment and OT research, if we want to build a solid theoretical background for interpreting OT, before IN-OT can be proposed as adjuvant therapy in insecure adolescents with emotion dysregulation. We hypothesize that IN-OT will promote a “momentary” subjective experience of attachment safety and proximity-seeking in insecurely attached adolescents during dyadic stressful interaction. We suppose that both anxious and avoidant adolescents under treatment will communicate with their parents with more self-disclosure, and will manifest less negative (eg, controlled, hostile, odd) behavior than under the PB condition. We suppose that IN-OT will positively bias parent-adolescent interaction, especially for highly anxious individuals, because it should attenuate their chronic concerns about distress, separation, and abandonment. OT might support parent-adolescent interaction and emotion regulation for highly avoidantly insecure adolescents (often suspicious and defensive), who avoid expressing emotions in the family context. We hypothesize that, in general, insecurely attached adolescents under IN-OT will process attachment-related emotional information in a more accessible and open way compared to PB conditions. Compared to the PB, IN-OT will increase exploration of distress and comfort pictures (eg, first fixation duration; number and duration of fixations). During visualization of distress pictures, IN-OT will attenuate emotional arousal by decreasing neurophysiological reactivity (amplitude of specific skin conductance response, SCR), and should also decrease SCR and decelerate heart rate (HR).

The main objective is to evaluate the effect of IN-OT on the behavior and discourse of insecure adolescents during a disagreement (stressful situation) with one of their parents. This evaluation will be based on the average score on the scales (4 negative and 1 positive) of the Goal-Corrected Partnership in Adolescence Coding System (GPACS) (personal communication from K Lyons-Ruth, author of the GPACS). Secondary objectives will be to evaluate the effect of IN-OT versus PB on visual exploration strategies for images eliciting attachment-related emotions (ie, distress and comfort). This characterization will be based on the study of eye parameters (ie, total picture fixation time, and average number of fixations per emotion) assessed by a remote eye tracking (ET) device.

We are conducting a randomized, double-blind, placebo-controlled protocol in which insecure adolescent male participants received IN-OT or PB. The protocol was designed to investigate the effects of IN-OT on emotion regulation strategies, by testing whether attachment anxiety and avoidance moderate the effects of IN-OT during a conflict discussion with a parent, and during the visualization of attachment-related pictures.

Emotion (dys)regulation strategies will be evaluated through a multi-dimensional approach: ET and neurophysiological measurements including electrodermal activity (EDA) and HR. These neurophysiological measurements have recently received a great deal of attention as potential biological markers of individual differences in affective response [64,65]. HR reflects the continuous interplay between the sympathetic and the parasympathetic nervous systems and is regarded as a measure of autonomic flexibility and even as a biological marker of emotional response [65]. SCR, a form of EDA, is supposed to primarily reflect autonomic arousal, regardless of the state induced by the stimulus, whether negative or positive [64]. Changes in SCR represent activity within the sympathetic axis of the autonomic nervous system (ANS), measured by autonomic innervations of the skin sweat glands [66]. The neurobiological underpinnings of SCR are widespread and not exclusively related to the defense network. Likewise, HR is also under the control of the ANS [67]. Deceleration of HR is often shown in response to affective stimuli, and is more pronounced in reaction to negative stimuli [67]. HR deceleration seems to be more closely related to attentional processing regarding changes in affective states rather than to valence or arousal [68]. In summary, SCR and HR represent different neural systems, thus allowing us to assess more accurately the effects of OT in response to attachment-related emotions.

We present a randomized, double-blind, placebo-controlled, parallel clinical trial protocol for IN-OT versus PB. We will argue in favor of a multi-method, multi-dimensional (individual and interpersonal strategies) approach to emotion-regulation assessment. We will use two types of attachment paradigms: a conflict discussion with the parent (attachment figure), and the visualization of standardized attachment-related emotional pictures (ie, distress, comfort, joy-complicity, and neutral stimuli) to assess behavioral, neuropsychological, and neurophysiological responses. Although some studies [69-72] have underlined the importance of investigating the effects of IN-OT on adult females, given the risk factors (eg, uterine contraction and menstrual cycle disturbance) for female adolescents in non-clinical environments, this experimental research will be conducted exclusively on healthy male adolescents.

This study will be conducted in Besançon, France at the University Regional Hospital, in the Department of Adolescent and Child Psychiatry.

The study will be carried out in accordance with the principles of the Helsinki Declaration. This protocol is governed by French legislation concerning interventional biomedical research and was submitted to the local ethics committee (ie, Comité de Protection des Personnes-EST II) and approved in 2013 (number 2013-000029-29). The study was also approved by the French Agency for the Safety of Health Products (ie, Agence Nationale de Sécurité du Médicament et des Produits de Santé, ANSM) in 2013. The trial is registered with the Clinical Trials Register (NCT02301715). An anonymous identification code will be attributed to each participant in the study, and the list identifying participating patients with their personal data will be stored by the investigator and kept strictly confidential.

For this study, 60 French-speaking male adolescents, aged 13 to 20 years, each accompanied by one parent, will be recruited from secondary schools in Besançon, France. Compensation will be given to each adolescent (a €20 gift voucher, Ticket Kadéos Universel, Edenred, France) and to the parent accompanying him (a €20 check).

To assess eligibility, a pediatric psychiatrist will interview candidate participants using the Attachment Scale Interview (ASI) [73] to determine attachment style. Adolescents diagnosed as insecurely attached will be accepted in the study. Signed consent will be required from both parents and adolescents. The following inclusion and exclusion criteria will be used (Textbox 1).

Information packs will be given to adolescents on the day of the first meeting. Information packs will contain two detailed study statements: one for parents, and one for the adolescent. The adolescent’s consent form must be signed by both parents and by the adolescent. The participating parent must sign the parental consent form. The investigator will sign informed consent forms, with a copy for each participant, and the original stored by the investigator. Participants will be informed, in the participant’s information leaflet, that the data generated by the study will be accessible only to the investigator, or to the relevant health authorities for official inspection.

Eye-tracking (ET) data will be recorded using the Remote Eye-Tracking Device (RED), a non-invasive, contact-free, automatic ET and head movement compensation solution developed by SensoMotoric Instruments (SMI, Teltow, Germany). This system will capture data with temporal sampling at 250 Hz, spatial resolution of 0.03° and high accuracy of visual angle, 0.4°. The RED system will provide reliable binocular and pupil gaze data and allow subjects to wear glasses or contact lenses. The RED system, including Experiment Center 3.0 to control stimulus presentation, iView X 2.8 to control ET data acquisition, and BeGaze to record data, will be interfaced with a Dell laptop. Pictures will be presented on a stand-alone 20 inch monitor (1680 by 1050 pixel screen resolution) placed approximately 60 cm in front of the participant.

A BIOPAC 5 channel acquisition system (BIOPAC System Inc. Model MP 36, Goleta, CA) and a Dell Pentium computer will be used to collect neurophysiological data. AcqKnowledge 4.3 software (BIOPAC Systems, INC. Goleta, CA) will be used to obtain continuous recordings of the participant’s neurophysiological responses (EDA and HR). Two electrodes (BIOPAC Systems Inc., Model EL 507), placed on the second phalanges of the index and middle fingers of the non-dominant hand [76] will be used to record EDA. The EDA data will be digitized at 1000 samples per second, with a gain of 1000. Low (35 Hz) and high (.05 Hz) pass filters will be applied.

Electrocardiograms (ECG) will be collected, together with cardiac impedance, using a two-electrode (2560, 3 M RED DOTTM) configuration with the bio-impedance module for ground referencing. The electrodes will be connected to a BIOPAC ECG module with the gain set to 1000. The waveform will be used to estimate HR using the AcqKnowledge “Hemodynamics” function. HR will be calculated from the R-R intervals in an ECG.

Attachment will be assessed using a modified ASI, with questions adapted to be applicable to adolescents. This interview elicits the adolescent’s current state of mind regarding the quality of relational experiences with parents and peers. The ASI determines the adolescent’s ability to access and use social support with 3 confidants (ie, parents and peers). This tool assesses relationship quality, social support, and security of attachment style. Several dimensions (ie, mistrust, constraint of closeness, fear of rejection, self-reliance, desire for company, fear of separation, and anger) were coded to determine the attachment profile as either secure or insecure (ie, enmeshed, fearful, angry-dismissive or withdrawn).

The quality of the parent-adolescent interaction will be assessed with the French version of the GPACS. The development of the GPACS drew on prior literature describing behavioral manifestations of security, insecurity, controlling behavior, and behavioral disorganization among younger children toward their parents in stressful situations. It has already been used in several studies [77-79]. The GPACS coding system includes the rating of each interaction on 10 5-point scales from 1 (not at all) to 5 (very much). There are 2 positive scales (1) “Collaborative Communication” indexing cooperative goal-corrected partnership and parent/adolescent carefulness; and (2) “Caregiving Validation of Adolescent’s Voice” indexing the caregiver’s support for the adolescent. There are 8 negative scales characterizing 3 subtypes of disorganized interaction (ie, punitive, caregiving/role confusion, and disoriented behavior). These 8 scales are (1) parental punitive behavior (eg, angry, critical or mocking comments about the adolescent); (2) adolescent punitive behavior (eg, angry, critical or mocking comments about the parent); (3) adolescent disoriented-distractible behavior (eg, suddenly stopping in midsentence and “freezing” with hand in midair or pausing abruptly in mi-sentence); (4) parental disoriented behavior (eg, the same as adolescent disoriented behavior); (5) adolescent odd, out-of-context or contradictory behaviors, which may seem disjointed, startling, or inexplicable to an observer (eg, using a forced, high-pitched, or childish tone of voice, shifting into unusual, fantasy-based topics); (6) parent odd, out-of-context behavior (eg, using a forced, high-pitched, or childish tone of voice, wandering around the room, stiff, usually shifting away from the topic); (7) adolescent attempts to manage or take care of the parent or modulate the parent’s behavior (eg, offering guidance, defusing tension with over-bright, entertaining behavior); and (8) parental role confusion (eg, asking for advice on topics typically discussed with partner or other adult).

To evaluate psychiatric traits in participants we will use an abbreviated version of the Beck Depression Inventory (BDI) [80], the Spielberger State-Trait Anxiety Inventory (STAI, Forms Y A and Y B) [81], and the Toronto Alexithymia Scale (TAS-20) [82].

The whole of each picture will be treated as a single area of interest (AOI). Prior to detailed statistical analyses, AOIs will be divided into 4 categories (distress, comfort, joy-complicity, and neutral) and each category will be analyzed separately. Fixations will be defined temporally and spatially, using a pre set minimum fixation duration of 80 milliseconds and a maximum dispersion value of 100 pixels. In order to explore different gaze patterns, various parameters (ie, fixation order, first fixation duration, fixation duration, number of fixations, and entry time) will be analyzed in relation to attachment style.

The mean fixation order of an AOI category will be calculated by averaging the fixation order of each AOI in the same category of pictures. The mean first fixation duration (ms) in the AOI will be calculated by averaging first fixation duration for each AOI in the same category of pictures. The mean duration of a fixation (ms) in the AOI will be calculated by averaging the fixation duration for each AOI in the same category. The mean number of fixations will be calculated by averaging the number of fixations in each AOI category. The mean sum of fixation duration will be calculated by averaging the sums of duration of all fixations in an AOI category. The mean entry time for an AOI will be determined by averaging entry time for each AOI in the same category.

EDA will be measured by skin conductance level (SCL) and SCRs. The SCR will be defined as the maximum change in conductance (in µSiemens) in the 0.1 to 6 second window after stimulus onset. The change in neurophysiological response will be calculated by subtracting mean levels of neurophysiological response SCRs (ie, latency and amplitude) during the baseline from mean levels during each slide viewing, a practice commonly used in neurophysiological research [65,67]. HR will be calculated from the R-R intervals in the ECG.

First, a psychologist will evaluate the participant’s diagnostic eligibility. Attachment style will be determined using the ASI [74]. Participants who are classified as insecurely attached will return 1 week later for the OT trial (see Figure 2). Randomization will be centralized and performed after the inclusion of the eligible participants. Participants will be randomly assigned to receive either OT or PB intranasally, with both investigators and participants blind to condition. The randomization code will be kept secret by the pharmacy responsible for dispensing the corresponding medication. To reduce bias, randomization will be performed in blocks, with stratification for confounding factors (eg, age). Experiments will be conducted in a comfortable laboratory during the mid-afternoon hours, in accordance with previous research suggesting stability in the diurnal cycle of plasma OT [84]. Each adolescent will first be asked to choose a topic of disagreement to discuss during the parent-adolescent interaction. Then, 45 minutes before the interaction, the adolescent will be instructed to sit and self-administer 1 puff every 30 seconds, alternating nostrils. The adolescent will take each dose in front of the investigator, to assure correct administration and tolerability. The dosage and timing of the nasal spray administration were chosen based on published research on IN-OT, behavior, and emotions in humans [85], and results on cerebrospinal fluid levels after intranasal vasopressin administration [86]. Dyadic interaction will consist of a 5-minute free conversation and a 10-minute conflict discussion that will both be videotaped. The entire interaction session will be transcribed verbatim, and then coded offline for verbal interaction behavior with the GPACS by coders blind to the study hypotheses. After the dyadic stressful interaction, the adolescent will then be isolated in the ET room for 15 minutes, in order to adapt to experimental conditions. The adolescent will be seated in front of the screen in a comfortable viewing position. After this adaptation phase, the EDA electrodes will be fastened to the second phalanges of the index and middle fingers of the non-dominant hand and 2 ECG electrodes will be placed in a bipolar configuration on interior sides of the participant’s wrists [87]. The RED will perform a 9-point calibration procedure. Once this calibration has been successfully accomplished, the 3 minute baseline responses will be recorded with the BIOPAC system. The participant will be informed that the test is due to start, and that he will freely observe a series of 80 pictures. The neurophysiological monitoring AcqKnowledge 4.3 software will be synchronized with Experiment Center 3.0 (RED SMI) software by event markers representing the beginning of each picture. The participant’s eye movements and neurophysiological recordings (EDA and HR) will begin when the first red dot (approximately 0.95° of visual angle) appears. The task will last for about 15 minutes.

Entry into the study (T0) is the first study period and will take place in agreement with the French Agency for the Safety of Health Products and the local ethics committee. The adolescents’ parents agree to be included and sign the informed consent to enter into the clinical active phase of the study. The ASI results will be assessed, taking into account that an insecurity score on the ASI is a prerequisite for entry into the study. The intermediate study period (1 to 24 months, T1) will involve the inclusion criteria described in Textbox 1. In the third (end) study period (30 months, T2) the AOI parameters and neurophysiological data will be analyzed for the adolescent population. The GPACS analyses will be carried out for dyads included in the study. The results will be valorized.

In order to investigate the potential clinical implications of OT as an add-on treatment in psychotherapy, it is necessary to establish a well-designed, randomized, controlled protocol to assess the impact of OT on emotion regulation skills in adolescents. Research shows that OT might induce a momentary emotional state of trust and safety, while enhancing self-confidence. Administration of a single dose of OT might ameliorate emotional regulation skills in adolescents, and thus promote their capacity to connect with parents and peers. Given that adolescence is a critical period for the refinement of interpersonal and intrapersonal competences in close relationships, emotion regulation skills in conflict situations may have direct implications for experiences in close friendships and romantic relationships. Experimental research on OT and emotion regulation in adulthood has provided a scientific basis for the administration of OT, and has already had a positive impact on many short- and long-term health outcomes. The proposed research should contribute significantly to the understanding of the role of OT in emotion regulation, and represents a critical domain that might contribute significantly to improving health and well-being. This protocol could also be used in clinical interventions to increase positive life outcomes for insecure adolescents. Additionally, IN-OT may hold promise for the therapeutic neuroenhancement of parent-adolescent close relationships.

A multi-dimensional approach should prove valuable in characterizing the intra- and inter-individual differences underlying both emotions and attachment, and their impact on social function. Studying attachment-related emotion regulation via oculomotor assessment and neurophysiological strategies during emotional picture viewing might provide additional evidence for the implicit defensive and/or approach strategies of emotion regulation.

Methods traditionally employed in emotion regulation studies [90] require conscious effort and self-monitoring processes from participants. In this study, participants will not be instructed to exert effortful regulation of their emotions. We seek to investigate emotion regulation mechanisms that will be automatically triggered by the stimulus itself and will take place without monitoring or awareness [91].

This is the first correlational study of a healthy adolescent population that explicitly investigates the role of OT in stressful interaction, combined with analysis of oculomotor behavior and neurophysiological reactions in attachment-related pictures (the BAPS [75]). There are, however, some potential limitations (1) only one dose of OT is administered; (2) only the healthy male adolescent population is studied; and (3) only static images are used.

Future investigation should seek to determine the effects of higher doses, administered in both clinical and healthy populations. Researchers should also examine the effects of OT on emotion regulation with different types of social support (eg, peers or strangers). Dynamic stimuli could also be used for ET studies, as they provide more naturalistic interpersonal situations than static pictures. Gaze pattern modulation could then be analyzed over time. Given the central role of OT in human social relationships, further research should integrate longitudinal data on the biological and family and/or environmental systems, focusing on triadic interactions between the healthy male adolescent and both of his parents.