Contrasting the Neural Systems of Social and Non-social Threat Regulation in the Adolescent Brain

Adolescence is a unique physical, social and cognitive transitional period from
childhood to adulthood (Sisk & Foster, 2004) that involves developmental
alterations on multiple levels. Adolescence constitutes a window of heightened
vulnerability to psychiatric disorders (Giedd et al., 2008), such as threat and
anxiety-related disorders (Kessler et al., 2005; Kim-Cohen et al., 2003;
Merikangas et al., 2010). Previous research has shown that adolescents show
deficits in learning about threats and safety relative to both children and
adults (Drysdale et al., 2014; Ganella et al., 2018; Mclaughlin et al., 2015;
Pattwell et al., 2012; Waters et al., 2017). However, our understanding of
these developmental variations in threat regulation is still limited. In order
to further understand the differences in threat regulation, this study will
focus on three core-aspects of development: brain development, sex-steroid
hormones and social sensitivity.

The general aim of this project is to understand developmental differences with
respect to threat regulation, by focusing on brain development, sex-steroid
hormones and social sensitivity. To achieve this, we will focus on two
fundamental forms of threat regulation: implicit safety learning and explicit
cognitive threat regulation. For both the implicit and explicit forms of threat
regulation, we will compare the social and non-social threat regulation in the
transition into an out of adolescence.

Implicit safety learning
Learning which objects or situations you should fear is critical for survival.
However, it is equally important to learn which objects or situations you
should not fear or not fear anymore, in light of new information, also called
implicit safety learning (i.e., fear extinction). A key feature of people with
anxiety disorders is a deficit in safety learning (Graham & Milad, 2011).
*Fears are educated into us, and can, if we wish, be educated out* (Karl
Augustus Meninger). This is the goal of exposure therapy (ET), which is the
first-line psychological treatment for anxiety disorders (Abramowitz et al.,
2012). In this fear conditioning paradigm, the presentation of a neutral cue
(e.g., a green light; conditioned stimulus (CS)) predicts the occurrence of an
aversive event (e.g., an electric shock or loud noise, unconditioned stimulus
(US)). After systematically pairing the CS and the US, the presentation of the
CS alone will elicit a conditioned response (CR), which is indicative of fear.
Repeated presentations of the CS in the absence of the US typically lead to a
reduction in fear. This involves the formation of a safety association and can
be referred to as safety learning (Vervliet et al., 2013). We will compare both
the social and non-social forms of implicit threat regulation.

Explicit cognitive threat regulation
Emotion regulation allows us to adaptively respond to affective events, such as
experienced threats. There are various emotion regulation strategies. The best
studied strategy is cognitive threat regulation, which involves changing one*s
interpretations of affective stimuli. It typically involves instructing
participants to *downregulate* their emotions to feared stimuli by thinking
about them in a way that makes them feel less negative (Gross, 1998). This
strategy appears to be highly effective at regulating affect and physiological
arousal (Buhle et al., 2014). Moreover, the core elements of cognitive threat
regulation are applied in many forms of therapies that are effective in
treating anxiety and threat related disorders, for instance in cognitive
behavioral therapy (Beck, 2005). In this study, next to focusing on the
implicit forms of safety learning, in which participants more implicitly learn
which situations are safe by observing the safe experiences of others, we will
focus on explicit forms of threat regulation as well. Specifically, we will
focus on the cognitive regulation that taps into more effortful control
processes aimed at reducing experienced threat (Buhle et al., 2014; Golkar et
al., 2012). We will consider the effectivity of instructed downregulation of
emotional reactions (through cognitive reappraisal, i.e. *please try to
downregulate your emotions* and by social encouragements as *you can do this!*)
in response to viewing threatening stimuli. We will compare both the social and
non-social forms of explicit cognitive threat regulation.

The general aim for this project would be to understand the neurodevelopmental
differences with respect to implicit and explicit forms of threat regulation,
by focusing on brain development, sex-steroid hormones and social sensitivity
in adolescents.

The first aim would be to directly contrast the neural systems of two
fundamental forms of threat regulation: (implicit) safety learning and
(explicit) cognitive threat regulation. In other words, we want to contrast the
neural correlates of both implicit and explicit forms of social and non-social
threat regulation in the transition into adolescence. This way, we can
investigated whether differences in the neural network of either social or
non-social threat regulation can successfully predict threat regulation.

The second aim would be to investigate to what extent there is a difference
between young adolescents (9-12 years) and mid-adolescents (13-16 years). In
line with the literature, we expect that mid-adolescents have more difficulties
with safety learning than young-adolescents (Pattwell et al., 2012; Waters et
al., 2017). Moreover, we want to investigate whether there is a difference in
effectivity between social and non-social safety learning between both
agegroups. We expect, in general, based on various researches by Golkar et al.,
that social safety learning is more effective than non-social safety learning
for both groups. In addition, in line with the social sensitivity of
mid-adolescents, we expect that for this group, social safety learning is more
effective than non-social safety learning.

The third aim (exploratory) would be to investigate whether the experience of
early stressful life experiences modulates the neurodevelopmental bases of
implicit safety learning and explicit threat regulation. The fourth goal
(exploratory) would be to investigate to what extent the experience of
early-life stress influences the neural correlates of implicit and explicit
forms of threat regulation.

This is a between-subject design and there is one session. This study combines
the implicit and explicit forms of social and non-social threat regulation with
structural and functional neuroimaging (fMRI) in related sets of experiments
(described below). One group participants will undergo social threat regulation
and one group participants will undergo non-social threat regulation. Saliva
will be obtained from all participant to measure sex-steroid hormone levels. In
addition, several questionnaires will be administered outside the scanner to
assess e.g., the level of pubertal development, the experience of early-life
stress and state anxiety. All of these are non-invasive measures.

Implicit safety learning
The participants first undergo a standard threat conditioning procedure in the
acquisition phase in which they learn to discriminate between two stimuli (2 CS
- twee neutrale gezichten) presented in counterbalanced order. The US will be a
loud scream (90 dB), based on other validated threat learning paradigms in
children and adolescents (Britton et al., 2013; Lau et al., 2011). In the first
phase they learn to distinguish between the CS+ and the CS-. De CS+ is a
neutral face that is followed by an anxious face, coupled to the US. The CS- is
a neutral face that is never followed by an anxious phase and also never
coupled to the US. During the safety learning phase, one group of participants
will view a pre-recorded video which shows another peer (age and gender-matched
learning model) who acts very calmly when he/she is exposed to one set of the
CSs (CS+) (social safety learning). The other group of participants will view
the same video, but in absence of the peer (non-social safety learning). In
this phase, the CS+ will not be coupled to the anxious phase / US. Lastly,
participants are re-exposed to all pseudorandomized CSs, in the absence of the
learning model. In this final stage, the efficacy of the safety learning can be
assessed by the SCRs (MR compatible) and online ratings of subjective fear
(self-reported arousal and emotional valency). The SCRs will be measured by a
pair of electrodes attached to the distal phalanges of the index finger of the
left hand.

Explicit cognitive threat regulation
A paradigm of social and non-social cognitive regulation will be developed
inspired by previously established protocols (Buhle et al., 2014; Golkar et
al., 2012; Myruski et al., 2019; Reeck et al., 2016) and will be further
adapted for the use in children and adolescents (McRae et al., 2012). The
procedure will exploit the influence of social support given by another
individual through pre-recorded videos and contrast cognitive threat regulation
in the presence of social support (social regulation group). The other
participants will undergo the pre-recorded videos in absence of social support
(non-social regulation trials). Neutral and negative pictures will be presented
to the participant. At the start of each trial, an instruction word or video is
presented in the middle of the screen (*decrease*, *look* or the pre-recorded
video is shown). In response to *decrease* and the video, the participant
should do try to downregulate the emotion. Then, a picture is presented
(negative if instruction is *decrease* or when the video is shown), negative or
neutral if instruction is to look. Thereafter, there is a self-report rating of
the strength of the negative affect (scale from 1 to 4) and the self-reported
success of downregulation (scale from 1 to 4). Then, the word *relax* is
presented. There are four trial types: decrease negative (non-social cognitive
threat regulation), look negative (non-regulation), look neutral
(non-emotional), decrease negative - video (social cognitive threat
regulation). The pre-recorded video is a video of a race, sex and age-matched
learning model who instructs the participant to downregulate their emotional
reactions (through cognitive reappraisal, i.e. *please try to downregulate your
emotions* and by encouragements as *you can do this!*). This will only be shown
to the participants in the social regulation group: half of the participants.
The efficacy of the cognitive threat regulation can be assessed by the SCRs (MR
compatible) and online ratings of subjective fear (self-reported arousal and
negative affect). The SCRs will be measured by a pair of electrodes attached to
the distal phalanges of the index finger of
the left hand. Pictures stimuli will be taken from the International Affective
Picture System (Lang et al., 1997). Several other researchers, for instance
McRae et al. (2012) and used these pictures as well, in children and in
adolescents and adults (Stephanou et al., 2016). In line with, McRae et al.
(2012) we will remove the *worst* pictures, to make sure that these are
suitable for all ages.

No known risks are associated with the participation in the proposed
measurements. MRI scanning is a non-invasive technique without catheterizations
or introduction of exogenous traces. Many children and adults have undergone
MRI studies without any apparent harmful consequences, as long as the absolute
contra-indications are adhered. Absolute contra-indications include the
presence of intra-cranial or intraocular metal or a pacemaker. A relative
contra-indication is claustrophobia. Some people might experience
claustrophobia while inside the scanner. If this happens during the study, it
will be terminated immediately at the subject's request.

Nevertheless, we are aware that participating in an fMRI study might be a
challenge for children. However, we consulted with other research groups and
they were able to successfully conduct similar procedures in children. Based on
that, we argue that the research plan is feasible. Moreover, the tasks are not
performance-based, i.e. participants are encouraged to do their best but they
will not be evaluated on their performances.

Although the participants do not receive direct benefits from participation,
there are greater benefits for society from the knowledge that could be gained
from the proposed research. This knowledge is needed to better understand the
neural systems involved in threat regulation in adolescents. We focus on the
broad range of adolescent years because of the fact that there is a large
increase in prevalence of anxiety disorders in (early) adolescence. We are
convinced of the need to take a developmental approach to study social and
non-social threat regulation in adolescence. Specifically, given that our
understanding of these developmental variations in threat regulation is still
limited. Moreover, if it appears that adolescents show a greater reduction in
fear in response to social threat regulation compared to non-social threat
regulation, or to the implicit instead of the explicit form of threat
regulation, this could be adopted in future therapies.