Lecture notes with The Adolescent Brain at Leiden University - 2015/2016

Lecture 1: The Adolescent Brain

Adolescent development

With adolescence we refer to the period after childhood and before adulthood, approximately between the ages of 8 to 25. Puberty is considered to be the start of adolescence, so the bodily changes that come along with the hormonal changes which we refer to as puberty can be seen as the starting point of adolescence. It’s more difficult to determine where adolescence ends, because it’s much more culturally determined, so there are many cultural variations in that.

The brain

Neurons are brain cells, and from the neuron there are axons that extend from the cell body. The axons are nerve fibers that conducts electrical impulses. There are also dendrites, which are branched projections of neurons that receive electrical impulses. There’s a distinction between grey matter and white matter. The gray matter consists of neuronal cell bodies and can be mainly found on the surface of the cerebral cortex and cerebellum and in subcortical structures, for example the amygdala, nucleus accumbens, thalamus, hypothalamus, putamen, etc. The white matter consists of bundles of myelinated nerve cells (axons) connecting gray matter areas of the brain. The white matter carries nerve impulses.

In general we can distinguish between four main lobes of the brain: the frontal, parietal, temporal and occipital lobe.

How do we study the brain?

One of the important ideas when looking at the brain was proposed by Joseph Gall. He came up with the idea of localizing functions of the brain according to the form of the brain, which is referred to as phrenology. He proposed that you can talk about a math-region and language-region, dependent on the skull surface. The idea of localizing functions to certain brain parts is still valid today, although his ideas are not valid anymore.

In the 19^th centuries there have been many post-mortem studies that were very useful. Paul Broca was working with a patient referred to as ‘tan’, because this was the only thing he could say. After he died, they found a lesion in the left inferior frontal cortex. They concluded that this was an important area for language production. In 1870, Wernicke found that the left posterior temporal cortex is important for language understanding. These ideas are confirmed through fMRI.

Another case study is the study of Phineas Gage (1868). He had an accident in which his medial prefrontal cortex and orbitofrontal cortex were damaged. After the accident he had a lot of changes in his behavior. He became impulsive, started gambling and showed inappropriate social behavior.

Today we have a lot of different methods to understand how the brain works. Within this course the focus is on Magnetic Resonance Imaging (MRI) to study the brain structure. When you use functional MRI (fMRI), the focus is on brain functions. The fMRI technique is based on the BOLD-effect (Blood Oxygenation Level Dependent). This is based on the assumption that whenever a part of the brain is active, it needs more oxygen. When you get more oxygen, there is more blood flow to that part of the brain. This changes the amount of hemoglobin in red blood cells. The magnetic properties of hemoglobin are measurable. So this is an indirect measure of brain activation. The advantage of fMRI is that it’s a non-invasive technique and that it has a high spatial resolution (in mm), but the disadvantage is that it has a low temporal resolution, so you can’t measure processes that go very fast. In contrast to this, the EEG has a very high temporal resolution, but a low spatial resolution. Another advantage of EEG is that you can measure for a long period of time, for example during sleep.

It’s important to remember that you can never talk about brain activation just as it is. It’s always in contrast to something else.

It’s important to remember that a BOLD signal is not an absolute, but a relative measure of brain activation. This means that a control condition is crucial. You need a lot of trials per condition so that you can average across these trials. You need at least 20 trials per condition. A fixation between trials is also needed, to slow down the BOLD signal between trials. You also need to be sure that your task is engaging and not to boring for the participants.

Brain development

A lot of brain development takes place prenatally. In the postnatal development there’s an increase in the amount of connections between neurons and there’s an increase in myelination in the cortex. This improves the transmission of information between brain cells. These are two processes that increase after birth. The amount of white matter increases with age, especially in adolescence.

Overall there is a decrease in grey matter with increasing age. This decrease is not linear. Gray matter goes through different processes. First there’s an increase, and after that there’s a decrease. The amount in increase in grey matter is related to a process called synaptogenes. This means that there’s an increase in the amount of synapses. But at one point there’s a process called pruning, which means that there’s an elimination of excess synapses. This happens to connections that are not used. Prefrontal and temporal regions are the latest in development.

Social Information Processing Network (Nelson et al., 2005)

The social information processing network says that we can look at the brain in terms of network. This is a useful approach, because it simplifies the way the brain works. It has a specific focus on what happens in adolescence, and they emphasize that one of the biggest changes we see in behavior is in the social domain. There’s a shift in focus on interaction with parents to the larger social context with peers and romantic relationships. So peers and friends become more important in this period. They propose a neurobiological model, which may serve as a naturally occurring model for the study of developmental changes in brain-behavior relations. It also helps to understand the development of psychopathology. Within this model they suggest three broad networks in the brain: the detection node, the cognitive regulation node, and the affective node.

The detection node is important for perceptual processing. The brain regions that they focus on here are mainly the temporal cortex, but also occipital regions such as the fusiform face area. They suggest that this is where the first detection of social stimuli occurs. Anything that is related to facial expressions, understanding emotions, but also when you see something moving, is related to the detection node.

The affective node is mainly focusing on emotions. There’s a big focus on the amygdala. The insula plays an important role in negative emotions, but it’s also involved in empathy and perception of fairness/unfairness. The nucleus accumbens, part of the striatum, is also important for the processing of reward. All these regions are subcortical brain regions. These regions are the more earlier developed brain regions in terms of evolution.

The third network is the cognition regulation node. In general this has to do with regulation and control of behavior and emotions. Reflection, mentalizing, and perspective taking are more complex forms of processing that are more focused in the prefrontal cortex. The anterior cingulate cortex is a part of the brain in the prefrontal cortex that is related to the regulation of behavior. The detection brain regions show early maturation, while the affective node matures in early adolescence, and the cognitive node is the last one to mature.

The brain regions within the affective node show increased maturation in early adolescence, which has to do with puberty and the hormonal changes in the body and the brain. This is why adolescents might get easily excited, and that they’re very sensitive to emotional and social stimuli. They propose that this is partly explained by the hormones that influence these brain regions. The more regulatory brain regions develop much slower, and their development is less directly related to hormonal changes. So they mature much later in adolescence, and their development goes on into early adulthood. The suggestion is that the affective node is directly related to pubertal hormones, while the development of the cognitive node is independent of the hormones and is related to myelination and pruning, and develops more slowly.

The role of motivation (Crone & Dahl, 2012)

The main point is that there’s a lot of evidence that the prefrontal regions develop much later, yet the evidence is not strong enough to say that children show certain behaviors due to this immaturity of this regions. They group the neuro-imaging studies that have been done across adolescence in three groups.

The first group is cognitive control, like working memory, inhibition of behavior, switching between tasks etc. Some studies find an increase in cognitive control in adolescence, while other studies find a decrease in cognitive control. This means that the evidence is mixed. They suggest that this is because aspects of the tests might influence the motivation of adolescents, which can determine task performance. The type of feedback you give to adolescents during the task might also influence their performance.

The second group of studies they reviewed is about affective processing. They suggest that the evidence seems to be in favor of a peak in mid-adolescence. They stress the fact that motivation might influence performance. Just the prevalence of a friend during the test can influence the activation of brain areas during the test.

The third grouping that they make is related to social reasoning. They suggest that there’s an age related increase in the activity of TPJ, this is a brain region that will come back in the following weeks.

This study emphasize how important it is for adolescents to be really engaged in something to be able to measure what we want to measure.

Lecture 2: Hormones and adolescent brain development

The research of Jiska Peper

Peper studies impulse control in her lab. This contains risk-taking, delay of gratification and impulsive aggression. They measure this in relation to brain development and hormonal levels. During puberty there’s a massive increase in testosterone in boys and estradiol in girls. But both sexes actually produce both hormones, in contrast to what most people think. The first symptoms of psychiatric disorders become apparent during adolescence. Depression, anxiety disorders and schizophrenia are the most prevalent disorders among adolescents. It’s important to study healthy pubertal brains to detect the underlying mechanisms of their behavior and to invent early interventions for psychiatric disorders.

Hormones usually have a bad name, because people think that hormones contribute to yelling girls, aggression etc. According to Peper, this picture is not completely true. From animal studies, we know that these hormones actually have an important organizing function in the brain. Estradiol influences cell proliferation, the branching of neurons and the myelination of axons. The more myelin there is, the more efficient neurons communicate with each other. To study this, they removed the testicles of rats, so that they didn’t enter puberty.

They also use longitudinal research, so that you can follow children over time and see what happens during their development.

The healthy pubertal brain

Neurons and myelin

You’re born with an overflow of neurons. You don’t need all these neurons, so during puberty the gray matter (neurons) decrease. This process is called pruning. During puberty, the brain is still under construction. When you’re growing older, the cerebral cortex is getting thinner. This process starts at the back of the brain, and the front of the brain develops last. This process parallels with puberty, which suggests that there’s a link with pubertal hormones.

Cognitive vs. emotional circuits

Cognitive development is thought to take place in a linear way, whereas the emotional development is thought to develop in a non-linear way. Peters et al. (2014) used a cognitive task to differentiate between negative and positive feedback and how this relates to brain activity. They found that the learning rate increases with age, and that the frontal and parietal part were used while learning. The activity in these areas increased with age.

With another task, they measured reward sensitivity. They found that when participants win money, the striatum is highly active. They also found that there’s a peak in striatum activity during mid-adolescence. This demonstrates the disbalance in brain regions during adolescence.

The research of Peters focuses on the connection between brain areas. This is measured with two techniques. The first is called Diffusion Tensor Imaging (DTI), and the second is called Functional Connectivity (FC). DTI is based on the biological principle of diffusion. In the brain, there’s lots of water, but there’s restriction of movement. DTI measures the displacement of water molecules. This is a measure of the quality of the white bundle tracts, so it’s a measure of structural connectivity. FC is a measure of functional connectivity. This correlates the brain activity in different brain areas. The more the brain areas are synchronized, the higher the functional connectivity is between both regions. This can also be done during rest.

Research showed that the white matter in the brain develops gradually up to 35 years of age. This might have something to do with specialization of neuronal networks. Peper et al. (2013) measured fronto-striatal connectivity. This tract is developing with age. The quality of this connection increases with age. The implications of this development were measured using a delayed-gratification task. The children were the most impulsive, and the young adults were the least impulsive. The quality of this tract predicted their behavior two years later. This demonstrates that you really need MRI or DTI to explain behavior in the future. The better this tract is developed, the better adolescents and children are able to control their impulses.

The role of hormones

Hormone production

Sex hormones are produced in the HPG-axis. The hypothalamus is the starting point, this is a very small but very important gland. Underneath the hypothalamus lies the pituitary gland. The ovaries and the testis are the glands that are responsible for the production of the male and female sex hormones. These sex hormones in turn influence the secondary sexual characteristics that develop during puberty. To measure this, a self-report is used. There are two important periods of hormone production. The first period already takes place before birth, and this starts with the development of the testis. Around the age of nine, the production of LH starts, and this is the first indication that puberty has started. You can only measure this during the night (using morning urine). In girls, this peak is one or two years earlier than in boys. The more hormones you produce at nine years of age, the more white matter you have.

Prenatal hormones

It is known that sex hormones before birth influence the sexual differentiation of the brain. When these sex hormones are absent, the brain develops along female lines. The effects of prenatal sex hormones are permanent. It recently has been found that the sex hormones during puberty have an organizational function as well, instead of just an activational function what was previously thought. This is known from animal studies.

It is difficult to study this in humans, but this can be done with twin studies. In twins, hormones from one fetus go to the other fetus and vice versa. So they exchange their hormones. The brains of boys are always 10% larger than the brain of girls. Opposite sex twins were compared, looking at their total brain volume. This study showed that if you’re a girl and you share the womb with a brother, you have a larger brain than when you’re a girl and you share the womb with a girl. The opposite is also true. This doesn’t lead to an increased IQ, but it can serve as a model for autism.

A way to measure prenatal testosterone exposure is to measure the digit ratio, by measuring the difference between the length of the index finger and the ring finger. This is explained by shared genes. There is a sex difference: females have a higher digit ratio than boys.

Hormones at puberty

The organizational/activational hypothesis should be revised, because puberty hormones actually do have organizational effects, in contrast to this hypothesis. To study hormones at puberty, the braintime-study was carried out, using 300 participants between 8-25 years. The measurements took place at three different time points. The balloon-task was used to measure risk-taking behavior. With every mouse click the balloon inflates, and you earn five cents. But if you go too far, the balloon explodes and you lose all your money. They also measured real-life risk-taking behavior with a questionnaire, to make it more ecologically valid. They also measured the personality of the participants. The balloon-task showed a slight increase in risk-taking, with a peak in risk-taking in females at age 14, and in boys at age 16. This shows the difference in puberty onset in boys versus girls. It was shown that boys with more testosterone had more balloon explosions, while girls with more testosterone won more money. This suggests that in girls, more testosterone might lead to more advantageous risk-taking because they earn more money, while in boys it might lead to disadvantageous risk-taking. With another study, they measured the recent alcohol use. The prediction was that testosterone relates to less brain connectivity, and that this would lead to more risk-taking and in turn to more alcohol consumption. Socially, more testosterone makes you more loose, and at the brain level it makes you more loose as well. That was particularly true for the connectivity between the prefrontal cortex and the amygdala.

White matter and aggression

Pepers et al. (2015) measured external and internal aggression using a questionnaire. The ones who score high on physical and verbal aggression show less connectivity between prefrontal and subcortical areas, while the inverse is true for the ones that score low on aggression. They found that hostility and anger correlate highly with anxiety and depression, and that this correlates with the testosterone level.

The peak of testosterone production seems to be after twenty years of age. This means that we should maybe redefine the period of adolescence, at least when you look at the hormones.

Lecture 3: Social Behavior

The social brain (Blakemore, 2008)

Social cognition is a uniquely human trait and is the ability to understand each other as conscious beings with internal mental states.
The social brain means the network of brain regions involved in social cognition, for example the understanding of others and recognizing and evaluating mental states, feelings and goal-oriented actions. Among the brain regions of the social brain, a few of them will be explained today.

Adolescence

During adolescence there are changes in social-emotional behavior and social cognition. Across adolescence, there’s an increasing amount of time spend with peers. This is also related to an increasing peer sensitivity. This is also related to increasing perspective-taking skills, which are part of the cognitive development. The relationships with the peers also get increasingly complex. Research has shown that these changes are not directly related to an increased cognitive capacity.

Social decision-making (Crone, 2012)

If you’re trying to make decisions in a social context, one of the typical processes that go on is that decisions often involve yourself and another person, and decisions very often involve processes related to thinking about consequences for yourself and consequences for others. To study this process, we look at economic games that are related to fairness, trust and reciprocity. These are all very social contexts. The games involve two players: yourself and the other person. There are very strict and specific rules in these games, and there’s always a specific asset involved in the game. We use these kind of games because you have a simple paradigm that can be used across a wide age range. Since these games are very structured, they enable us to quantify certain behaviors. They’re also very easy to use in neuroimaging studies and in this way you can examine the underlying neural substrates of social behavior. The last advantage is that you can look at individual differences in behavior and neural activation, which can be related to measures of cognitive control or perspective-taking. Being fair means that all parties involved in the interaction have the same amount, whatever the outcome measure is. Central concerns in interactions are self-interest and the interests of the other person.

Dictator game

The dictator game is a game between two players. One player is called the allocater and the other one is called the recipient. The allocator, player one, makes the offer, while player two passively receives. You just ask the participants how they will divide a certain amount of money for example. Whatever the first person decides is how the money will be distributed. The game theory predicts that the allocator keeps everything to him/herself, because we as human beings want to maximize our self-interest and there are no consequences of this action. But according to some studies, one of the most occurring offers is that the allocator gives 50% to the recipient. So actually they will share the money with someone they don’t know.

When you look at the decisions made by the allocator in the dictator game, this is an indicator of prosocial behavior. You see that these concerns for others’ outcomes emerges at very young ages. The size of the donations increases with age between ages 3 and 8, but after the age of 9 the donations no longer differ from adults.

Ultimatum game

In the ultimatum game, two players have a ‘say’ in the decision. The first player makes an offer, and the second player can accept or reject the offer. If the second player rejects the offer, neither of them gets anything. The economic theories suggest that the proposer should offer the smallest possible, because the logic is that the other person otherwise will go empty-handed. But what mostly happens is that proposers very often make an equal split. One thing that can influence their decision is what they think is fair, and what do I think what the other person thinks is fair, which is strategic thinking.

Steinbeis et al. (2012) looked at the dictator game versus the ultimatum game. In the dictator game, you don’t need to have a strategy. In the ultimatum game you’re likely to increase your offer compared to a dictator game, so the difference between the dictator game and the ultimatum game is seen as a measure of strategic social behavior. They found that as individuals get older, they give more in the ultimatum game compared to the dictator game. So individuals start acting more strategically when they get older. They also let children do a stop-signal test. This measures how well children can inhibit a motor response. They found a link between this strategic behavior and reaction times on such a motor task. So they suggest that strategic behavior is also related to your ability to control your impulses. This means that children with a lower stop-signal reaction time (SSRT), showed more strategic behavior.

Sally and Hill (2006) looked at the relationship between Theory of Mind and offers in these dictator and ultimatum games. They compared normal developing children of 6, 8 and 10 years old with children with autism disorder. They found that there were no differences between these two groups on the dictator game. For the ultimatum game, you see that the 5/5 offer (so an equal offer) is the most frequent distribution. So the majority of the typically developing children make this offer. In the children with autism, this option also occurs, but a lot less than in typically developing children. You see that 0/5 distribution also occurs a lot in children with autism. So they suggest that Theory of Mind deficits might play an important role in how you adept your behavior in a social context in order to maximize your outcomes. This is the case because it’s important to take other people’s perspectives into perspective in such strategic behavior.

Ultimate game – responder

The responder also has an important role in the ultimatum game. If you look at the economic theories, you would expect the responders to accept every offer, because otherwise they will get nothing. But what we typically find is that responders reject any offer that is less than 20% of the stake. The decision that they are making might be closer to fairness considerations than in the proposer, because the proposer also has to make a strategic decision. You see that if the offers are equal split, individuals are likely to accept these offers. The lower the offers are for the second player, the more likely they are to reject the offers. This is only true if they think that the offer is made by a human-being, because they’re more likely to accept an unfair offer if they think it’s made by a computer.

Role of intentionality in fairness

Mini-ultimatum game

In the mini-ultimatum game there are fixed distributions that they can choose from, so the first player has to choose between these different options. There are different alternatives: the fair alternative, the hyperfair alternative and the no alternative condition. In the no alternative condition, there’s no choice: the offer is not fair in both cases. For 9 year old children, the responses to an unfair offer is the same regardless of which condition. The older the participants get, the more differences there are between the different conditions. The older the participants get, the less likely they are to reject an unfair distribution in the no alternative condition than when they no someone had an alternative offer. This is because they can take the perspective of the proposer. Another thing is that older participants have more impulse control.

Regarding the brain regions, there’s a positive correlation between age and brain activation in the Temporal Parietal Junction (TPJ) in the no alternative condition. This brain region is related to mentalizing skills. There are also correlations between age and the dorsolateral prefrontal cortex in the no alternative condition. It could be that age related changes in brain functioning might lead to age related changes in behavior.

Steinbeis et al. (2012) found that there are age-related changes in the dorsolateral prefrontal cortex activity related to ultimatum game and dictator game decisions. So with increasing age there’s more activity in the dorsolateral prefrontal cortex. They also found that this change is related to behavior. So the more strategic the participants are, the more activity they found in the dorsolateral prefrontal cortex. Finally, they found that DLPFC activity correlates with impulse control. So maybe there’s a mediating role of brain activity.

Trust game

In the trust game, the first player is given two choices. They can either share a certain amount of points. They can share it more or less equally (no trust condition), or they can give everything to the second player. If they go for the second option, the experimenter triples the amount of points in the game. This is the trust condition. The second player also has two options. They can either give 40 points to the first player, and keep 290 points to themselves (defect option) or they can give 160 points to the first person and keep 170 points to themselves (reciprocate option).

Van den Bos et al. (2009) showed that there’s an age related increase in the trust condition for the first players. For the second players, there’s an age related increase in reciprocity: the older they get, the more likely they are to reciprocate. Important to remember is that when the second person can choose, they can also see the different options the first person had. When you look at young adults, they reciprocate more when they see that the other person has taken a risk to trust you whereas there’s not a significant difference in behavior across the conditions.

The older participants were, the more activity there was in the DLPFC and the TPJ when they saw that the other person trusted them. For older participants, there’s more activity in the medial prefrontal cortex when they defect than when they reciprocate. The medial prefrontal cortex is related to thinking about the self. In younger children there is high activity in the medial prefrontal cortex regardless of the decision that they make.

Summarizing we can say that strategic behavior develops with age across adolescence. Furthermore we found that perspective-taking and impulse control abilities are important for strategic and prosocial behavior. Behavioral changes in strategic/prosocial behavior are accompanied by an increased recruitment of brain regions implicated in impulse control (dlPFC) and perspective taking (TPJ).

Lecture 4: Learning, unlearning and controlling fear in the developing brain

There’s a distinction between biological/innate fears and acquired fears. To study fears, mice and human are compared. The kind of behaviors that could be conserved across species are social nurture and fear learning.

What is fear?

Fear is an emotion which involves neural, physiological, subjective and behavioral signatures. The amygdala is very important in fear, and the activation of the amygdala leads to heightened autonomic arousal and release of the stress hormone cortisol. You also become more heightened to processing stimuli in a vigilant way and it causes you to ‘feel’ afraid. All of this leads to defensive behavior (freezing/flight).

Fear learning and unlearning

In the lab, you can’t expose people to real fear stimuli. What you usually have in fear learning is an unconditioned stimulus, you would be afraid of this whenever you experience this stimulus, for example a shock. There’s also a conditioned stimulus, for example a sound. When these stimuli are connected, the tone becomes predictive of the upcoming painful stimulus. People start to show the response to the unconditioned stimulus to the conditioned stimulus, so they start to sweat in response to the tone. This is called a conditioned response. Fear extinction means that you present the conditioned stimulus without the unconditioned stimulus. What actually happens is that you form a new memory, so it’s not the same as the unlearning of fear. This helps us understand the process of spontaneous recovering. If you actually unlearned the fear, you wouldn’t see any return of the fear response when the conditioned stimulus is presented. But what you actually see is that you give them the conditioned stimulus, and that the fear response comes back immediately.

The neural mechanisms of fears

There are two routes to the amygdala to process fears. There’s a fast route which goes through the visual thalamus and then directly to the amygdala. The slower route goes to the visual cortex and is transmitted more slowly to the amygdala. The amygdala is important not just when you see something fearful, but also in forming the association between something that you are afraid of and something that you haven’t been afraid of in the past. The specific nuclei that are important are the Lateral nucleus, the basal nucleus, the accessory basal nucleus and the central nucleus. The lateral amygdala is important in forming the association between the US and the CS, so it’s important in the plasticity.

In extinction learning it’s not about the unlearning of fear, but it’s about the formation of a new memory. Damage to the ventromedial prefrontal cortex leads to inhibition of conditioned responses. Conditioned responses can reappear after passage of time (spontaneous recovery), re-exposure to US (reinstatement), or exposure to the original context (renewal).

How does fear change with age?

With age, brain matter decreases. In some parts it decreases faster than in others. It decreases most slowly in prefrontal cortical brain regions. There are also large changes in brain connectivity. It’s always about networks of brain regions. These developmental changes in the structural connectivity are strong predictors for how these brain regions work when you’re doing a task. Researchers have shown that brain regions that are far away from each other become more connected with age. During late adolescence there’s a decrease in neural synchrony. Subcortical structures develop early, in particular the ventral striatum and the amygdala. These regions develop earlier than the cortical structures, in particular the prefrontal cortex. This causes an imbalance between brain regions because of the different time courses. There are various models to explain this: the dual system model, a triadic model and an imbalance model.

This means that there’s a developmental gap between early developing brain regions of emotion and motivation, and later developing regions of control and regulation. This should manifest itself in clear age-related differences related to the interplay of cognition and emotion. Adolescence is typically considered a storm-and-stress period of development. Adolescence is also a critical period with important ramifications for the development of mental disorders, especially affective disorders.

The interplay of emotion and control

Hare et al. (2008) presented participants with stimuli of fearful faces and sometimes amongst these fearful faces there was a neutral face. Participants had to press a button when they saw a fearful face, but they had to withhold this button press when they saw a neutral face. What they looked at was the amygdala respons to fearful faces. What you see is an increase when they see a fearful face, especially in the adolescence. The more quickly the amygdala habituated to fearful faces, the lower the trait anxiety of the participants. This was particularly pronounced for children and adults, but adolescents didn’t show this habituation system.

Pattwell et al. (2012) looked at the fear extinction in mice and humans. They showed pretty much the same pattern in mice and in humans. The study shows a lack of fear extinction in adolescents in comparison to children and adults. The question is if there’s a molecular basis that is shared across the two species. The hypothesis is that single-nucleotide polymorphisms in the brain derived neurotrophic factor (BDNF) gene is involved in anxiety-related behavior in humans. They found a difference between the MET-carriers and the VAL-carriers, which are different genes.

CBT is the only evidence-based treatment for anxiety disorders, but it works in only 40 to 50% of the cases. In adolescence, CBT works less, because extinction learning works less as well in adolescents.
 

Neurobiological informed treatments

The mechanism of memory reconsolidation is effectively used in the blocking of fear. It is based on the fact that memories are consolidated each time they are retrieved. The stored information is rendered labile after being retrieved. This means that you can add things to it, it’s not a fixed memory. This ability to change memories is a potentially effective mechanism by which new information can replace old information. The memory reconsolidation happens in the amygdala, not in the prefrontal cortex. The extinction works much better for adults than for adolescents. The extinction works better for adolescents if they got a reminder 10 minutes before.

Stressor control

Stressor control refers to the perceived ability that we are in control of stressful experiences in our own life. The subjectively experienced efficacy is a strong determinant how we perceive threats. It predicts how well one bounces back from stressful life events. Believing oneself to be not in control is a hallmark of affective disorders such as depression and anxiety. Only in the escapable stressor group there was no spontaneous recovery. This means that stressor control is really important in stress related effects.

Brain regions involved in learning control over a stressor

Collins et al. (2014) found that the ability to learn control over a stressor was strongly predicted by the connectivity between the amygdala and the anterior cingulate cortex. The stronger the connectivity between these regions are, the better the participants are in learning that they’re in control of a particular stimulus. Learning control over a stressor strengthens the connectivity between the amygdala and the ventromedial prefrontal cortex. So to summarize, learning fear occurs in the lateral amygdala, while regulating fear is mediated by the ventromedial prefrontal cortex.

Lecture 5: Social decision making in the peer context

Decisions in interaction with peers

Decisions involving others involve several cognitive processes. Important processes are for example self-regulation. You need to find a balance between your own interest and the interests of someone else. It’s also very important to understand other people’s state of mind, so you need to take the perspective of someone else, this is called mentalizing. You also can have different motivations to interact with someone, so there are different emotional processes involved.

Adolescence: re-orientation to peers

Adolescence is a period of a re-orientation to the peer-group, so peers become more important. The presence of peers influences decisions. For example it influences risky decisions, but also more healthy decisions such as working together to achieve something. Especially friends become more important during adolescence, and these relationships also become more complex. You’re also interacting with other people. The degree of involvement of underlying processes in social interactions might be influenced by the type of relationship people have with their peers, and the age.

Having friends

Having friends is associated with psychosocial adjustment, cancer survival, well-being in old age, decreased cortisol under stress and protection against mortality. A lack of friends has been associated with cancer mortality and depressive symptoms. Having friends is a proxy for being socially skilled (making/keeping friends). You can distinguish between the identity of friends and the quality of friendships.

Prosocial behavior in the peer context

An example to measure prosocial behavior is the dictator game. A fair distribution of coins is the prosocial decision. So the participants are asked whether they are willing to share their coins with the other person. The other person cannot respond. Güroglu et al. (2014) looked at different interaction partners and the prosocial behavior. For nine years old children, they don’t really make a difference between different types of partners. The same goes for the twelve years old, but for the 15 and 18 years old children, the behavior starts to change. They start to make a difference between different interaction partners. With increasing age, there’s an increasing differentiation between interaction partners. With increasing age, the non-costly prosocial behavior for friends and neutral classmates starts to increase.

When participants are asked whether they are willing to share their coins, so costly prosocial behavior, you see again that the participants of 15 and 18 years of age, there’s an increase in differentiation between the interaction partners. You can also see that there’s a decrease in costly prosocial behavior for unfamiliar peers.

Perspective taking and prosocial behavior

We’ve learned that prosocial behavior change with age. Also, perspective taking skills increase with age, and there was a mediating role of perspective taking to explain age. So the type of game matters: whether you give or share coins, and it’s determined by the type of person with whom you’re interacting.

Brain areas involved in social interactions

The striatum and the dorsal and ventral medial prefrontal cortex and the temporal parietal junction are important areas in social interactions. The striatum is sensitive for rewards. The TPJ and the medial prefrontal cortex are involved in mentalizing skills. The TPJ is involved in the ability to draw your attention to another person and the medial prefrontal cortex is very important in self- and other regarded thinking.
The heads-or-tails gambling game measured how interaction partners modulate reward and mentalizing processes. In this game the person wins or loses money for the self, a friend or a disliked (unfamiliar) peer.

fMRI task: reward sensitivity

When winning or losing money for another person (a friend or a disliked peer), this yielded activation in the TPJ, precuneus and mPFC. The action in these brain areas is much lower when they won or lose money for themselves than when this was for another person. There was also an interaction between the person they’re playing for and the outcome. Winning for the self and a friend was associated with heightened stratial activation. When compared to winning, losing was associated with heightened stratial activation for a disliked peer. Maybe because winning for the self and a friend is more rewarding, and losing for a disliked peer is more rewarding.

Mid-to late adolescents activated the striatum more when winning for themselves versus losing for themselves, versus early adolescence and adulthood. This is probably because they’re more sensitive to rewards. There was also a peak in medial prefrontal cortex activation when winning for disliked peers versus losing.

We can conclude that the TPJ, precuneus and mPFC activity is modulated by the beneficiary. Striatum responses to rewards are modulated by the beneficiary. There’s increased striatal activation when winning for the self and for a friend, and there’s increased striatal activation when losing for disliked peers. Striatum responses to reward for the self peaks in mid-adolescence.

Adolescent behavior in social context

Peer influence on risk-taking

There’s an increased reward sensitivity during adolescence. When you’re winning money versus losing money for yourselves there’s a peak for striatal activity during adolescence. Adolescence is a very important period for peer relationships. If we combine these aspects, we might think that interaction with friends might be very rewarding. The risk-taking behavior maybe becomes even more rewarding when with peers.

Chein et al. used the stoplight-task to look at risk-taking behavior in a peer context. The researchers asked the participants to drive a car (on the computer), and they had a limited amount of time to come to the endpoint. When they were driving the car there were a lot of traffic lights. Sometimes the light turned orange, and then the participants had to make the decision if they would stop or if they would take the risk to drive through. First they looked at the number of crashes that they made, and this was used as an index of risk-taking behavior. The participants had to play this game two times. The second time they had to take a friend with them, and the friend had to sit next to them and don’t talk. They did this study with three age groups: 14, 19 and 37 years old. When the participants were asked to play the game alone, the age groups didn’t differ in the car crashes that they made. But when a friend was sitting next to them, the 14-years old had significant more car crashes than the 19 and 37 years old. So when a friend is present, this might cause adolescents to take more risk. In the lateral prefrontal cortex, there’s more activity in the adults in both conditions, than in the two younger age groups. So this means that the lateral prefrontal cortex is an important area for the regulation of behavior. They also found an interaction between age and the context of alone versus peers in the ventral striatum. The ventral striatum was significant more active for 14-years old when they were in the peer condition. This difference did not exist for the other age groups. For the older age groups the striatum activity was much lower, and did not differ between the two contexts.

So we can conclude that the presence of friends might influence risk-taking behavior. For the young adolescents, there’s higher striatum and lower lPFC activation at the presence of friends. Risk-taking in the peer context is highly rewarding for adolescents, and with age there’s an increasing activation of the control/regulation network.

Friendships and well-being

Güroglu et al (2008) did a study with an orchestra. In advance they asked them all to indicate who they liked the most, who they liked the least and who they were neutral about. In the fMRI scanner they were asked to imagine that they enter a room and the room was empty except for one person standing in the room. They were asked to indicate if they would like to go to this person, or go away from this person, or that they feel neutral and don’t have a strong preference. For every participant, the pictures that they saw were different, based on who they liked and disliked. In order to control for the role of emotions, they included celebrities, from which they knew that the participants liked or really disliked these celebrities.

Peer > celebrities

All the brain regions were more active when they saw a picture of a peer than when they saw the picture of the celebrity. All the brain regions that are related to mentalizing are more active when you’re interacting with someone you personally know. The more subcortical regions that are known as the affective network and the cognitive network was also more active when someone was interacting with a person that’s personally relevant to them.

Interactions with friends

After this, they looked at the brain activity when the participants were interacting with friends specifically. These brain regions were specifically more active for friends compared to neutral or disliked orchestra members and the celebrities. They saw more activity in the amygdala (emotional arousal) and the hippocampus (memory retrieval). They also saw that the nucleus accumbens is also more active specifically for interactions with friends. This can be because the interaction with friends might be a social reward. The ventro-medial prefrontal cortex was also more activity. The study shows that it’s important for emotion-regulation, empathy and reward processes as well, because it has close relationships with the striatum. It could be that this region is active because you’re trying to find out what your friend is thinking.

Prosocial peer influence

Participants were asked to come with a friend to the lab. In the study were 30 participants of 12-13 years old and 30 participants of 15-16 years old. They were told that they were going to play a public goods game. They were told that they were playing this game with three other people, which were from their age and sex. In every round, every player is given a number of coins. They can decide if they wanted to keep these coins themselves or if they wanted to donate it to the public pot. For the best outcome for the complete outcome, everyone should donate the money to the public pot, because the money would then be doubled and distributed. There was a feedback condition, in which they were alone and the participants were told that they wouldn’t receive any feedback. In the spectator condition, there were confederates, which they had met before the game. They were shown pictures of the people they just met, so that they knew that these others kids are watching their decision. In the evaluation condition, everything is the same as in the second condition, but now the peers could give their thumbs-up for the donation. So they were evaluating the decisions.

It was a fMRI task. It was done in two runs. Across the conditions, there’s an increasing level of prosocial behavior. So in the evaluation condition there was a higher number of donations compared to the other two conditions. So positive feedback from peers can influence your behavior in a positive way. The second important thing is that younger kids donated more tokens than the older participants.

On the stimulus onset (the moment the tokens are given to you and you have to make a decision) we see greater activation of the mentalizing network (taking other’s perspective etc) in the evaluation condition than when the participants were alone. In terms of age differenes, an effect was found in the dorsal medial prefrontal cortex. The younger participants had more activation in the dMPFC in the evaluation condition than in the spectator condition and the alone condition. This difference is higher for younger participants than for older participants. There’s also a positive relationship between the activity of the TPJ and the number of tokens that were donated by the participants. This brain area is also very important for mentalizing, probably because when they are able to take the other’s perspective more, they also donate more tokens. So to conclude, younger adolescents are more susceptible for peer influence.

Social stress: exclusion

Social exclusion is very important, because when you’re excluded, this damages one of the very basic needs that we have, and that is the need to belong. This might explain the links between social exclusion and negative outcomes such as loneliness and depression. One study found that loneliness at the age of nine was related to depression at the age of 13, and this link was also there for age 5 to age 13. The highest rate of depression was found for children who felt lonely at the age of 5 and the age of 9.

Social exclusion is measured by the cyberball task. In this task, three participants are playing it. The participant in the study is told that they are going to play the game with two other players. First of all, every player gets the ball 33% of the time. In the exclusion condition, participants receive the ball once at the beginning of the game and after this get ostracized. After the exclusion block, participants of all ages show a significant decrease in mood and self-esteem and belonging. After an inclusion block, all the participants feel good about themselves. This has been replicated in a lot of studies. The mood after observed exclusion, so when they observe others being excluded, their mood also drops significantly. Two important brain regions in social exclusion are the anterior cingulate cortex and the insula. These regions are also more active in situations where physical pain is involved. So maybe exclusion can be compared to physical pain. The anterior cingulate cortex is also involved in the regulation of conflicts, and the insula is involved in negative affect.
In another study, two groups were compared. In one group of kids, there were positive and in another group there were negative experiences with their peers. There were higher levels of activity in the cingulate cortex in participants who were chronically rejected compared to the participants who were stably accepted.

Another study showed a negative relationship between the time spent with friends and insula and dACC, so the regions associated with social exclusion. So it might be that social contact with friends also acts as a protective factor against effects of social exclusion. But there can be other factors that can also explain these findings.

Lecture 6: Self-concept development

Definitions of self

It’s different to distinguish between different aspects of the self. Self-representations are characteristics of the self that are consciously acknowledged by the individual through language. We often talk about this in terms of our identity. The second definition of the self is about self-esteem. Self-esteem is the overall evaluation of one’s worth or value as a person. This is very important for performance on all kinds of tests. So the first type is more in terms of cognitive approaches, while the second type has more to do with affective dimensions.
Self-concept has already been studied for centuries. The self can among others be measured through self-reports and through behavioral tasks.

Construction of self

We have changes in the structure and organization of the self. We develop our self as a cognitive construction. We develop certain cognitive skills which allow us to look at ourselves in different ways. The second way of the construction of our self is the developing of the self as a social construction. For example, children adopt opinions of others and they internalize these opinions. Our caregivers have a strong influence on the development of our self-construct. Later in life the peer group influences the construction of our self.

Normative liabilities of the self-concept

The self has multiple positive functions. The first one is the organizational function. The self provides expectations and a predictive structure, and the consequence is that we maintain a coherent picture of ourselves. It also has a motivational function, because it energizes us to pursue goals and to achieve ideals in self-improvement. It also has a protective function, because an overall sense of who you are helps you to go on after a threatening situation for example.

Development of the self-concept

Early to middle childhood

In early to middle childhood, children develop the ability to categorize self-descriptions. It’s still very general and often related to certain actions. The idea of who you are is only based on your own experiences. The self-concept is in terms of labels rather than traits. Children also have a strong positivity bias, so they describe themselves in very positive terms. The reason for this is that children base their self-concept on temporal comparisons, so for example they think: ‘Today I did this better than I did two weeks ago’.
Young children start to understand that others evaluate them, but this is still limited.

Middle to late childhood

In middle to late childhood, children tend to have more global self-descriptions. They focus on interpersonal characteristics in describing themselves. They also start to compare themselves with others, and as a consequence self-evaluation emerges and they start to make social comparisons. They may start to see the difference between the real self and their ideal self.

Early adolescence

In early adolescence, higher-order thinking emerges, but this is not distressing yet. The self-esteem is largely based on appearance and differs per relational context. The self-serving bias is the process where you attribute positive events to yourself, and negative events to others. This is a protective factor which is very important in early adolescence.

Middle adolescence

In middle adolescence, adolescents are very preoccupied with the opinions of others. In middle adolescence there’s a lot of introspection, and the global self-esteem declines with regard to early adolescence. Some researchers have also argued that it’s not a constant decline in self-esteem, but that the self-esteem is more vulnerable during middle adolescence so that it can be high and low. There’s also often a discrepancy in the ideal self versus the actual self. Two types of egocentrism occur during middle adolescence. The personal fable is the idea that you are unique as a person and that no-one can understand what you are going through. The second type is the imaginary audience effect, this means that adolescents think that everybody is watching them all the time. This is also seen as very adaptive, because it helps you to develop your self-identity and become independent from your parents, and it may also help to get a more realistic idea of who you are.

Late adolescence

Late adolescence is the time where the self is more tested against your personal values, beliefs and morals. There’s often a strong focus on our future self, this is often quite idealistic. It gives a sense of direction for a person. Contradictory traits are no longer seen as opposing in late adolescence. There’s even more higher-order abstraction. It’s normal to be a different person in a different context, which individuals start to realize in late adolescence and feel okay about that. Egocentrism becomes less during late adolescence, and perspective taking skills are better developed. But there are large individual differences in egocentrism and perspective taking skills. The personal fable continues in late adolescence. The self-concept becomes more realistic in late adolescence due to the ability to think abstractly. It also becomes more balanced. The self-concept becomes more balanced, but is still a bit overly idealistic.

Self-conscious emotions

Self-conscious emotions are affective reactions to the self, when people fail to live up to the ideal of the actual self-representations. Pride is a positive self-conscious emotion, but shame, guilt, embarrassment and humiliation are negative. Pride is a positive emotion but it differs from other positive emotions in the sense that it is sensitive to self-evaluations. Pride is often considered adaptive, because it propels people to act prosocially and it motivates moral actions. It’s not always considered adaptive, because in the case of hubris or narcissism it’s considered non-adaptive. Hubris means that you are proud of who you are, instead of that you’re proud of what you did.

Shame is an emotion which results from moral wrongdoings that violate moral standards. It can result from achievement failures, a lack of competence. It results in a desire to hide from others, to distance oneself from people who are able to evaluate the self. The associated emotions are disappointment and depression and most of the time shame is not seen as adaptive emotion.

Shame and guilt are both moral emotions and they’re both related to internal attributions. They’re also both experienced in interpersonal situations. They differ in cognitive attributional features. Shame is experienced in moral wrongdoings that violate other’s ideals of the self, and it threatens one’s social status and identity. Guilt is experienced when you violate standards and moral rules of how one ought to behave towards others. The attribution is internal, but unstable and more controllable than shame.

Embarrassment and shame use the same word in Dutch, but they have a very different meaning. Embarrassment is experienced in case of public violence of social norms of civility, awkward and unintended accidents observed by others. It can also be experienced when one is complemented or praised. In embarrassment there’s a lower level of responsibility and self-blame, and the causes are viewed as accidental. It’s also related to the imaginary audience effect. It may evoke sympathy.

Adolescent brain development

Thinking about the self relies on several regions in the brain, for example the medial prefrontal cortex. This area is more advanced in humans than in animals, and this regions helps to think about ourselves in the past, present and the future. Another region is the ventral striatum, which is very active and sensitive in adolescence. The last region is the temporal parietal junction, which is part of the social brain and is often involved when we think about intentions of others. So there are distinct regions in the brain that are active when you think about yourself versus when you think about others or about how others view you.

Adolescents and adults rely on the same areas of the brain in self-evaluative tasks, this is the ventral medial prefrontal cortex, but this activation is larger in adolescents than in adults. This indicates more self-involvement when you think about yourself in adolescence. The dorsal medial prefrontal cortex is more active when you think about others versus when you think about yourself in adolescents and adults.

When adults think about themselves in the perspective of another person, they also show activation in the temporal parietal junction instead of when adults think about themselves. When adolescents think about themselves, you again see activation in the medial prefrontal cortex, and also when they think about themselves in the perspective of others. But in adolescents, the temporal parietal junction is also active when they think about themselves, regardless of this is in the perspective of others or not. In adolescents, it makes a big difference if they think about themselves from their own perspective versus when they think about themselves from the perspective of others in terms of the activation of the dorsal medial prefrontal cortex. 

Only for adolescents, thinking about yourself from the perspective of your friends results in activation in the ventral striatum, but only for social traits such as popularity. The ventral striatum is sensitive for pubertal hormones.

These different studies show that the parietal temporal junction becomes more selectively active when you get older, and that the medial prefrontal cortex is more active in adolescents than in adults. The last study is about self-conscious emotions. This study shows that around mid to late adolescence, participants feel most embarrassed when they think they are being videotaped in an MRI-scanner. There’s more activity in the medial prefrontal cortex when participants think that they are being videotaped than when they thought the system was off. People who felt more embarrassed, showed stronger activity in the ventral medial prefrontal cortex. 

Lecture 7: the adolescent brain – problem behaviors

Adolescent development

Adolescence is a period where identity development takes place, where there are physical changes, where peer interactions are salient and there’s increased independence and increased exploratory behavior. In the behavior of adolescents, there’s high risk-taking, high exploration, high novelty and sensation seeking, and high activity and social behavior. The function of these behaviors is to develop social skills and to prepare for an independent adult life. So the risk-taking is also functional, because individuals need to explore their capacities et cetera.
The paradox is that adolescence physically is one of the healthies periods in lifetime, but it’s associated with a 200% higher mortality compared to other age groups. There’s a very striking increase in the number of traffic accidents adolescents were involved in.

Addiction

The percentage of those who started drinking or smoking before the age of 16 is significantly higher than those who started after the age of 16. Across the lifespan, we see that adolescence is a sensitive period where adolescents start to engage in these kind of behaviors.
Can these age differences be explained by poor cognitive control? Or can increased risk-taking and impulsivity explain these changes? Can adolescent brain development explain these changes? Was the research question of one of the articles for this week. The idea is that with increasing age, we see slower development of the prefrontal cortex whereas you see a sudden increase related to puberty in the subcortical brain regions such as the striatum. The idea is that the connection between these two brain regions also change with age. In children there are low levels of connectivity between the striatum and the prefrontal cortex, which get more strengthened with age. The links of the striatum with the prefrontal cortex is much stronger than the links of the prefrontal cortex leading to the striatum in adolescents. So there’s less control of the prefrontal cortex over the striatum.

Model frontostratial connections

In making their point they mentioned several studies. The striatum is very important for reward processing and learning based on rewards. The prefrontal cortex is involved in cognitive control (resistance to temptation, delay of gratification). The study of Galvan et al. (2006) looked at rewards and risks. The participants are shown a stimulus, and participants are told in advance that these different cues are associated with different kinds of rewards. They are asked to give a response as to which side of the fixation cross they saw the cue. If you do it correctly, then you get the reward. They find activity in the ventral striatum, and that this activity is higher in adolescents compared to children and adults. Participants are also asked to fill out questionnaires about their risk behavior, and they found that the activity in the striatum is significantly related to self-reported risk-taking behavior. They didn’t find a relationship between impulsivity and striatum activity, and they found decreased impulsivity with age. So the point is that the age-related changes in risk-taking behavior might be related to the sensitivity of the brain to reward-related cues, but stratial activity is not related with impulsivity. Impulsivity is more related to prefrontal cortex activity and cognitive control.

Cognitive control and motivation

Casey and Jones (2012) talk about the importance of social context/motivation in modulating cognitive control. How do the social context influence the level of cognitive control that you exert? This can be in two ways. First, the social context can improve cognitive control, but it can also decrease cognitive control. So one idea is that rewards (i.e. motivation) can improve performance that requires cognitive control. They used a anti-saccade task for this. They are shown a dot, and they’re supposed to look immediately to the other side, which requires cognitive control. For adolescents there was a significant difference if there was a reward involved, while for adults there were no significant differences. For adolescents there were also higher levels of ventral striatum activation and higher prefrontal cortex activation in the rewarding trials, which caused better performance. 

The second point is how performance can suffer when there’s a reward involved in the situation, which was shown by Sommerville et al. (2012). Participants had to press the button when they saw a neutral face, but not press the button when they saw a happy face. The hits are correct responses, and false alarms are when they’re not supposed to press the button but they did press the button. What you see is that when you compare mistakes when you’re supposed to control your impulses is higher compared to when you see a neutral face for teenagers, compared to children and adults. This is because a happy face is a rewarding cue. In the ventral striatum there are higher levels of activation when looking at happy faces for the adolescent group, and they find that how well you can control your impulses is related to the prefrontal cortex. They also found that the frontostriatal connectivity improves with age. So the main point is that there’s an increasing communication between these two brain regions with increasing age, which is related to performance on the task.   
The point is that we’re always taking into consideration the importance of the context.

Moreover, substances of abuse have dopaminergic effects. So they are rewarding in itself because they activate the brain regions related with reward. So it may exacerbate ventral striatum activity, which is densely packed with dopaminergic receptors, resulting in heightened rewarding property during adolescence. There are studies that show that adolescents seem to be less sensitive to negative effects of alcohol, so they miss the negative feedback of substance use. Finally, the salience of peers for adolescents may increase substance use.

Conclusions

Cognitive control in itself is not an explanation for understanding why adolescents are more engaged in risk-taking. We do see that they’re very sensitive for reward cues, and that the connectivity between the frontal cortex and the striatum is very important and developing across adolescence. With increasing age, these connections become more efficient. Combined with less negative effects of alcohol use and the influence of peers can explain why we see an increased substance use in adolescence.

Affective disorders (Davey et al.)

The second paper focusses more on the onset of affective disorders. Around adolescence we see an increase in the onset of affective disorders like anxiety, depression and schizophrenia. There’s a reorientation from parents towards peers. These social interactions, especially with peers, are related to a very strong social motivation. If there are negative interactions with peers, such as rejection, it can form a major risk factor for the onset of depression. So the social context is highly relevant. This is maybe even more the case for those adolescents who don’t have a social network from their peers or at home to buffer against the negative social interactions. They refer to the mismatch between the cognitive and affective related brain regions. They propose that this mismatch might be a further risk for adolescents. In the Davey et al. paper, they have a slightly different view on this problem, because they don’t see the mismatch per se as a problem.
Davey et al. tried to understand the emergence of depression in adolescents, and they focus on the prefrontal cortex and the reward sensitive areas. They suggest that adolescence onset depression is a forerunner of adult depression. They strongly emphasize the role of reward processing. They propose that the vulnerability to depression is increased by the capacity to anticipate on different kind of rewards than they had in childhood. Adolescents understand rewards that are much more abstract and distant. This capacity to have a very different perspective on rewards, is the change that is closely related to understanding adolescent onset depression. They emphasize the social factors again.

Reward and depression

Especially, positive affect is closely related to the development of depression, and not to negative affect. So it’s really about not having the positive affect. Positive affect is generated by rewarding engagement with the environment. Interpersonal rewards are very rewarding to humans. The social environment is changing in adolescence, and this might be related to changes in how you perceive social reward.
Reduced positive affectivity is important for depression (e.g. reduced energy, reduced appetite). Interpersonal rewards are important contributor to positive affectivity. The social risk hypothesis is that increased perception of risk/failure leads to interpersonal withdrawal and reduce this desire to get these social rewards. You do this to protect yourself and to have less expectations from your social environment.
The nucleus accumbens is part of the ventral striatum, and also an important region of the reward system. They also emphasize other brain regions. The dorsolateral prefrontal cortex is very important in having certain goals and keeping in mind goal-related information, planning, maintaining goal-related information in your memory. The anterior cingulate cortex is important for monitoring your own behaviors and the behavior of others, and monitoring of errors. The hippocampus is important for memory formation and retrieval. Impaired reward functioning is related to all these brain regions in depressed patients. Epstein et al. (2006) showed that depressed patients showed less striatum activity to positive stimuli. This correlated with reduced interest and pleasure.
Davey et al. suggest that there are certain biological and social changes. These are related to psychological changes.

Biological changes (neural)

They emphasize that the development of limbic structures and dopaminergic system with puberty. They also emphasize the improved connectivity between the prefrontal cortex and the striatum across adolescence. Importantly, prefrontal cortex development is related to understanding of reward. The PFC development is also related to more distal and complex rewards (such as social rewards). So the cognitive changes that are directly related to the development of the PFC, is related to the increased reward sensitivity in adolescence.

Social development

There’s a very strong emphasize on peer relationships and romantic relationships. And there’s a high sensitivity to being accepted by peers, your status in the peer group.

Psychological/cognitive aspects

Adolescents are increasingly better in hypothetical situations, reasoning, abstract logic, alternatives, consequences. There’s a focus on identity and future identity. Ruination also emerges: ‘what if’?

Taken together

The model proposes that you have to take all these aspects into consideration to understand the basis of depression in adolescence. Taken together we can say that the key changes that occur in adolescence are the development of the prefrontal cortex, and its tighter integration with the dopaminergic reward system, and the profound social changes that mediate the passage from childhood dependence to adult independence. The development in the prefrontal cortex and the dopaminergic system plus the social changes are so intertwined, that you can’t consider them apart from each other. Initial episodes of clinical depression during adolescence will often result from the frustration, or omission, of a highly anticipated reward: a reward such as romantic love or belonging to a social group. Prefrontal cortex development comes along with a lot of advantages, including the ability to make complex decisions in social environments that take into account the consequences of decisions in the future, but this development sometimes comes with certain costs. There’s a heightened vulnerability to depression when anticipated future rewards are disappointed.

Source

These lecture notes are based on the subject The Adolescent Brain from the year 2015-2016