Introduction

Current understanding of the adolescent brain has been shaped by over a decade’s worth of psychological, physiological, and neurobiological studies. Along with continued scientific efforts, what has been unique about the last decade has been the refreshing narrative shift about adolescence. Whereas early brain studies fueled the notion that the adolescent brain promoted problematic behavior, the contemporary perception is one of respect, awe, and appreciation for the opportunities, innovations, and adaptability that characterize adolescent brain. Both the scientific and nonscientific communities have been plagued by predictable clichés about rebellious attitudes, invincibility, and poor decision-making. But the tide is turning. While still being mindful of the unique challenges some adolescents face, there is increasingly greater appreciation for what has always been known but seldomly been celebrated. Adolescence is a developmental window ripe with opportunity and creativity. From the onset of puberty through the early to mid-twenties young people explore, adapt, and learn through active engagement with the environment, relationships, and trial-and-error learning. Identities are shaped and reshaped, new types of decisions are made, and social bonds are forged. These developmental milestones are at the essence of adolescent development and the adolescent brain is well-suited for these developmental tasks. The normative developmental activities that define adolescence include exploration of new activities and desires, social influence, taking risks, and renegotiating relationships with caregivers. Rather than viewing these behaviors in a negative light, leaders in the field have called for a more nuanced understanding that celebrates these behaviors and conceptualizes their importance in navigating changing social landscapes during adolescence (Crone & Dahl, 2012; Galván, 2014). The Life-span Wisdom Model, for example, frames the “apparent increase in adolescent risk taking as an adaptive need to gain the experience required to assume adult roles and behaviors” (Romer, Reyna, & Satterthwaite, 2017) and aims to move beyond unfair and inappropriate stereotypes of adolescents by highlighting the importance of experience and wisdom gained through exploration.

The goal of this review is to synthesize the last decade of research on the adolescent brain that has led to this narrative shift. This is an exciting task given the rich data, analytics, and energy that has emerged in the past 10 years on this topic. Changes that characterize adolescence are profound, from puberty-related physical maturation to shifts in social dynamics, and risk-taking behaviors. Developmental cognitive neuroscience has made significant strides in uncovering the neural mechanisms and changes that support these behavioral alterations. Regrettably, it is not possible to include the multitude of topics and studies that have contributed to this progress but general themes are used to illustrate current understanding. The overarching framework of this review is inspired by this narrative shift and by the notion put forth by Smith and Thelan (2003) and others that development is a dynamic process both within the individual and between the individual and external inputs. As such, this review will focus on research that has provided evidence, showing that the development of the brain is influenced by multiple elements and shown that no single element (e.g., one neural region, system, or experience) drive its development; rather, the ongoing and dynamic interactions between the elements is what undergirds brain development. The main themes and findings referenced throughout are highlighted as exemplars of how the field arrived at this renaissance in current developmental thinking about adolescence. The highlighted studies come from functional magnetic resonance imaging (fMRI) research, as this method is best suited to examine functional interactions among brain regions and networks. A glossary of terms is provided in Box 1. The first section focuses on advances made in understanding of the development brain networks and puberty across adolescence. The second section examines the dynamic and integrative behaviors that characterize adolescence and the neural systems that support them. The closing section offers opportunities for continued advancements in policies and programs that build on the rich literature to date by adopting a dynamic systems framework and positive view of adolescence.

Box 1. Glossary of Terms

This glossary includes brief descriptions of the most frequently referenced brain regions and terms that may be unfamiliar to nonneuroscientists. It is not exhaustive and for all brain regions there are numerous other functions not described here.

• Amygdala: a cluster of neurons in the medial temporal lobes of the brain that is implicated in emotion processing and learning. It is also associated with emotional memory and monitoring affective information.

• Cingulate cortex: this brain region is located in the medial portion of the cerebral cortex. It receives and sends inputs from numerous brain regions and has been implicated in a range of behaviors, including conflict resolution, decision-making, and emotional processing.

• Default mode network: a large-scale brain network primarily composed of the medial prefrontal cortex, posterior cingulate cortex/precuneus and angular gyrus. Studies show robust activation of this network in the absence of external stimuli, during “wakeful rest” such as daydreaming, or concentrated thinking or planning.

• Diffusion tensor imaging (DTI): a neuroimaging tool used to examine white matter tracts in the brain that relies of diffusion of water molecules.

• Dorsolateral prefrontal cortex: a brain region that undergoes protracted development in humans. It has been implicated in many executive functions, including goal-oriented behavior, decision-making, working memory, and cognitive flexibility.

• Face processing network: a network of brain regions, including the fusiform gyrus, superior temporal sulcus, and occipital face area, responsive to distinct aspects of processing and recognizing faces.

• Graph theory: a method of describing the relationships between two or more objects. When applied to neuroimaging data, the objects are individual voxels or clusters of voxels derived from fMRI studies. It is broadly used to explain how the brain is organized.

• Grey matter: brain tissue that consists of neuronal cell bodies

• Hippocampus: a region that plays central roles in memory and learning

• Insula (insular cortex): a region in the cerebral cortex found within the lateral sulcus that displays activation in response to various stimuli (e.g., emotional images) and behaviors (e.g., decision-making, risk-taking)

• Machine learning: the implementation of computer algorithms that use data to identify relationships among multivariate features (e.g., brain activation) and to make predictions about brain states, different populations of research participants, or brain-behavior relationships.

• Network module: a group of densely interconnected nodes, which often are the basis for specialized subunits of information processing.

• Nucleus accumbens (ventral striatum): a brain region that exhibits significant functional development during adolescence, particularly as related to reward, reinforcement learning, motivation, risk-taking, and salience

• Posterior superior temporal sulcus (pSTS): primary functions of this region include social cognition, perspective-taking, processing of biological motion, faces, voices, and language.

• Precuneus: a brain region it is involved with episodic memory, visuospatial processing, reflections upon self, and aspects of consciousness.

• Reinforcement learning: a learning system that adapts future responses in response to feedback from the environment.

• Resting state connectivity: represents the temporal coherence of functional activation within or between regions or networks during period free from external stimuli (“at rest”).

• Social brain network: this network includes brain regions, including the amygdala, pSTS, TPJ and others, that work in concert to represent social behavior and social information processing.

• Somatosensory cortex: a region included in the “social brain” network that is primarily implicated in balance, body position, and touch.

• Temporoparietal junction: a region included in the “social brain” network that is primarily implicated in social cognition such as perspective-taking and empathy.

• Ventromedial prefrontal cortex: this region is located in the frontal lobe at the bottom of the cerebral hemispheres and is implicated in the processing of risk and fear, emotion regulation, and social cognition.

Part I: Trajectories of Brain Development

The brain is a complex and dynamic functional system, characterized by constant activity and change (Goldenberg & Galván, 2015). Billions of neurons form intricate patterns that can flexibly merge based on shared function, forming networks that are constrained by, but not limited to, direct structural connections of the brain (Vincent et al., 2007). In early adolescent brain research, there was a focus on understanding how individual brain regions supported and interacted with psychological domains but in the past decade there have been methodological and conceptual strides in the understanding of the developing brain through the use of neural connectivity and functional network approaches. These approaches start with the premise that brain regions and the networks they contribute to do not work in isolation—that is, no one brain region or network governs any one psychological process and therefore there is no “reward center,” for example. More specifically, neural connectivity refers to the anatomical and functional connections among brain regions, whereas functional networks refer to the extent to which brain regions/systems coactivate (“work together”) to enact a particular output.

Functional networks have the capacity to support complex thought and action, more so than any single element of the system would be able to support alone. This fact is consonant with the principles of dynamic systems theory (Smith & Thelen, 2003), which states that development can only be understood as the multiple, mutual, and continuous interaction of all levels of the developing system. With advances in neuroimaging and machine learning, as applied to brain data, in recent years scientists have gained traction on understanding the topology of functional networks (Baum et al., 2017; Cui et al., 2020; Dosenbach et al., 2010; Fair et al., 2007, 2008, 2009). Functional connectivity and neural network tools measure the temporal correlation between remote neurophysiological events (Sporns, Honey, & Kötter, 2007) and can identify the influence one neural system exerts over another (Friston, 2009).

Detailed review of the various connectivity techniques has been described elsewhere (e.g., Goldenberg & Galván, 2015 for greater detail and discussion of limitations). However, a few popular approaches are highlighted. The first method, resting-state connectivity, is acquired during fMRI and is used to examine intrinsic functional connectivity among brain regions. Brain regions that often work together form a functional network with a high level of ongoing, strongly correlated spontaneous neuronal activity, without the presence of a task or stimulus (Fox & Raichle, 2007). Resting-state provides a method with which to measure connectivity by examining the level of coactivation between the functional time-series of brain regions during rest (Biswal, Zerrin Yetkin, Haughton, & Hyde, 1995). These patterns of resting-state correlations are hypothesized to reflect the stable and intrinsic functional architecture of the brain (Buckner et al., 2009). Another tool, graph theory, has been extensively used to examine complex systems and reveal landscapes about the local and global organization of functional brain networks (Bullmore & Sporns, 2009). Functional networks can be described as graphs that are composed of nodes (i.e., brain regions) that are linked by edges, representing functional connectivity (i.e., correlations between time series) (Bullmore & Sporns, 2009). Neuroimaging research applying graph theory on resting-state data has revealed small-world architecture of functional brain networks across development (Bassett & Bullmore, 2006; Power, Barnes, Snyder, Schlaggar, & Petersen, 2012).

Highlights of these approaches have yielded insights into: the brain’s default mode network (Fair et al., 2008), the development of distinct control networks through segregation and integration (Baum et al., 2017; Cui et al., 2020; Fair et al., 2007) and the developmental process that proceed from a local to a distributed organization (Fair et al., 2009). Attempts to predict individual variation of brain maturation have also relied on connectivity techniques (Dosenbach et al., 2010). Growing understanding of the brain as an interconnected system is based on research showing that across age, simpler networks are organized into increasingly complex networks (Power, Fair, Schlaggar, & Petersen, 2010), as based on environmental inputs and experience.

Dynamics of Brain Network Development

One of the most novel contributions to adolescent brain development in recent years has been the identification of how large-scale brain networks dynamically integrate with and segregate from one another to support cognitive and behavioral development (Cole, Bassett, Power, Braver, & Petersen, 2014). Functional neuroimaging studies have demonstrated that the human brain has a well-defined modular organization, as reflected in the presence of large-scale functional networks (Biswal et al., 1995; Power et al., 2010). During childhood and adolescence, functional modules become more distinct: connectivity within modules increases while connectivity between modules declines (Dosenbach et al., 2010; Gu et al., 2015; Satterthwaite et al., 2013). Such development allows for functional specialization, reducing interference among systems, and facilitating cognitive performance. The identification of how development of structural brain development supports the maturation and refinement of these functional modules has provided insight into the notable advances that occur in cognitive development during adolescence. A few key findings have emerged.

First, using functional MRI data and diffusion-weighted imaging, we have learned that structure–function coupling undergoes massive remodeling during adolescence that supports age-related improvements in cognitive maturation (Baum et al., 2020). Data from a sample of over 700 participants aged 8 to 23 years was used to examine the degree to which a brain region’s structural connections support coordinated fluctuations in neural activity. Results revealed that brain regions exhibiting high structure-functional coupling were localized in segregated regions of primary sensory and medial prefrontal cortex (Baum et al., 2020). These findings provide greater understanding of how brain “architecture develops during adolescence to support coordinated neural activity underlying executive processing” (Baum et al., 2020) and suggests that the individual variation in structure-functional coupling (and in the timeline of the development of structure-functional coupling) contributes to individual differences in executive functioning skills.

Second, age-related changes within structural brain networks strengthen the global efficiency of the brain and support cognitive performance (Baum et al., 2017). These changes include the increasing segregation of structural network within-modules and increased coordination between modules, which is essential for dynamic processing if cognitive inputs. These changes are associated with improvements in executive functioning across childhood and adolescence because they reflect functional specialization. For example, associations between module segregation and cognition were domain-specific: segregation of the default mode mediated age-related improvements in social cognition, which is reliant on regions within that network (Baum et al., 2017). The process of structural network segregation may allow for functional specialization and reduce competitive interference between brain systems (Fornito, Harrison, Zalesky, & Simons, 2012). More broadly, these findings are important because they provide greater depth of understanding of the type of changes that occur during adolescence that impact behavioral maturation.

Third, using machine learning approaches in large samples of youth, researchers have found support for dynamic systems theories (Smith & Thelan, 2003): “cognitive systems closely interact with one another in transient processes that collectively produce the complex landscape of brain dynamics that supports cognition” (Chai et al., 2017; Cui et al., 2020). In other words, the distinct cognitive systems that have been identified in the brain (e.g., attention, executive, sensorimotor) draw on collaborative coupling, whether locally or distributed across the brain, to maximize cognitive performance. This also highlights the importance of flexibility among these networks and supports the notion that dynamic responding to the environment (and within brain systems) is a key element to yielding a mature state of the brain.

Fourth, individual variation in the development of functional topography of brain networks is associated with individual differences in executive functions across early to late adolescence. Using machine learning on fMRI data collected in a large sample (n = 693) of individuals aged 8–23 years and building on findings in adults (Braga & Buckner, 2017), Cui et al. (2020) report that the distributed regions and systems that act in concert to support executive tasks exhibit remarkable inter-individual variability in development. They report that at any given age, a greater cortical representation of control networks is associated with improved executive performance (Cui et al., 2020). This finding helps explain why, despite similarities in age, some children and young adolescents exhibit more mature executive functioning than others, particularly in the domains of memory, inhibitory control, mental flexibility, reasoning, and motor speed.

Together, these findings underscore a few principles of brain development that have found empirical support in the past decade: (1) the importance of individual differences across brain development, (2) the dynamic and interactive nature of brain systems, and (3) continued development and refinement of the brain through early adulthood has implications for integrative behaviors.

Part II: The Dynamics Between Puberty and Brain Development

In early adolescence, the release of pubertal hormones initiates the process of sexual maturation, yielding significant physical, psychological, social, and neurobiological changes. The multiple stages of puberty are beyond the scope of this review but greater detail (Vijayakumar, Op de Macks, Shirtcliff, & Pfeifer, 2018) and the challenges in measuring puberty in humans (Dorn, ) are provided elsewhere. Here, the focus is on the complex relationship between brain development and puberty. Although there is agreement that puberty and concurrent brain development are intricately linked, the numerous inconsistencies among studies indicate that more research in this area is needed. In particular, studies that have greater consistency in study design, measurement of puberty, and which factors that are statistically accounted for are warranted. For example, questions remain regarding the correspondence between physical assays of puberty and hormonal measures (Shirtcliff, Dahl, & Pollak, 2009). The variability in precision of pubertal measurement is also a concern (Dorn & Biro, 2011; Dorn & Dahl, 2006). Despite the lack of consensus and limitations, there are some noteworthy findings that represent the important relationship between the two developmental events (puberty and brain development) that are characteristic of adolescence.

Structural Brain Development

Many studies have identified negative associations between global grey matter volume and pubertal stage and gonadal hormone levels (Bramen et al., 2011; Paus et al., 2010; Peper et al., 2009; Pfefferbaum et al., 2015), which means that as puberty progresses, there is a decrease in grey matter (comprised of neuronal cell bodies) across the brain. This pattern is consistent with research examining the relation between grey matter and age (Mills et al., 2016; Vijayakumar et al., 2016) but it is important to note that there do exist some inconsistencies across studies in this finding based on which factors are included in statistical models.

In a sample of over 280 4–22-year-olds, Nguyen et al. (2012) found significant negative associations between testosterone and cortical thickness1 in the cingulate, precuneus, and dorsolateral prefrontal cortex in males and right somatosensory cortex in females (Nguyen et al., 2012), which may be driven by testosterone (Vijaykumar et al., 2018). Conversely, the authors found a positive relation between dehydroepiandrosterone (DHEA), a pubertal hormone that is a biological precursor of testosterone, and cortical thickness in the dorsolateral prefrontal cortex and temporoparietal junction (TPJ), regions implicated in executive functions and social cognition, respectively. It is unclear why these relationships were observed in these specific regions but may be related to the significant plasticity of these regions during adolescence and their sensitivity to environmental input. Estradiol has also been found to be associated with decreased thickness in the left middle temporal cortex in females (Herting, Gautam, Spielberg, Dahl, & Sowell, 2015). However, a different study did not find any associations between grey matter development and changes in testosterone or estradiol over the span of 3 years in which youth were examined (Brouwer et al., 2015).

Research on subcortical brain regions that are rich in pubertal hormone receptors and which undergo substantial structural development during adolescence (Mills et al., 2016) has also been of great interest. The reason these regions, including the amygdala, hippocampus, and nucleus accumbens,2 have been extensively researched is because of their importance in socioemotional and reward processing. The amygdala, a key circuit in emotion processing, learning, and social development, evinces interesting sex differences as related to pubertal hormones. Whereas amygdala volume is positively associated with testosterone (Bramen et al., 2011), and hair and skin changes (Hu, Pruessner, Coupé, & Collins, 2013) in males, it is negatively associated with testosterone (Bramen et al., 2011) and breast development (Blanton et al., 2012; Hu et al., 2013) in females. Research on the hippocampus and nucleus accumbens is less conclusive. Although there is some evidence for a decrease in hippocampal volume with increasing pubertal stage (Blanton et al., 2012; Neufang et al., 2009), other studies have found opposite or inconclusive results (Hu et al., 2013). The relationship between estradiol and amygdala and hippocampal volume is also unclear. Most studies on the nucleus accumbens as it relates to puberty find minimal support for a relationship (Brouwer et al., Koolschijn, Peper, & Crone, 2014; Neufang et al., 2009; Peper et al., 2009). However, Uros\̌ević, Collins, Muetzel, Lim, and Luciana (2014) reported smaller accumbens volume in pubertally mature females and larger volumes in more mature males (Uros\̌ević et al., 2014), which may be explained by earlier sexual maturation in females.

Longitudinal research finds a positive association between pubertal development as assessed with Tanner staging, and amygdala volume (Goddings et al., 2014), along with significant interactions between age, testosterone levels, and sex (Herting et al., 2014). Herting and colleagues report that more pubertally mature boys and those with higher testosterone levels exhibit greater amygdala volume, whereas the opposite is true for girls, with girls who are more pubertally mature have smaller amygdala volumes than those who are less pubertally mature (Herting et al., 2014). The authors note that these findings mirror those observed in animal models, which have shown an increase in new cell growth in the amygdala in male rats compared with females during puberty (Ahmed et al., 2008) and that estrogen increases dendritic spine density in the medial amygdala in female rats (de Castilhos, Forti, Achaval, & Rasia-Filho, 2008). Collectively, these results suggest that estradiol and testosterone may contribute to sexually dimorphic changes in amygdala volumes across adolescence and into adulthood in both human and animal models (Herting et al., 2014) that have implications for sexually dimorphic behaviors as well.

The brain’s white matter also undergoes significant development during adolescence and there is some evidence that this maturation coincides with puberty-related changes. In general, there is trend showing that increased white matter, which is comprised of myelination and axons, is associated with greater pubertal maturation (Chavarria, Sanchez, Chou, Thompson, & Luders, 2014; Paus et al., 2010; Perrin et al., 2008; Pfefferbaum et al., 2015). Using Diffusion Tensor Imaging (DTI), a commonly used technique in cognitive neuroscience to assay myelination, studies have reported greater white matter volume in the internal capsule, temporal and frontal lobes with increasing pubertal maturation (Herting et al., 2012, 2014) and testosterone levels in boys (Menzies, Goddings, Whitaker, Blakemore, & Viner, 2015). In girls, there is a negative association between white matter development and physical pubertal development and estradiol (Herting et al., 2012). It is unclear if these changes are due to bidirectional influence of white matter and puberty or to unexamined factors that influence both white matter and puberty.

In general, the majority of findings showing a negative association between puberty and grey matter (cortical thickness) and a positive association between puberty and white matter fall along the same lines as those observed with age. However, these effects remain even after controlling for age, suggesting that the effects of puberty on brain development are distinct from ontological ones. Furthermore, this suggests that sex hormones may have a particularly unique role in contributing to brain development, particularly as it relates to brain function.

Functional Brain Development

In the past decade, there has been a sharp increase in the number of studies examining whether pubertal development relates to how the brain functions. Although there is significant variation in the level of precision used to assay puberty (i.e., ranging from self-reported markers of pubertal maturation to hormone assessments), some trends have emerged. A thorough explication of these findings can be found in a review by Vijaykumar et al. (2018). An abbreviated summary follows of the reward processing and socioemotional development as related to puberty as these have been the focus of the majority of studies.

In studies of reward processing, some investigations have found increases in activation of regions typically associated with reward (e.g., ventral striatum, orbitofrontal cortex) with increasing pubertal maturation as assayed either using physical development (Braams, van Duijvenvoorde, Peper, & Crone, 2015) or hormone levels (LeMoult, Colich, Sherdell, Hamilton, & Gotlib, 2015; Op de Macks et al., 2011, 2016). Others have found less activation in the frontal cortex (Morgan et al., 2013) and the caudate (Forbes et al., 2010). Estradiol has been found to be positively associated with activation in frontostriatal regions during reward processing (Op de Macks et al., 2011, 2016). Other studies have reported no effects of reward-related brain activation and puberty when using self-reported measures of pubertal stage (van Duijvenvoorde et al., 2014; Op de Macks et al., 2016).

Investigations into the relation between social-affective processes and puberty have primarily focused on the amygdala. Using emotional faces as stimuli, these studies find that more pubertally-mature adolescents exhibit less amygdala activation than less pubertally-mature adolescents to emotionally neutral faces (Ferri, Bress, Eaton, & Proudfit, 2014; Forbes, Phillips, Silk, Ryan, & Dahl, 2011) but more activation to angry faces (Forbes et al., 2011). A separate study found a positive association between advancing pubertal stage and activation in the amygdala and ventral striatum to opposite-sex faces in a sample of 9–16 year olds (Telzer et al., 2015). In terms of hormones, neural responses in the anterior cingulate cortex to emotional conflict were positively associated with estradiol but responses in the striatum and frontal gyrus were negatively associated with testosterone in 10–15 year old males (Cservenka, Stroup, Etkin, & Nagel, 2015). Tyborowska, Volman, Smeekens, Toni, and Roelofs (2016) reported a negative association between testosterone levels and amygdala activation in 14-year-olds (Tyborowska et al., 2016). Findings from longitudinal work indicate that there is an increase in amygdala and ventral striatal activation (Spielberg, Olino, Forbes, & Dahl, 2014), as well as reduced amygdala-orbitofrontal functional coupling (Spielberg et al., 2015), to angry and fearful faces with increases in testosterone in 11–15-year-olds. A similar positive association has been observed between testosterone and activation in the anterior temporal cortex during mentalizing tasks (Goddings, Burnett Heyes, Bird, Viner, & Blakemore, 2012).

In sum, this research on the biological changes that emerge during adolescence highlight: the dynamic nature of pubertal and brain development with bidirectional influences one another; that puberty is related to both structural and functional brain development; and the need for greater rigor in measuring both puberty and brain outcomes.

Part III: Dynamic and Integrative Behaviors in Adolescence

Along with significant physical changes in the body and neurobiological maturation in the brain, there are notable changes in adolescent behavior. Consistent with dynamic systems theory, psychological domains co-development with one another dynamically, constantly changing in response to the individual’s behavior, experience and ontogenetic brain maturation, as well as a reaction to external inputs. This section highlights findings from studies of reward sensitivity, learning and decision-making, and social behavior, because these developing psychological domains: (1) exhibit the greatest change during adolescence (relative to other behaviors), (2) represent a unique shared experience among most typically developing adolescents relative to younger children or adult, (3) have implications for policies and programs aimed at helping and honoring young people, and (4) are those that have generated the majority of adolescent brain research so a focus on these domains best reflects the existing scholarship. Although typically investigated in isolation, these behaviors share neural mechanisms and it is in part through these shared brain systems that they are behaviorally integrated.

Adolescent brain research over the last decade has helped identify the neural regions and systems that support the development of these psychological transformations, building on previous research that determined the brain development that supports cognitive regulation (Casey, Jones, & Somerville, 2011; Rosenbaum & Hartley, 2019). This section reviews current brain research on reward sensitivity, learning and decision-making, and social behavior.3 Although their development is dynamically integrated, the discussion of each in separate subsections reflects the isolated manner in which they have been studied. However, the reader will note how the studies highlighted below illustrate the integration of multiple brain systems. Box 2 provides a discussion of efforts to study the developing brain in context by creating ecologically relevant assays to test these behaviors.

Figure 1 Open in figure viewer PowerPoint

Neurosynth, an online tool for large-scale, automated synthesis of functional magnetic resonance imaging (fMRI) data, was used to identify regions central to reward processing and reinforcement (n = 922 studies): (a) ventral striatum (crosshairs at x = −10, y = 10, z = −6); (b) hippocampus and amygdala (crosshairs at x = 24, y = −12, z = −20); (c) medial orbitofrontal cortex (crosshairs at x = 0, y = 46, z = −12). [A, anterior; P, posterior].

Figure 2 Open in figure viewer PowerPoint

Neurosynth was used to identify regions central to processing social behavior (n = 1302 studies): (a) temporal gyrus and amygdala (crosshairs at x = 24, y = 0, z = −20); (b) medial frontal cortex (crosshairs at x = 4, y = 50, z = −16); (c) temporoparietal junction (TPJ) (crosshairs at x = 52, y = −46, z = 8). [A, anterior; P, posterior].

Box 2. Ecological Validity

A big contribution to adolescent neuroscience research has been the efforts to study the developing brain in context. Development clearly does not occur in a vacuum, and the very behaviors germane to adolescent development occur in the context of relationships, social interactions, and systems. For example, risk-taking is a very social behavior, particularly during adolescence, when thrill-seeking often occurs with friends. This poses a methodological challenge for researchers eager to uncover the neurobiological response to peer-influenced risk-taking. To circumvent this issue, a research team led by Laurence Steinberg and Jason Chein created a clever fMRI experiment in which participants played a risk-taking video game in the presence of their peers while undergoing fMRI (Chein, Albert, O’Brien, Uckert, & Steinberg, 2011). The risky component was designed to mimic real-life driving circumstances in which each intersection contains a stoplight that turns yellow as the car approaches and participants must decide whether to make a risky choice by running the yellow light or make the nonrisky choice of stepping on the brakes (and thus incurring extra time to get to the finish line). Each participant played the game either alone or in the presence of two same-aged, same-sex friends. College students and adults exhibited the same behavior (that is, made the same number of risky and nonrisky choices) whether or not there was a peer watching them but adolescents made significantly more risky choices when a peer was watching than when they were alone and showed differential brain responses as well (Chein et al., 2011). In addition to the findings, this study was seminal in demonstrating the feasibility of conducting an fMRI study that better approximated how adolescents make decisions “in real life”—in the presence of peers—than previous studies had done. This ecologically-informed approach has since been adopted in many other studies (e.g., Defoe, Dubas, Dalmaijer, & van Aken, 2020; van Hoorn, McCormick, Rogers, Ivory, & Telzer, 2018; Op de Macks et al., 2018) and forced the field to call into question whether the standard imaging approaches whereby participants are asked to complete cognitive tasks alone were actually capturing neural response to more ecologically-relevant behaviors such as risk-taking.

Reinforcement Learning and Decision-Making

The integration of multiple processes, including reward sensitivity, valuation, decision-making, and exploration, as well as sensory and motor information, and the brain systems that support them, is central to reinforcement learning (Figure 1). Research in the past decade has revealed that each of these processes, and the neural systems that support them, are exceptionally dynamic during adolescence. As such, reinforcement learning is a quintessential behavior that embodies the dynamic and integrative nature of adolescent brain development.

The quest to understand the very basic aspect of reinforcement learning—reward sensitivity—in the adolescent brain began over a decade ago. Early studies consistently found that adolescents exhibited greater activation in the ventral striatum as compared with children and adults when receiving a reward (Ernst et al., 2005; Galván et al., 2006; van Leijenhorst et al., 2009; May et al., 2004) (but see Bjork et al., 2004 for research showing that adolescents have less engagement of the ventral striatum as compared with adults); this finding has been replicated in more recent research with larger sample sizes (Braams et al., 2015) and been found to be associated with pubertal development (Uros\̌ević et al., 2014). This adolescent-unique profile is consistent with current understanding of plasticity in the brain regions (e.g., striatum) and neurotransmitter systems (e.g., dopamine) central to reinforcement learning during adolescence (Galván, 2010). Evidence from rodent, nonhuman primates, and human adolescents shows that there are developmental differences in dopaminergic innervation of the striatum that peaks in adolescence and declines into adulthood (Galván, 2010). Using a novel magnetic resonance tool, R2′, a measure of tissue iron which colocalizes with dopamine vesicles and is necessary for dopamine synthesis, Larsen et al. (2020) confirmed that dopamine availability increases through the adolescent period before stabilizing by early-adulthood (Larsen et al., 2020). This methodological advance in adolescent neuroscience research is significant because it provides converging evidence with the fMRI studies to support the notion that in addition to reward sensitivity, dopamine plays a key role in feedback and reinforcement learning during adolescence.

The hyper-excitability of the dopamine-rich striatum in adolescence was initially interpreted rather narrowly as evidence for “poor decision-making” and impulsivity in adolescents. Fortunately, evidence from several recent studies has demonstrated the advantages of this adolescent-unique phenotype (Telzer, 2016). A series of both cross-sectional (Cohen et al., 2010; Davidow, Foerde, Galván, & Shohamy, 2016) and longitudinal (McCormick, Peters, Crone, & Telzer, 2021; Peters & Crone, 2017) studies using feedback-based learning tasks have discovered that increased activation in the striatum benefits learning in adolescents as compared with adults (Davidow et al., 2016). Research also suggests it benefits other aspects of cognition, as adolescents who showed the greatest ventral striatum activation during a difficult working memory task had the highest memory performance (Satterthwaite et al., 2012). These findings are noteworthy because they challenge earlier notions that enhanced striatal activity in adolescents serves maladaptive purposes and instead demonstrate how it is adaptive for learning. Together with previous work showing the continued development of frontoparietal regions that also support feedback-based learning (Peters, Braams, Raijmakers, Koolschijn, & Crone, 2014; Peters, Van Duijvenvoorde, Koolschijn, & Crone, 2016), the results demonstrate (1) that complex learned behaviors involve the integrated and dynamic coordination of distributed brain circuits (Gerraty et al., 2018), and (2) adolescence may be a unique life phase for increased feedback-learning performance, particularly when it interacts with prior skill or experience (McCormick et al., 2021).

Recent research in the past decade has built on the reward sensitivity findings to better understand how the brain supports valuation and decision-making in the context of reinforcement learning. Reinforcement learning refers to the mechanisms by which future decisions are informed by the ability to associate previous actions with outcomes and the valuation assigned to such outcomes (Raab & Hartley, 2018). In a reinforcement learning task, participants use feedback over many trials to associate choices with probable outcomes (Shohamy et al., 2004). Similar to adults, a network of cortical and subcortical structures play distinct but intertwined roles in the affective and cognitive processing involved in learning from reinforcement (O’Doherty, 2016). In this process, the striatum has long been thought to serve an integrative role (Alexander, DeLong, & Strick, 1986), given the many inputs it receives from cortical areas and the projections it sends back to motor cortex, rendering it anatomically positioned for integration (Haber & Knutson, 2010). Research in adults also shows that increased dynamic connectivity between the striatum and large-scale circuits is associated with learning performance and suggests that network coordination centered on the striatum underlies the brain's ability to learn to associate values with sensory cues (Gerraty et al., 2018). What is distinct from adults is that these networks, which also include the hippocampus, anterior cingulate cortex, medial cortical regions, ventromedial prefrontal cortex (vmPFC), orbitofrontal cortex (OFC), and associated regions implicated in sensory processing and motor output (Haber & Behrens, 2014); see DePasque & Galván, 2017 for a review) are undergoing continued refinement during adolescence. Research also shows that there is continued structural maturation of the hippocampus and its connections to the prefrontal cortex, as well as between the amygdala and PFC throughout adolescence (Tottenham & Galván, 2016). Integration between the striatum and PFC also helps facilitate value-based cognitive control (Davidow et al., 2016) and predicts developmental changes in reinforcement learning (van den Bos, Cohen, Kahnt, & Crone, 2012).

A greater understanding of the dynamics of reinforcement learning has simultaneously led to insight into adolescent decision-making (see Nussenbaum & Hartley, 2019 for a thorough review on this topic). Borrowing from reinforcement learning models in adults and using a variety of reinforcement learning tasks, research in adolescents has uncovered how the adolescent brain uses past positive and negative experiences to guide behavior (Rosenbaum & Hartley, 2019). Although there are nuances to the findings and interpretations, including heterogeneity of the task structures and reward probabilities, some general patterns have emerged. First, data suggest that individuals improve with age in their ability to advantageously integrate recent outcomes into their estimates of the value of different decisions (Nussenbaum & Hartley, 2019); in other words, there is evidence for increasingly improved value updating with age. Second, children and adolescents are more likely than adults to switch responses and “test out” new response options (i.e., explore their decision options; Christakou et al., 2013; Decker, Otto, Daw, & Hartley, 2016) based on the benefits of strategic exploration (Somerville et al., 2017) and perhaps based on interest in continuously discovering new rewarding options (Javadi, Schmidt, & Smolka, 2014).

Given the relevance for public health (Centers for Disease Control & Prevention, 2020) and uptick in prevalence of risk-taking during adolescence (Steinberg, 2011), the majority of decision-making studies in adolescents have focused on the neural correlates that support risk-taking, defined by behavioral economists as choosing the option with the highest outcome variability (Figner & Weber, 2011). A comprehensive survey of this research, including the many parameters that have been examined from a neuroeconomics framework such as value, utility, risk, ambiguity, probability structures, and affect, calls for a separate review paper (but see Rosenbaum & Hartley, 2019 and Van Hoorn, Shablack, Lindquist, & Telzer, 2019 for a meta-analysis of neuroimaging studies that examine adolescent decision-making vis a vis the social context). Here, converging themes are highlighted.

First, risk-taking behavior varies as a function of affective context (Casey, 2015; Crone & Dahl, 2012; Duckworth & Steinberg, 2015). A meta-analysis by Defoe, Dubas, Figner, and van Aken (2015) revealed that risk-taking tendencies in adolescents change as function of context: whereas adolescents take more risks than adults on tasks with immediate outcome feedback on rewards and losses, they take fewer risks than children on tasks with a sure/safe option (Defoe et al., 2015), which may explain the finding that in studies examining risk-taking in adolescents and adults, over half do not report greater risk-taking in adolescents (Rosenbaum, Venkatraman, Steinberg, & Chein, 2017). Dual systems frameworks, which focus on the interactions between cognitive and affective networks, posit that under affectively laden situations, adolescents will engage in greater risk-taking compared with children and adults but in affectively neutral contexts they will make less risky choices than adults (Casey et al., 2011; Steinberg, 2008). This framework has been refined over time to be inclusive of other factors that contribute to risk-taking such as flexibility, individual differences, and motivational states (Crone & Dahl, 2012; Pfeifer & Allen, 2012).

Second, findings across neuroscientific studies consistently report that the dynamics of frontostriatal network development, along with changes in pubertal status (Peper & Dahl, 2013) relate to risk-taking (Peper, Koolschijn, & Crone, 2013). When making a risky choice, adolescents exhibit heightened activation in the ventromedial prefrontal cortex (van Duijvenvoorde et al., 2015; van Leijenhorst et al., 2010), anterior insula (Smith, Steinberg, & Chein, 2014), and dorsomedial prefrontal cortex (van Duijvenvoorde et al., 2015) in addition to the ventral striatum (Braams et al., 2015; Chein et al., 2011), as compared with adults. The distributed network of regions that are associated with risk-taking in adolescents suggests that the motivation to take a risk is influenced by more than simply the potential rewards to be gained (as would be expected if only the ventral striatum was associated with risk-taking), but rather by a host of factors, including potential loss (Barkley-Levenson, Van Leijenhorst, & Galván, 2013), physiological signals (Smith, Steinberg, et al., 2014), cognitive appraisals (van Duijvenvoorde et al., 2015), and stress (Uy & Galván, 2017). This is noteworthy because it challenges modular views of the brain and instead provides evidence that no one brain region works in isolation, and that behaviors are the product of multiple, integrated systems.

Third, there are vast individual differences in risk-taking among adolescents. Several studies have found that greater activation in the ventral striatum to rewards is associated with increased risk-taking propensity in adolescents (van Duijvenvoorde et al., 2014; Galván, Hare, Voss, Glover, & Casey, 2007; Telzer, Fuligni, Lieberman, & Galván, 2013). However, other studies find declines in risk-taking as a function of greater activation in the ventral striatum but this varies based on task. For example, adolescents who showed the greatest ventral striatum activation when being prosocial (Telzer et al., 2013; Telzer, Fuligni, Lieberman, & Galván, 2014), when processing emotions (Pfeifer et al., 2011) or were the most influenced by their prosocial peers (Cascio et al., 2014) showed the greatest declines in risk taking. In addition, individuals who expressed the greatest distress over societal circumstances exhibited the greatest decline in depressive symptoms over time if they had high neural response to reward in the ventral striatum (Tashjian & Galván, 2018), suggesting that the ventral striatum can serve a protective effect in buffering against internalizing symptoms. Individual differences in brain connectivity-behavior associations have also been observed. For example, neural connectivity evinces dynamic patterns as based on affective context. Utilizing multivariate methodology and two distinct MRI measurements (i.e., function and structure), one study manipulated emotional context with affective cues. They found that neural connectivity exhibited less mature brain states under high-emotions states, which was associated with greater risk preference in adolescents and young adults (Rudolph et al., 2017). In a study examining risk-taking under self-reported high-stress conditions (as indexed with daily diary), multilevel logistic regression analyses revealed that although all participants were more likely to take risks as expected reward value increased, this behavior was greater under high versus low stress for individuals with low accumbofrontal tract integrity (i.e., less connectivity between the ventral striatum and frontal cortex; Uy & Galván, 2010). Results suggest that individual differences in brain structure and function are just as germane to characterizing risky decisions in adolescents as ontogeny.

Social Behavior

Maturational changes in the network commonly referred to as “the social brain” (Kilford, Garrett, & Blakemore, 2016) is paralleled by shifts in social behavior during adolescence. Using computer tasks designed to capture the social cognitive processes that underlie social behavior, research has revealed how increasing integration and refinement of regions in the social brain, including the amygdala, inferior frontal gyrus, dorsomedial prefrontal cortex (dmPFC), and regions in the temporal cortex, have direct implications for the increasing sophistication and nuance observed in adolescents’ ability to navigate increasingly complex social interactions. Although relationships with caregivers, teachers, and relatives remain important (Guassi Moreira, Tashjian, Galván, & Silvers, 2018), there is a unique fascination with same-aged peers. As such, prevailing studies on social behavior in adolescents have focused on neural responses to social rewards, social rejection, and prosocial behavior.

Forming and maintaining friendships is paramount to the adolescent experience. Numerous studies show that hyper-responsivity in mesolimbic brain circuity among adolescents (as compared with children and adults) to monetary reward (Galván et al., 2006) is also evident in response to social images (images of friends, peers, or other social rewards (Chein et al., 2011). This is unsurprising given that adolescents place high value on friendships and peer acceptance and this circuitry is crucial for monitoring value (O’Doherty, 2016). However, there is also extensive variability in friendship stability and quality that the ventral striatum also tracks, such that those with more stable friendships across adolescence show greater activation in the ventral striatum as compared with children and adults and as compared with those with less stable friendships (Schreuders, Braams, Crone, & Güroğlu, 2021). There is also evidence that cortical thickness in regions commonly associated with social processing, including the medial PFC, posterior superior temporal sulcus (pSTS), and TPJ is also greater among those who report stronger friendship quality (Becht et al., 2021). This suggests that individual differences in both the structure and function of the social brain contribute (or result from) normative variation in friendship quality.

One way to maintain and grow friendships is to exhibit prosocial behavior and, indeed, there is evidence that prosocial behavior becomes increasingly sophisticated during adolescence, with a greater focus on peers and close others as recipients of prosocial actions. Studies show that when tasked with making prosocial decision in fMRI studies, adolescents predictably make the most prosocial choices when the recipient is a friend versus when it is a disliked peer. This preference is accompanied by greater activation in mesolimbic circuitry (i.e., putamen) and regions of the social brain that are involved in social value and judgment (Schreuders, Smeekens, Cillessen, & Güroğlu, 2019). Interestingly, studies that presented adolescents with prosocial scenes find robust engagement of TPJ that varies as a function of self-reported empathy (Overgaauw, Güroğlu, Rieffe, & Crone, 2014) and charitable giving behavior (Tashjian, Weissman, Guyer, & Galván, 2018). There is also evidence that levels of testosterone and cortisol impact prosocial behavior, such that adolescents with high levels of testosterone and low cortisol donate to charity to a greater extent when viewing others doing so, and exhibit greater activation in the TPJ, striatum, and orbitofrontal cortex (Duell, van Hoorn, McCormick, Prinstein, & Telzer, 2021). This is noteworthy because it provides empirical evidence for the notion that, during adolescence, some of the most considerable changes in the brain are associated with some of the most significant changes in behavior. So, while the particular neural changes are important as they relate to specific psychological domains, the broader meaning of this work is that brain development during adolescence is systematic and relevant for the behaviors that support the transition into adulthood.

Perhaps because of the intense focus on forming social bonds, social exclusion is particularly painful during adolescence. Numerous studies have examined the neural correlates of both the experience of social exclusion (as simulated in fMRI tasks) and of being chronically excluded in real life. Chronically rejected adolescents exhibit greater activation in the cingulate cortex (Will, van Lier, Crone, & Güroğlu, 2016), a region activated during social rejection and distress (Cacioppo et al., 2013; Eisenberger, 2012), when experiencing social exclusion. They also show robust engagement of regions typically associated with cognitive control when they refrained from punishing a peer who excluded them (Will, van Lier, et al., 2016). Longitudinal work in adolescents finds that greater activation in the cingulate cortex during exclusion is associated with increases in depressive symptoms one year later (Masten et al., 2011) but that greater time spent with friends (as assessed using a daily diary) related to less activation in this region and in the insula—regions previously linked to negative affect and pain processing—during an experience of peer rejection two years later (Masten, Telzer, Fuligni, Lieberman, & Eisenberger, 2012). In a recent meta-analysis, Vijayakumar et al. report that social exclusion or rejection elicited engagement of the subgenual anterior cingulate cortex, ventrolateral PFC, posterior midline regions, and ventral striatum (Vijayakumar, Cheng, & Pfeifer, 2017) in numerous studies. There is also some preliminary evidence that activation in these regions in response to rejection is associated with pubertal development (Masten, Eisenberger, Pfeifer, & Dapretto, 2013; Silk et al., 2014) but more research linking pubertal development with social exclusion and rejection is necessary.

As adolescents expand their social network there is a change in their sensitivity to fairness, trustworthiness and cooperation. Although a basic sense of fairness is observed in young children, a complex understanding of fairness, trust, and reciprocity develops gradually over adolescence and into early adulthood (Fett, Gromann, Giampietro, Shergill, & Krabbendam, 2014; Güroǧlu, van den Bos, & Crone, 2009), which is accompanied by adolescent-unique functioning of the TPJ and dlPFC (van den Bos, van Dijk, Westenberg, Rombouts, & Crone, 2011). This underscores the notion that the same brain regions within the social brain circuit have multiple roles in social behaviors and, conversely, that the development of social behavior is served by multiple brain regions and circuits.

Social cognitive skills

The maturation of social cognitive skills, which include face processing skills, emotion recognition, and perspective-taking, draw on increased maturation of cognitive development and rely on some of the same neural circuitry in prefrontal and parietal cortices. Face processing skills–including the ability to extract identity, emotional expression, and direction of eye gaze—expand to include judgments of attractiveness or social status during adolescence (Kilford et al., 2016). Regions in what has been termed the “core face-processing network” change significantly around the time of puberty just as these skills are becoming more sophisticated (Cohen Kadosh, Cohen Kadosh, Dick, & Johnson, 2011). As adolescents spend increasingly more time with peers, they develop increased sensitivity to social affective input and are better able to “read” other people’s emotions (Garcia & Scherf, 2015), which is associated with activation of the social brain network, including the dmPFC, pSTS and TPJ (Sebastian et al., 2012).

Perspective-taking (or mentalizing), the ability to make attributions about others’ beliefs, thoughts, desires, intentions, and feelings, changes dramatically across development (Frith & Frith, 2007). Although some mentalizing is observed in infancy (Baillargeon, Scott, & He, 2010) and early childhood (Frith & Frith, 2007), adolescence is when mentalizing advances sharply and coincides with increasing activation in the dmPFC, TPJ, and pSTS (e.g., Burnett, Bird, Moll, Frith, & Blakemore, 2009; Overgaauw, van Duijvenvoorde, Moor, & Crone, 2015; Pfeifer et al., 2009). For example, using the Director task, which requires participants to take another person’s perspective (the Director) to move objects around a screen, Dumontheil, Hillebrandt, Apperly, and Blakemore (2012) found developmental differences between adults and adolescents (11–16 years); adolescents showed greater activation whenever social information was present, whereas adults showed greater activation only when the directors' viewpoints were relevant to task performance (Dumontheil et al., 2012), suggesting that social information may be over-represented in the adolescent brain, even when it is not germane to the present goal.

Collectively, these studies lend support for the notion that adolescence is a sensitive period for social development (Blakemore & Mills, 2014). They also showcase how multivariate inputs, including structural brain development, functional neural activation, and hormone levels as well as dynamic social behavior, contribute to variation in behavior and experience. A common theme throughout these studies is the vast individual variation in how youth respond to social contexts, at both the neurobiological and psychological levels. The neurobiological susceptibility to social context framework (Schriber & Guyer, 2016), which is based on neurobiological susceptibility models more broadly (Ellis, Boyce, Belsky, Bakermans-Kranenburg, & Van Ijzendoorn, 2011), encapsulates this idea. It highlights that the manner and extent to which social contexts—including relationships with parents/family and peers—shape developmental outcomes is moderated by adolescents’ susceptibility to social context, as indexed by brain characteristics (e.g., function and structure). This model underscores the importance of individual differences in brain sensitivity to perceiving, processing, and responding to social information. At least one other model proposes that puberty, via physical, neural, and interpersonal changes, “dramatically alters adolescents’ social perceptions and experiences with peers” (Pfeifer & Allen, 2021). Furthermore, Pfeifer & Allen argue that social experiences contribute to a cascading series of developmental processes that build on preceding biological and psychosocial maturation (Pfeifer & Allen, 2021), highlighting the dynamic nature of how social interactions, past and present, contribute experiences to the developmental process.

Social influence

There is a rich literature examining social influence and associated neural correlates during adolescence. Initial studies of social influence focused on the role of peers, given the significant uptick in the time spent with them. These studies generally find that the mere presence of a peer impacts how adolescents make reward-related (O’Brien, Albert, Chein, & Steinberg, 2011; Smith, Steinberg, Strang, & Chein, 2015), risky (Smith, Steinberg, et al., 2014), and cognitive (Silva, Shulman, Chein, & Steinberg, 2016) decisions. Peer influence has also been studied extensively with neuroimaging studies using, for example, preferences for music or art. In one study, both adolescents and adults adjusted music preferences if they differed from peers’ ratings of music (Berns, Capra, Moore, & Noussair, 2010). In a separate study, adolescents adjusted ratings of artworks after seeing their parents and peers’ ratings (Welborn et al., 2015). In both studies, this social influence was associated with changes in brain activation in the anterior cingulate cortex (Berns et al., 2010) and regions active during mentalizing, including the medial prefrontal cortex and TPJ (Welborn et al., 2015), suggesting that the prolonged development regions implicated in social processing may be associated with increases in susceptibility to social influence.

More recent work has examined social influence beyond the peer network to determine the role that family and parents have in making decisions. Contrary to the popular assumption that peers have greater influence over adolescents than do others, studies have shown that parents continue to exert significant sway (Guassi Moreira et al., 2018; Guassi Moreira, Tashjian, Galván, & Silvers, 2020). In a risky decision-making task performed during fMRI, adolescents exhibited greater activation in reward-related neural regions when making nonrisky choices in the presence of their mother versus an unknown adult (Guassi Moreira & Telzer, 2018), challenging the assumptions that reward circuitry is more response to risky vs. nonrisky choices and that it does not discriminate in favor of parents versus others.

The term “peer influence” or “social influence,” has colloquially had a negative connotation. Recent thinking, however, conceptualizes social influence “as an opportunity for promoting social adjustment, which can redirect negative trajectories and help adolescents thrive” (Telzer, van Hoorn, Rogers, & Do, 2018). In other words, it is now understood that “social influence” is not synonymous with “negative influence.” In fact, research in recent years has demonstrated that although greater susceptibility to social influence may in some cases lead to negative or antisocial behavior in adolescence, it can also yield positive and adaptive adjustment (Telzer et al., 2018). Drawing on the neuroscience research of social cognition in adolescence, this framing promotes the view that social influence, from both peers and family, can have an impact on positive adjustment during adolescence through neural mechanisms that exhibit heightened sensitivity in response to social others (Qu, Jorgense, & Telzer, 2020). Furthermore, it beneficially contributes to the narrative shift underway in adolescent neuroscience research.

Digital age

The rapid rise in digital technology and social media worldwide has also led to increased attention to the relationship between social media and adolescent brain development. An important task for developmental scientists is to ascertain how the digital world is integrated into the dynamic developmental processes that characterize the adolescent brain.

Prevalence rates of technology use among adolescents, currently at 95% in the United States (Rideout & Robb, 2019), are high, leading to vigorous investigation into the relation between social media and adolescent development. One primary topic of interest is whether social media is responsible for increases in anxiety and depression observed among young people (Baker & Galván, 2020). A detailed review synthesizing key findings on this topic was published recently (Odgers & Jensen, 2020). In brief, existing literature demonstrates inconsistent and primarily small associations between the quantity of digital technology use and mental health, with no support for causal claims (Odgers & Jensen, 2020). In general, there is a lack of evidence for strong connections between time spent on social media and mental health and meta-analyses on the topic are currently mixed.

From a neurobiological perspective, parents, educators, policymakers, and adolescents themselves are keen to know: What is social media “doing” to adolescent development and the developing brain? Unfortunately, the answer is not simple for a few reasons: First, social media platforms change rapidly. Today, WhatsApp, Instagram, and TikTok are popular but tomorrow these will be replaced by new apps and tools and none are used exclusively so knowing the effects of any one social media platform on brain development is challenging to discern. However, a few studies have found that receiving “likes” in a functional magnetic resonance imaging (fMRI) task that mimicked the Facebook (Sherman, Payton, Hernandez, Greenfield, & Dapretto, 2016) and Instagram (Sherman, Greenfield, Hernandez, & Dapretto, 2018) experience yielded robust activation in the ventral striatum—a region associated with receiving reward—in adolescents. Second, no study has examined brain functioning while the adolescent is engaged in social media use in real-time. This may be because the majority of adolescents engage with social media through smartphones and that is not currently possible to simulate in the scanning environment. Third, because social media use is so prevalent, there is no obvious “control” condition or comparison group of people who do not use social media. In other words, to truly isolate the effects of social media use on the brain, researchers would need to compare brain function in a group of adolescents who heavily consume social media and compare them to a group of adolescents who rarely or never use social media—the latter group is challenging to find and unlikely to be representative of contemporary young people.

Despite these limitations, there is a rich literature examining psychological experiences that are relevant to digital interaction, including social rejection/acceptance and peer influence. Being socially accepted, for example, through “likes” in a chat room fMRI task yielded greater activation in the ventral striatum in children (Achterberg et al., 2017), adolescents (Guyer, Choate, Pine, & Nelson, 2012; Moor, van Leijenhorst, Rombouts, Crone, & van der Molen, 2010), and adults (Davey, Allen, Harrison, Dwyer, & Yücel, 2010). Likewise, social rejection in the virtual world mimics negative feelings experienced in real-life following rejection and engages neural circuits implicated in greater arousal, negative affect, and pain (Guyer, McClure-Tone, Shiffrin, Pine, & Nelson, 2009; Vijayakumar et al., 2017; Will, van Lier, et al., 2016), an effect that is exaggerated in adolescents compared with adults (Crone & Konijn, 2018). These findings offer a few insights: first, engagement of a broad network of circuits is necessary to represent the broad range of psychological and visceral reactions to negative social experience. Second, greater activation of these brain networks in adolescents offers opportunity for future research to disentangle at least two developmental reasons for this observation: adolescents exhibit greater activation than adults because (1) social rejection is more painful for them or (2) the ongoing developmental of these regions necessitates greater engagement to achieve similar neural representation as adults.

Clearly, there is much more work to be done in the area of social media and adolescent brain development. Only through longitudinal approaches will we learn whether the adolescent brain is particularly sensitive to social media and whether there are particular neural networks that are advantaged or disadvantaged by intense media use. Designing experiments that truly capture the interactive developmental processes involved—included pubertal changes, brain development, and shifts in social behavior—with engagement with the digital world will be paramount. But perhaps more interestingly, the study of social media and brain interactions provides opportunity to examine how cohort-unique experiences influence developmental outcomes of generations to come.

Part IV: Opportunities and Adolescent Brain Research

With increasing collaboration between scientists and policy scholars, there is greater opportunity to incorporate the principles that have emerged from research on the adolescent brain the last decade into the policies and interventions that impact young people. In this section, there is discussion of how these principles, including (1) the importance of individual differences, (2) the dynamic and interactive nature of brain systems, and (3) the role that continued plasticity and refinement of the brain through early adulthood has on behavior, can help inform three themes that broadly impact adolescent development: defining adolescence, “one size doesn’t fit all” and inclusion of adolescents that represent the diversity that comprises this population.

Defining Adolescence

One surprisingly unresolved issue in adolescent research is “who is an adolescent?”. When does childhood end, when does adulthood begin and what exactly happens in the middle? (Ledford, 2018). Specifically, when does the developing brain cease developing and start aging? This is important because inherent in the understanding of the “developing brain” is the notion that the brain retains its plasticity so long it is in the developing phase. Despite considerable discourse to define “adolescence” the field does not have a universally-adopted definition for this phase of life. There is consensus that adolescence is a transition period between childhood and adulthood; is initiated at puberty; encompasses striking physical, emotional, neurobiological, and psychological changes; is characterized by dynamic shifts in brain networks; and varies widely intra-individually (i.e., there is no “average” adolescent). But a precise onset (or what triggers it; Dorn, ) and endpoint remain elusive. One glaring challenge is that the boundaries are a moving target. Across successive generations, the “onset” of adolescence—puberty—has started earlier and earlier in many countries, including the United States (Ledford, 2018). This trend is most notable in girls, with breast development (Aksglaede, Sørensen, Petersen, Skakkebæk, & Juul, 2009) and first menstrual period (Zhai et al., 2015) occurring approximately one year earlier than in previous generations. Whether this means there are generational differences in when brain changes begin is unknown. In the United States, epidemiological data suggests breast development has been observed to occur at increasingly earlier ages since the late 1990s in white girls as compared with African American girls, who had already exhibited breast development at a younger age (Biro et al., 2006) and that Latinx girls are more likely to have menarche earlier than Black girls, whereas white and Asian girls are more likely to have menarche later than both Latinx and Black girls (Biro et al., 2018).

The end of adolescence (and end of brain plasticity) is equally blurry. There is currently no physical metric that can precisely represent the beginning of adulthood. Instead, the end of adolescence has historically been based on social roles. Marriage, college, financial security, and parenthood have all been informally used as “adult milestones” indicating that a person has reached adulthood but clearly these behaviors vary across individuals, cultures, and generations. Even these historical markers of adulthood are shifting to later in life. In the United States, the median age of marriage was 20.3 for females and 22.8 for males in 1950. Since the late 1980s, youth have remained more financially dependent on their parents for longer and are marrying later. In 2015, the median age of marriage jumped to 27.6 for females and 29.5 for males. One reason for this trend is that there has been an increase in the number of people (particularly women) who are attending college or university. Fall enrollment in American colleges and universities increased 26% between 1997 and 2007 and 8% between 2007 (18.3 million) and 2017 (19.8 million (Snyder, de Brey, & Dillow, 2018). In 1970, 42.3% of the college-attending population was women; today, women make up 57% of the college population (Snyder et al., 2018). This increase makes individuals more dependent on parents or government loans for financial support, particularly as the costs of college continue to rise steadily. The economic crisis in 2008 and the COVID-19 pandemic in 2020 further contributed to continued financial dependence on parents, as college graduates moved back home in record numbers (Fry, 2020). Whether delays into “adulthood” have an impact on brain development remains an open question empirically but insofar as experience changes the trajectory of brain development, the brain retains its plastic state if adulthood is delayed, as shown in animal models (Johnson, Loucks, et al., ). The specific outcomes will vary by individual but will be the product of each individual’s particular circumstances that delayed the transition into adulthood (e.g., education, financial, emotional) interacting with their environment. Crucial brain systems that are the last to mature dynamically respond to these environmental inputs and to each other as each develop independently and interactively.

Why is a definition important? Conventional wisdom and scientific evidence suggest that the vast individual variation inherent in development will preclude identification of an exact definition of adolescence. However, parameters to describe a window of life that spans 15+ years will help inform policies and programs. It will also help neuroscientists design studies that aim to capture the nuance of brain development during this time window. For example, identifying particular points in adolescence when the brain is most modifiable to particular interventions or programs will help guide healthy developmental trajectories. There are other practical implications of defining adolescence. Judges and doctors lean on neuroscientists and psychologists to tell them when a young person is competent to make adult decisions. Additionally, funding for research on adolescence lags behind that of early childhood, likely due to the challenge in supporting efforts for which there is no clear definition (Ledford, 2018).

“One Size Doesn’t Fit All”

Myriad rearing environments and earlier-life experiences that contribute to individual-level developmental processes lead to vast variation in brain developmental outcomes and trajectories. A focus of the last decade has been on uncovering the role that individual differences play in current understanding of brain development. What has emerged is the understanding that the brain is influenced by multiple elements, including individual-level factors, and no single element (e.g., one neural region, system, or experience) drive its development. Neuroscience research demonstrates that patterns of brain function not only change across development, but within-individuals and within individuals under different contexts (Cohen et al., 2016; Rudolph et al., 2017), underscoring the point that any one snapshot in time of any one individual or developmental group will not wholly generalize to understanding the entire group.

This information is valuable for policies and programs that support young people in particular ways. First, it suggests policies and programs should consider intersectional approaches that target multiple developmental systems, including psychological, experiential, and neurobiological systems, in a coordinated manner. This could occur by bridging multiple programs to ensure continuous interaction of all levels of the developing system. In addition, findings from network analyses showing that the distinct cognitive systems in the brain (e.g., attention, executive, sensorimotor) collaboratively work to maximize cognitive performance should inform interventions and programs that target cognitive systems holistically, rather than in isolation. Second, research on individual differences argues against the “one size fits all” approach and suggests that practitioners should instead adopt different versions of policies and programs that vary by age, experience, or executive functioning, for example.

Intentionally Inclusive Research Practices

There is growing recognition that adolescent brain research needs to be more inclusive of all the types of diversity that represent our young people. Despite this recognition, we can do better. Currently, 90% of current evidence about adolescence comes from research in high-income countries (Blum & Boyden, 2018). Yet nine-tenths of people aged between 10 and 24 live in low- and middle-income countries (LMICs). This divide precludes generalization of the research to broader understanding of adolescents world-wide and muddles the path forward in adolescent neuroscience research.

The vast variation in cultural and environmental contexts has implications for adolescent neuroscience data. Most of this research has examined individuals in western cultures and very few studies have provided detailed information about socioeconomic status (SES), ethnicity, and other cultural factors (e.g., immigrant/refugee youth, youth of color, discrimination; Qu et al., 2020). Moreover, research shows that the specific stressors, risk factors, and environmental contexts that increase adversity in adolescence vary considerably across diverse contexts (Masten, 2014). Environmental stressors differentially impact brain processes in populations chronically exposed to stress versus populations with less exposure to adversity. For example, SES has been consistently associated with functioning across a wide variety of neurocognitive domains including language, memory, executive functioning, and socioemotional processing (Ursache & Noble, 2016) and significant mechanistic evidence suggests this occurs through different neurobiological pathways (Noble et al., 2015; Noble, Houston, Kan, & Sowell, 2012). This research points to the need to take contextual differences into account when implementing policies and services aimed at reducing maladaptive risk-taking and other negative outcomes (Uy & Galván, 2017). Generally, understanding research sample demographics and characteristics will be crucial to accurately interpreting and extrapolating adolescent brain research to youth policy. Qu and colleagues propose an interdisciplinary approach to fill the gap, combining developmental psychology, cultural psychology and neuroscience (Qu et al., 2020). Adopting this inclusive and transparent approach will help achieve greater parity in representation of the diverse experiences that contribute to the observed individual differences and dynamic developmental trajectories. Large-scale efforts, including the NIH-funded 10-year longitudinal Adolescent Brain Cognitive Development study, will advance these goals because of the multi-variate data collected across 21 research sites in the United States and the large sample size, at over 11,000 children and adolescents enrolled.

Conclusions

The study of the adolescent brain has evinced significant advances in the past decade. New directions in understanding brain network topography and how neural systems dynamically integrate to support behaviors that emerge or mature during adolescence has been a boon for the field. Together, they have helped address long-standing questions about how the brain develops vis-à-vis the vast diversity in rearing environments and paving the way for new areas of inquiry including in the social media domains. These investigations have led to new conclusions and iterative reframing of adolescence, such that deficit models are becoming less prominent in favor of ones that frame exploration, reward sensitivity, and risk-taking during adolescence as a key ingredient for the formation of psychological and neurobiological maturation. Importantly, these studies have demonstrated that the adolescent brain develops in a dynamic and integrative matter, calling on multiple brain systems and developmental processes to advance towards maturation. Dynamic systems theory served as a guiding framework in this review because it is useful for advancing beyond the notion that there is not a 1:1 correspondence between a particular brain region and a particular behavior. It is also helpful in describing why each brain region/circuit serves multiple functions and conversely, why behavioral changes are dynamically influenced by multiple brain circuits. A number of important topics supported by a rich literature were not included in this review that are highly relevant for adolescent brain development, including emotion regulation, the role of social contexts (including poverty, neighborhood, parenting), and early life adversity, but are described elsewhere in a more in-depth manner than they would have been here (Johnson, Loucks, et al., ; Tan, Oppenheimer, Ladouceur, Butterfield, & Silk, 2020; Tottenham, 2020). Collectively, these new research questions, methods, and models have shed an overdue spotlight on adolescence and have laid the groundwork for another decade of discovery.

Link to Article

The Adolescent Brain: A Decade of Progress and Future Directions

Introduction

The past decade has witnessed significant advancements in our understanding of the adolescent brain. Novel insights into brain network organization and the dynamic integration of neural systems have shed light on the emergence and maturation of behaviors characteristic of adolescence. These findings have challenged long-held deficit models, reframing adolescence as a period of exploration, reward sensitivity, and risk-taking that is crucial for psychological and neurobiological development. This review synthesizes key advancements in adolescent brain research, highlighting the dynamic and integrative nature of brain development during this transformative period.

Brain Networks and Puberty

Structural Brain Development

Studies have revealed negative associations between grey matter volume and pubertal stage, indicating a decrease in grey matter as puberty progresses. Conversely, white matter volume increases with pubertal maturation, reflecting enhanced myelination and axonal growth.

Functional Brain Development

Pubertal development has been linked to changes in brain function. For example, increased activation in reward-processing regions (e.g., ventral striatum) has been observed in response to rewards and social stimuli as puberty progresses.

Dynamics of Brain Development

Network Connectivity

Functional connectivity and graph theory analyses have demonstrated that large-scale brain networks dynamically integrate with and segregate from one another to support cognitive and behavioral development. These networks exhibit increased coordination and efficiency as adolescence progresses.

Individual Variation

Individual differences in brain development are substantial. Structural and functional topography of brain networks varies based on factors such as age, sex, and pubertal status. These variations contribute to the wide range of behaviors observed during adolescence.

Integrative Behaviors in Adolescence

Reinforcement Learning and Decision-Making

Adolescents exhibit enhanced activation in the ventral striatum during reward processing, which contributes to increased learning from feedback and exploration. However, this heightened reward sensitivity can also lead to increased risk-taking. Studies have shown that adolescents engage in more risky behaviors under affectively charged contexts and in the presence of peers.

Social Behavior

The "social brain" network, including regions such as the amygdala, temporal cortex, and temporoparietal junction, undergoes significant maturation during adolescence. This maturation supports the development of social cognitive skills, such as face processing, emotion recognition, and perspective-taking. Adolescents become increasingly sensitive to social rewards, rejection, and influence from peers and family.

Digital Age

The rapid rise of social media has led to investigations into its impact on adolescent brain development. While some studies have found associations between social media use and mental health, evidence for causal claims is lacking. Future research should explore the longitudinal effects of social media on brain development and the potential for both positive and negative outcomes.

Opportunities in Adolescent Brain Research

Defining Adolescence

Defining the boundaries of adolescence remains a challenge. The onset (puberty) and endpoint (adulthood) are fluid and vary across individuals and cultures. Establishing parameters for this developmental period would inform policies and programs targeting adolescents.

Individual Differences

The vast individual variation in brain development necessitates a shift away from "one size fits all" approaches. Policies and interventions should consider intersectional factors and tailor approaches to meet the diverse needs of adolescents.

Inclusive Research Practices

Adolescent brain research must become more inclusive to represent the diversity of experiences and contexts that shape development. Expanding research to low- and middle-income countries and underrepresented populations is crucial for generalizing findings and addressing disparities.

Conclusions

The past decade has brought significant advancements in our understanding of the adolescent brain. The dynamic and integrative nature of brain development, coupled with the influence of individual experiences, has reframed adolescence as a period of opportunity and growth. Future research should continue to explore the complexities of this transformative period, leveraging interdisciplinary approaches and inclusive practices to promote healthy adolescent development.

Link to Article

The Adolescent Brain: A Decade of Discovery

Introduction

The last ten years have brought exciting advancements in our understanding of the adolescent brain. Researchers have uncovered how different brain regions work together to support the unique behaviors and experiences of this transformative period. This new knowledge has shifted our perspective on adolescence, highlighting its importance as a time of exploration, learning, and growth.

Brain Networks and Development

One major discovery is that the adolescent brain develops through the dynamic interaction of multiple brain networks. These networks, which include regions involved in attention, decision-making, and social processing, work together to support the cognitive and behavioral changes that occur during adolescence.

Dynamics of Brain Integration

Another key finding is that the integration of these brain networks is not static but changes over time. This means that different brain regions may play a more or less important role in certain behaviors depending on the developmental stage and the specific situation. For example, the prefrontal cortex, which is involved in decision-making, becomes increasingly connected to other brain regions as adolescents mature.

Reinforcement Learning and Decision-Making

Adolescents are known for their increased sensitivity to rewards and their tendency to engage in risky behaviors. Research has shown that this is partly due to the unique way their brains process rewards. The striatum, a brain region involved in motivation and reward, is particularly active in adolescents when they receive rewards. This heightened sensitivity can lead to increased exploration and risk-taking, which are essential for learning and adapting to new environments.

Social Behavior and the "Social Brain"

Adolescence is a time of significant social development. The brain regions that support social interactions, known as the "social brain," undergo major changes during this period. These changes enable adolescents to navigate increasingly complex social situations, form meaningful relationships, and develop a sense of identity.

Social Influence and the Digital Age

Social interactions play a crucial role in shaping adolescent brain development. Studies have shown that the presence of peers can influence how adolescents make decisions, even when it comes to risky behaviors. With the rise of social media, researchers are now exploring how digital interactions affect the developing brain. While there is still much to learn, it's clear that social media platforms can both positively and negatively impact adolescent development.

Opportunities for Future Research and Policy

The insights gained from adolescent brain research have important implications for policies and programs that support young people.

Defining Adolescence:

Defining the boundaries of adolescence remains a challenge, but understanding brain development can help inform this process.

Individual Differences:

Brain development varies widely among individuals, highlighting the need for tailored interventions and programs that consider each adolescent's unique experiences and needs.

Inclusive Research:

To ensure that our understanding of adolescence is representative of all young people, research must include diverse populations and consider the impact of cultural and environmental factors.

Conclusion

The past decade has brought significant advancements in our understanding of the adolescent brain. The dynamic and integrative nature of brain development during this period has reframed our view of adolescence as a time of opportunity and growth. By continuing to explore the complexities of the adolescent brain, we can better support young people as they navigate this transformative stage of life.

Link to Article

The Amazing Adolescent Brain: A Time of Change and Growth

Part I: Brain Networks and Development

Your brain is like a supercomputer with different parts working together to help you think, feel, and do things. These parts, called brain networks, change and grow a lot during your teenage years.

One big change is that the connections between different brain networks get stronger. This helps you think better, make decisions, and control your emotions. It's like building a highway between different cities, making it easier for them to communicate.

Part II: Dynamic and Integrative Behaviors

When a teenager's brain networks develop, this causes some changes in how they act and feel.

Reward Sensitivity:

• Teens might feel more excited about things that make them happy, like hanging out with friends or playing video games. This is because the part of your brain that makes you feel good is extra active during your teenage years.

Learning and Decision-Making:

• Teenagers also get better at learning new things and making decisions. This is because their brains are using feedback from past experiences to help them figure out what to do in the future. It's like the brain is learning from its mistakes and getting smarter.

Social Behavior:

• Teens tend to spend more time with friends and thinking about relationships. This is because the parts of the brain that help you understand and connect with others grow and change during these years. Teenagers might also be more sensitive to what other people think and feel.

Part III: Opportunities for Growth

Your teenage years are a special time for your brain. Here are some things to keep in mind:

Defining Adolescence:

• It's hard to say exactly when people become adults, but the teenage years are a time of big changes in your brain and body.

Individual Differences:

• Everyone's brain develops differently. Your experiences, environment, and even your genes can affect how your brain grows.

Inclusive Research:

• It's important to study the brains of teenagers from all different backgrounds to understand how everyone's brain develops.

Conclusion

Your adolescent brain is a work in progress. It's changing and growing in amazing ways, helping you become the awesome adult you're meant to be. Embrace the changes and enjoy the ride!

Link to Article

The Amazing Changing Brain of Teenagers

Just like a child's body is growing and changing as they get older, their brain is too. But a teen's brain changes in some really cool and unique ways that make being a teenager a special time.

The Brain is Like a Team

Your brain is made up of different parts that work together like a team. These parts talk to each other and help you do all the things you do, like think, feel, move, and learn. As a child gets older, these parts start to work together better and faster.

The Brain Loves to Explore

Teenagers are known for being curious and trying new things. That's because the brain is really good at learning and adapting to new experiences. When you try something new, your brain makes new connections and gets stronger.

The Brain is Sensitive to Rewards

Have you ever noticed that you feel really good when you get a compliment or win a game? That's because your brain releases a special chemical called dopamine that makes you feel happy. Teenagers have more dopamine in their brains than adults, so they feel these good feelings even more strongly.

A Teen's Brain is Still Growing

A teenager's brain can do lots of things well, but it's still not done growing yet. In fact, some parts of the brain don't finish developing until you're in your 20s.

Everyone's Brain is Different

Just like everyone has a different body, everyone has a different brain. Your brain is shaped by your experiences, your environment, and even your genes. That's why it's important to remember that everyone learns and grows at their own pace.

Link to Article