Adolescence is a developmental period characterized by substantial neurobiological changes (Casey et al., 2008; Giedd et al., 1999; Somerville et al., 2010; Spear, 2000) and is associated with increases in engagement in risky behaviors (Dahl, 2004; Steinberg, 2008), including nonsuicidal self-injury (NSSI). NSSI is defined as direct and deliberate self-harm without intent to die (Nock, 2009). NSSI typically begins in early adolescence around ages 12–14 years old (Ammerman et al., 2018; Whitlock et al., 2011; Zetterqvist et al., 2013), and rates of NSSI are higher in girls as compared to boys (Swannell et al., 2014). Prevalence of NSSI peaks during adolescence, with high rates in both community (18%; Muehlenkamp et al., 2012) and clinical (50%; Asarnow et al., 2011; Nock, 2010) samples. NSSI is associated with significant distress and injury (Glenn & Klonsky, 2013; Klonsky et al., 2014) and is a prospective predictor of suicide attempts (e.g., Asarnow et al., 2011; Guan et al., 2012; Klonsky et al., 2013). Research has examined psychological correlates of NSSI, but few studies have investigated the neural underpinnings of NSSI, which may uncover novel mechanisms contributing to NSSI.

Prevailing theoretical models of NSSI focus on the reinforcing properties of NSSI. For example, the Four Function Model (FFM) specifically posits NSSI as a behavior that regulates internal/intrapersonal (automatic function) or interpersonal (social function) experiences (Nock & Prinstein, 2004). Within this model, automatic and social functions are reinforced by either generating desired cognitive-affective or social states (positive reinforcement) or by decreasing aversive cognitive-affective or social states (negative reinforcement). Research supports this model (Nock, 2009; Nock & Mendes, 2008; Nock & Prinstein, 2005), and implicates automatic negative reinforcement as the most prevalent function of NSSI (Bentley et al., 2014; Liu, 2017). Similarly, other theories highlighting the reinforcing properties of NSSI suggest that this behavior provides an affect-regulation function (Hooley & Franklin, 2018; Klonsky, 2007), or avoidance or escape from unwanted emotional experiences (Chapman et al., 2006). In sum, these theories suggest that NSSI may aid in reduction of negative affect and often serves a regulatory function among adolescents.

Neural structure in the prefrontal cortex and parietal cortex may be associated with NSSI given their substantial role in emotion regulation (Etkin et al., 2015; Goldin et al., 2008; Ochsner et al., 2004), self-regulation (Heatherton & Wagner, 2010; Langner et al., 2018; Palacios-Barrios & Hanson, 2019), and cognitive control (Luna et al., 2015; Vijayakumar et al., 2014). Initial research reveals structural neurobiological alterations in adolescents with a history of NSSI. A recent study found that self-injuring adolescent girls (ages 13-19) show regional reductions in gray matter volume in the insular and right inferior frontal cortex when compared to age-matched peers (N=40; Beauchaine et al., 2019). Another study comparing adolescent girls with NSSI history (ages 14-18, N=50) to an age- and gender-matched healthy control group found that NSSI history was associated with lower regional grey matter volume in the insular cortex and anterior cingulate cortex (ACC; Ando et al., 2018). Of note, only Beauchaine et al. (2019) reported covarying for age in their statistical models, and neither study explored whether NSSI altered age-related changes in the brain. Examining the role of age is paramount to further understanding of how the developmental pattern of cortical structure may differ based on NSSI history.

Prior research has quantified brain maturation during childhood and adolescence by measuring cortical thickness and surface area either across time or for participants of different ages (Mills & Tamnes, 2014). Cortical surface area and thickness are distinct structural components with unique cellular and genetic origins (Geschwind & Rakic, 2013; Kremen et al., 2013; Panizzon et al., 2009), that represent independent measures of cortical structure. Longitudinal magnetic resonance imaging (MRI) studies show that cortical thickness decreases approximately linearly across childhood and adolescence for most of the cortex (Giedd et al., 1999; LeWinn et al., 2017; Mills et al., 2014a; Tamnes et al., 2017; Wierenga et al., 2014). In contrast, cortical surface area increases during childhood but then decreases across adolescence (LeWinn et al., 2017; Raznahan et al., 2011; Tamnes et al., 2017; Wierenga et al., 2014), following a non-linear trajectory. Thus, typical neurodevelopment during adolescence is characterized by cortical thinning and decreases in cortical surface area. Additional evidence suggests that expansion of cortical thickness is complete by age 2, while the expansion of cortical surface area continues into childhood and early adolescence (Gilmore et al., 2018; Lyall et al., 2015), demonstrating that these measures of cortical structure are independent of one another, follow different trajectories, and develop at different rates.

Refinement of neural structure during adolescence, characterized by decreases in cortical surface area and thickness, may reflect improvement in behaviors including emotion regulation and cognitive control (Vijayakumar et al., 2014). However, excessive reductions in cortical structure during this sensitive developmental period may be detrimental. Although research has not directly examined whether NSSI is associated with age related changes in neural structure, related psychological conditions alter the association between age and cortical development. For example, in longitudinal analyses, depressive symptoms are associated with accelerated cortical thinning of the frontal lobe in adolescents (Bos et al., 2018; Luby et al., 2016). Given existing work suggesting that neurodevelopmental trajectories vary with behavior and symptoms, it is possible that similar associations may emerge with NSSI.

Developmental cognitive neuroscience research demonstrates that widespread decreases in cortical thickness and cortical surface area occur during adolescence. These changes may be particularly relevant to understanding adolescents' engagement in health risk behaviors, including NSSI. Few studies have explored the association between NSSI and brain structure in adolescence, and no studies have investigated whether there are differences in the relation between age and brain structure in NSSI populations.

In this preliminary study, we first examined associations between age and cortical thickness and surface area in a sample of adolescent girls. Consistent with prior research, we hypothesized that increased age would be associated with reductions in cortical thickness and surface area. Next, we investigated group differences in cortical thickness and surface area in adolescents with versus without NSSI history. We hypothesized that lifetime history of NSSI would be associated with less cortical thickness and surface area in the prefrontal cortex and parietal cortex. Finally, we examined the association between age and neural structure in adolescents with NSSI history versus without NSSI history. We hypothesized that NSSI history would moderate the association between age and cortical surface area and cortical thickness. Because cortical thickness and surface area tends to decline in the age range of our sample (LeWinn et al., 2017; Mills et al., 2014a; Tamnes et al., 2017; Wierenga et al., 2014), we expected that the relation between age and decreased cortical structure would be stronger among adolescents with a history of NSSI compared to those without a history of NSSI. We also expected these differences would remain significant while controlling for well-known correlates of NSSI, including depressive symptoms (Schreiner et al., 2017) and medication use (Quevedo et al., 2016).