If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Non-suicidal self-injury (NSSI), a transdiagnostic behavior, often emerges during adolescence. This study used the Research Domains Criteria (RDoC) approach to examine cognitive control (CC) with a focus on response inhibition and urgency relative to NSSI severity in adolescents.
Methods
138 adolescents assigned female sex at birth, with a continuum of NSSI severity completed negative and positive urgency measurements (self-report), emotional Go/No-Go task within negative and positive contexts (behavioral), structural and functional imaging during resting-state and task (brain metrics). Cortical thickness, subcortical volume, resting-state functional connectivity (RSFC), and task activation focused on an a priori defined cognitive control network (CCN). 84 participants had all these main measures. Correlations and stepwise model selection followed by multiple regression were used to examine the association between NSSI severity and multi-unit CC measurements.
Results
Higher NSSI severity correlated with higher negative urgency and lower accuracy during positive no-inhibition (Go). Brain NSSI severity correlates varied across modalities and valence: for right mPFC and right caudate, higher NSSI severity correlated with greater negative, but lower positive inhibition (No-Go) activation. The opposite pattern was observed for the right DLPFC. Higher NSSI severity correlated with lower left dACC negative inhibition activation and thicker left dACC, yet it was correlated with higher right rACC positive inhibition activation and thinner right rACC, as well as lower CCN RSFC.
Conclusion
Findings revealed multi-faceted signatures of NSSI severity across CC units of analysis, confirming the relevance of this domain in adolescent NSSI and illustrating how multimodal approaches can shed light on psychopathology.
Introduction
Non-suicidal self-injury (NSSI) is the intentional destruction of one’s own body tissue without suicidal intent in a manner that is not culturally sanctioned (
). NSSI, which is categorized as impulsive, as opposed to stereotypic or psychosis-related, is characteristically associated with tension release or emotion regulation (
Prevalence and function of non-suicidal self-injury (NSSI) in a community sample of adolescents, using suggested DSM-5 criteria for a potential NSSI disorder.
). To guide treatment development, advances are needed in understanding the complex mechanisms behind NSSI.
Recent work highlights the potential promise of a multiple-units of analysis approach to study functioning domains (psychological/biological systems outlined in the Research Domains Criteria (RDoC) initiative (
Klimes-Dougan B, Eberly LE, Schreiner MW (2014): Multilevel assessment of the neurobiological threat system in depressed adolescents: Interplay between the limbic system and hypothalamic–pituitary–adrenal axis. Development. Retrieved from https://www.cambridge.org/core/journals/development-and-psychopathology/article/multilevel-assessment-of-the-neurobiological-threat-system-in-depressed-adolescents-interplay-between-the-limbic-system-and-hypothalamicpituitaryadrenal-axis/BF2442694658D9483F8657B8978080AB
Coordination between frontolimbic resting state connectivity and hypothalamic-pituitary-adrenal axis functioning in adolescents with and without depression.
). One fundamental RDoC construct implicated in NSSI is Cognitive Control (CC), defined as follows: “A system that modulates the operation of other cognitive and emotional systems, in the service of goal-directed behavior, when prepotent modes of responding are not adequate to meet the demands of the current context. Additionally, control processes are engaged in the case of novel contexts, where appropriate responses need to be selected from among competing alternatives.” (
Development and Definitions of the RDoC Domains and constructs (n.d.): National Institute of Mental Health (NIMH). Retrieved February 21, 2023, from https://www.nimh.nih.gov/research/research-funded-by-nimh/rdoc/development-and-definitions-of-the-rdoc-domains-and-constructs
) While the broader construct of cognitive control encompasses many different facets, impulsivity is especially relevant to NSSI. Decades ago, “Deliberate Self-Harm Syndrome” was considered an impulse control disorder based on the conceptualization that self-injurers have difficulty resisting the impulse/urge to injure themselves (
). However, prior multi-unit studies of CC have found weak relationships between self-report and behavioral measures of impulsivity, perhaps because they tap distinct aspects of the construct (
Claes L, Muehlenkamp JJ (2013): Non-Suicidal Self-Injury in Eating Disorders: Advancements in Etiology and Treatment. Springer Science & Business Media. Retrieved from https://play.google.com/store/books/details?id=Jty4BAAAQBAJ
Testing the relations between impulsivity-related traits, suicidality, and nonsuicidal self-injury: a test of the incremental validity of the UPPS model.
). The RDoC sub-construct of CC most relevant to impulsivity is ‘Response Selection; Inhibition/Suppression’. Response inhibition is related to negative urgency more so than other impulsivity traits (
Validation of an emotional stop-signal task to probe individual differences in emotional response inhibition: Relationships with positive and negative urgency.
). Accordingly, here we focused on the relationship between NSSI severity and response inhibition (RI) and urgency (both in negative and positive contexts) within the CC framework. Since urgency is highly related to both NSSI and RI, but it does not itself fall directly within the RI construct, we broadly refer to CC as the domain encompassing both of these constructs of interest.
Prior studies using case-control designs to examine impulsivity within emotional context in NSSI have revealed mixed findings regarding CC performance (
). In accordance with RDoC, dimensional versus categorical approaches may be more useful. Furthermore, incorporation of additional units of analysis such as neuroimaging may shed light on the complexity of these mechanisms in relation to NSSI. A small but growing body of neuroimaging studies have indicated NSSI’s association with possible neural deficits underlying both emotion processing and impulse regulation (see 38–40). One prior study showed that compared to healthy controls, individuals with NSSI showed higher cingulate cortex activation and lower dorsolateral prefrontal cortex (DLPFC) activation during a CC task (
). However, no prior studies have comprehensively examined CC using a multiple-units approach in adolescents with NSSI.
Current study utilized an RDoC approach to examine CC using multiple units of analysis in adolescents with a continuum of NSSI severity. These units included self-report, behavior, and structural and functional neuroimaging characterizing the cognitive control network (CCN) (Table 1). We preliminarily examined how these CC variables relate to each other, and then tested how different units of CC measures relate to NSSI severity. We hypothesized that NSSI severity would correlate with greater dysfunction in CC across all units of analysis, and that considering all units together would shed new light on the nuances of this complex behavior.
Bilateral dACC, rACC, mPFC, DLPFC, Caudate, and Putamen activity during No-Go conditions within negative & positive contexts (Negative > Scrambled & Positive >Scrambled contrasts) during the emotional Go/No-Go Task
Connectivity
Mean Cognitive Control within Network Resting State Functional Connectivity
Behavior
➢ Accuracy in Go (no-inhibition) & No-Go (inhibition) conditions within positive & negative contexts ➢ RT in correct Go conditions ➢ d prime within positive & negative contexts as z(hit) - z(FA)
This work is part of a longitudinal study, the Brain Imaging Development of Girls’ Emotion and Self (BRIDGES), whose overarching goals are to examine Cognitive Control (CC), Sustained Threat (
). The current work is based on data collected primarily at the first assessment. The study was approved by the Institutional Review Board at the University of Minnesota (UMN).
Participants
Inclusion criteria were assigned female sex at birth, 12-16 years old with and without a history of NSSI, postmenarchal. Exclusion criteria were MRI contraindications, having a clinical condition that would potentially confound brain findings such as neurological disorders, major medical illness, and current substance abuse disorder, and (although NSSI can occur in the context of these disorders) having primary psychotic disorders, bipolar disorders, and autism that might confound findings due to associated significant neurodevelopmental abnormalities. Participants provided informed assent; parents/guardians gave consent; they were compensated for each visit.
Clinical Assessments
Diagnostic interviews were conducted separately with adolescents and parents using the Kiddie Schedule of Affective Disorders and Schizophrenia-Present and Lifetime Version (K-SADS-PL; 42,43), and the Self-Injurious Thoughts and Behaviors Interview (SITBI) (
) was administered to the adolescents. NSSI severity, our primary outcome variable, was defined as the lifetime number of NSSI episodes based on the SITBI. To address skewness, this variable was log transformed to allow for the data to be used appropriately in regression models.
Parents reported their gross income and whether adolescents were currently taking psychotropic medications. Intelligence quotient (IQ) was estimated based on the Vocabulary and Matrix Reasoning subtests of the Wechsler Abbreviated Scale of Intelligence (WASI-II) test (
), a 59-item inventory measuring dimensions of impulsivity. According to the rationale provided in the introduction, we focused on negative and positive urgency: the tendency to engage in impulsive behavior in negative or positive affect contexts.
MRI data acquisition
Neuroimaging was conducted at the Center for Magnetic Resonance Research at UMN using a Siemens 3 Tesla Prisma scanner (Erlangen, Germany) and a 32-channel receive-only head coil, using the Human Connectome Project (HCP) multiband sequences. See Supplement B for detailed information on the acquisition.
Emotional Go/No-Go Task
We measured RI in emotion contexts using a modified Go/No-Go task, which measures the ability to inhibit a dominant response in the context of affective pictures as task-irrelevant distractors (
). Letters were presented sequentially in a small box at the center of the screen superimposed on negative, positive, neutral, or scrambled images in the background, selected from the International Affective Picture System (IAPS (
Lang PJ, Bradley MM, Cuthbert BN (2005): International Affective Picture System (IAPS): Affective Ratings of Pictures and Instruction Manual. NIMH, Center for the Study of Emotion & Attention. Retrieved from https://play.google.com/store/books/details?id=VEW2PgAACAAJ
)) – that had an equal number of neutral, positive, and negative valence ratings. Participants were instructed to ignore the images and respond as quickly as possible with a button press to the presentation of each letter (Go), except the letter X (No-Go). The task was presented using E-Prime (Psychological Software Tools Inc., Sharpsburg, PA) in the MRI scanner. See Supplement D for more details on the task.
Behavioral performance was measured by accuracy on Go (no-inhibition) and No-Go (inhibition) trials across emotional backgrounds, and reaction time (RT) on accurate Go trials for each background. Additionally, an overall measure of behavioral performance was indexed by d-prime for negative and positive backgrounds, the standardized difference between the hit rate (accuracy on Go trials) and false alarm rate (commission errors on No-Go trials) distributions. Larger values of d-prime indicate better performance: higher hit rate and low false alarm rate (Supplement D).
Hereafter, we use “negative inhibition” to refer to cognitive inhibition in the context of negative backgrounds, and “positive inhibition” to in the context of positive backgrounds; and “inhibition” refers to No-Go trials, and “no-inhibition” to Go trials.
Neuroimaging data preprocessing
HCP pipelines were used to process neuroimaging data (
). See Supplement B for a detailed description of the processing steps, quality checks, and handling of head motion.
Defining the Cognitive Control Network (CCN) for Neuroimaging Metrics
We used an a priori defined CCN applied to all neuroimaging modalities to consistently examine the same network across different types of brain assessments. We selected cortical regions known to be crucial both for cognitive and emotion regulation, which are functionally inseparable (
): cognitive (dorsal) and emotional (rostral) subregions of anterior cingulate cortex (ACC), being a key region for response inhibition, especially when there is emotion interaction (
Defining biotypes for depression and anxiety based on large-scale circuit dysfunction: a theoretical review of the evidence and future directions for clinical translation.
). Since this study primarily focused on the response inhibition sub-construct of the CC domain, we limited our CCN selection around the middle and lateral frontal cortices and key basal ganglia regions, but for example not the dorsal parietal cortex which is sometimes included in cognitive control networks (
Defining biotypes for depression and anxiety based on large-scale circuit dysfunction: a theoretical review of the evidence and future directions for clinical translation.
Figure 1Locations of Glasser ROIs selected for the CCN in the current study: All ROIs were selected from both left and right hemispheres. For example, rACC includes 2 Glasser parcels (p24 & a24) as both left and right, thus rACC includes 4 Glasser parcels. Likewise, dACC includes 6 (left and right 33pr, a24pr, p24pr), mPFC includes 12 (left & right p32pr, a32pr, d32, p32, 8BM, 9m), and DLPFC includes 26 parcels (left & right 8C, 8Av, i6-8, s6-8, SFL, 8BL, 9p, 9a, 8Ad, p9-46v, a9-46v, 46, 9-46d). See Supplement C for the explanations of abbreviations.
For the cortical regions, cortical thickness values from Glasser regions in the CCN were extracted from the HCP-derived vertex-wise thickness maps, by calculating the weighted average of thickness according to ROIs’ surface area:
For the subcortical regions, volumes were calculated using FreeSurfer’s “asegstats2table” function, and then adjusted to participants’ intracranial brain volumes.
) was used to conduct a whole-brain regression analysis, measuring neural activation during the Go/No-Go task. Our contrasts of interest were Negative > Scrambled and Positive > Scrambled. The Scrambled condition was used as a comparison as it does not elicit an emotional response, allowing us to capture brain activations during cognitive inhibition specifically within an emotional context. Average z-scores within left and right CCN ROIs for these contrasts were extracted for further analyses.
Resting State Functional Connectivity (RSFC)
The CIFTI-space gray-ordinate-wise time series were used to create average time series for each of the Glasser (cortical) and Harvard-Oxford (subcortical) parcellations. CCN ROI time series were extracted, cross-correlated, and Fisher's z-transformed to yield z-scores representing RSFC in each connection. These were averaged across all CCN ROIs (Figure 1) to yield a single measure of CCN RSFC per person.
Statistical Analyses
As a preliminary step, we conducted correlation analyses with listwise exclusion to examine relationships among all study variables. To address our main aim (understanding how multilevel CC measures predict lifetime NSSI severity), we first applied a stepwise variable selection procedure that iterates through a large number of intermediate models to identify the model that optimizes Akaike information criterion (AIC). This optimized model consisted of the combination of variables that together best explained the variance of NSSI severity while penalizing overly complex models. We then tested how the variables in this “best explanatory” model predicted NSSI severity using multiple linear regression. To ensure our results were robust to outliers, we used the same variables in robust regression models (
) and found very similar results (see Table 3 for the variables whose significance changed with robust regression).
Table 3Multiple regression analysis results with multi-level Cognitive Control variables that made it to the “best explanatory model” as predictors and NSSI Lifetime Episodes (log transformed) as the outcome variable. The variables whose significance level changed with the robust regression are indicated with parentheses.
Follow-up analyses tested whether any significant effects from correlation and regression analyses above could be explained by age, income, IQ, depressive symptoms, and medication status. We also ran a regression model including only these covariates and compared that to our optimal model using an ANOVA test to make sure this model explains NSSI severity significantly better than the covariate-only model.
Different rates of missingness across variables posed a challenge for application of imputation methods. Therefore, we used all possible data available for each analysis; but because of the various missing data patterns across variables, sample sizes in each regression model (Table 3) changed depending on the combination of variables.
All statistical analyses were conducted in R (R Core Team, 2015). Figures were produced using the packages ggplot2(
Wickham H (2009): ggplot2: Elegant Graphics for Data Analysis. Springer Science & Business Media. Retrieved from https://play.google.com/store/books/details?id=bes-AAAAQBAJ
Demographic and clinical information for this sample is provided in Table 2.Figure 2 summarizes the activities completed by all participants in this study, capturing missing data and dropout, which is further detailed in Supplement A.
Table 2Demographic and clinical characteristics of BRIDGES participants. For categorical variables data are reported with sample size, and percentage in parentheses, whereas for continuous variables data are reported as mean, and standard deviation in parentheses.
Figure 2Brain Imaging Development of Girls’ Emotion and Self (BRIDGES) consort diagram for Cognitive Control. *130 MRI scans were conducted at the Year 1 time point. 8 individuals who did not have their scans conducted in the first year completed an MRI session at their second-year visit, bringing the total scans to 138.
We observed a consistent pattern of many strong correlations within each level of analysis. Accuracy and RT for Go trials during positive inhibition and age correlated positively with activation during negative inhibition in mostly left medial prefrontal regions; whereas positive urgency scores correlated negatively with: positive and negative inhibition performance (No-Go accuracy and d prime); right mPFC, right dACC, and right and left putamen activations during negative inhibition and left dACC during positive inhibition. CCN RSFC correlated negatively with negative inhibition performance (d-prime), negative no-inhibition performance (Go accuracy), age, medication status, and more importantly NSSI severity (Figure 3A). ACC stood out among other CCN regions: left dACC activation during positive inhibition correlated negatively with positive no-inhibition performance and both positive and negative urgency scores; left rACC activation during negative inhibition correlated positively with all positive performance measures. However, most of these findings did not remain significant when adjusted for age, income, BDI, WASI, and medication status; except correlations with left dACC activation during positive inhibition. Moreover, negative no-inhibition performance became significantly negatively correlated with activation during negative and positive inhibition in almost all CCN regions; and left DLPFC CT became positively correlated with activation during position inhibition in medial frontal cortex (Figure 3B).
Figure 3Correlations: A. Correlations between all variables. B. Correlations between all CC variables after controlling for the covariates.
Relationships between multilevel CC variables and NSSI Severity (N = 84)
Stepwise model testing revealed that the variables of the model that best explained the NSSI severity (AIC = -72.56) were: self-report: negative urgency; behavior: positive and negative Go accuracy; activation during negative inhibition: right DLPFC, right mPFC, right Caudate, right Putamen, left DLPFC, left dACC, left Putamen; activation during positive inhibition: right DLPFC, right mPFC, right rACC, right dACC, right Caudate, right Putamen; structural: right DLPFC, right mPFC, right rACC, and left dACC CT; connectivity: CCN RSFC.
When applied using only these variables, the linear regression model significantly predicted NSSI severity (adjusted R2 = 0.47, F(
) = 4.53, p < 0.001): Higher negative urgency; right mPFC and right caudate activations during negative inhibition and right DLPFC and right rACC activations during positive inhibition; and left dACC CT predicted higher NSSI severity. Furthermore, lower positive Go condition accuracy; right DLPFC and right putamen activations during negative inhibition, and right mPFC and right caudate activations during positive inhibition; and right rACC CT; and CCN RSFC predicted higher NSSI severity. Figure 4 shows the associations between NSSI severity and these key variables via regression plots.
Figure 4Significant associations between NSSI severity and the best explanatory variables. Capital letters refer to different units of analysis: A. Self-report; B. Behavior; C. Connectivity; D. Structure; E. Function. For all graphs, X axes represent log-scaled NSSI lifetime episodes and Y axes represent CC measures that form the best explanatory model. To highlight divergent patterns, graphs are grouped according to the brain region’s association with NSSI severity in the Left and Right hemisphere or in the Negative and Positive context.The asterisk denotes CC measures that lost their significant association with NSSI severity when the best explanatory model is controlled for the covariates.
) = 4.77, p < 0.001), the model was still significantly predicting NSSI severity: The significant associations with greater NSSI severity generally held, with some new relationships between NSSI severity and CC measures emerging as marginally significant or losing significance. Table 3 shows the results for the model predicting NSSI severity with the best explanatory variables both with and without covariates.
Furthermore, we found that our best-explanatory model, which accounted for 47% of the variance, predicted NSSI severity significantly better than a model consisting of only covariates (p < 0.05), which accounted for only 40% of the variance. This best explanatory model explained NSSI severity even better when controlling for the covariates, by accounting for 57% of the variance (see Table 3).
Discussion
This study used an RDoC approach to examine CC in adolescents with a continuum of NSSI severity. Key strengths of this work include an integrative multiple units of analysis, utilization of a transdiagnostic recruitment strategy, a relatively large (compared to prior CC/NSSI studies) sample of adolescents exhibiting a range of NSSI severity, and the consideration of impulsive behavior in both negative and positive emotion contexts. One key observation was that the CCN in adolescents with NSSI showed divergent activation patterns depending on the valence of the inhibitory context. Secondly, we observed a lateralization effect in both structural and functional brain measures, where significant associations between NSSI severity and CCN activations during positive inhibition emerged in the right hemisphere, the side of the brain most commonly implicated in emotion (
). Moreover, we demonstrated that a specific combination of CC measures from different units of analysis together explained NSSI severity significantly better than a combination of some key demographic and clinical measures. This combination of CC measures explained NSSI severity even better when it was controlled for those demographic and clinical measures. Overall, the findings revealed multi-faceted neural and behavioral signatures of NSSI severity across units of analysis, confirming the relevance of this domain in adolescent NSSI and illustrating how multimodal approaches can shed light on the complexity of how RDoC domains operate in the context of psychopathology.
Validation of an emotional stop-signal task to probe individual differences in emotional response inhibition: Relationships with positive and negative urgency.
), which found different association patterns between positive and negative conditions, negative urgency here was negatively correlated with performance during both positive and negative inhibitory conditions (No-Go) but not with performance in Go conditions. The more the adolescents viewed themselves as more impulsive in the negative context, the worse they performed in positive inhibition. Further, after controlling for covariates, negative Go accuracy correlated inversely with activation during negative and positive inhibition in almost all CCN regions. Moreover, when considered with variables from other units of analysis, higher NSSI severity was significantly associated with worse positive Go accuracy. Notably, for the same task in healthy adolescents, negative context impaired inhibitory performance (
) in adolescents with NSSI, may set the stage for impaired accuracy in the context of positive versus negative emotion due to less practice with positive emotion states. This could suggest an adaptive process at the neural level that does not rise to the level of awareness; in their daily lives, adolescents with severe NSSI still perceive themselves as more impulsive in the context of negative emotion.
With respect to brain activation correlates of NSSI severity, we also observed patterns that diverged according to context valence (Figure 4E). For example, higher NSSI severity was associated with greater activation of the right mPFC during negative inhibition, but lower activation during positive inhibition. In contrast, the opposite pattern was seen for the right DLPFC: higher NSSI severity associated with greater activation during positive inhibition, lower activation during negative inhibition. A similarly complex pattern was observed for the subcortical CCN regions: while the right caudate results mirrored those of right mPFC, the nearby region of right putamen was inversely associated with NSSI severity during negative inhibition. These findings highlight the complex interactions of emotion, cognition, and psychopathology in the brain: when tasked with inhibiting impulses, valence impacts how adolescents recruit their neural resources, and this varies by severity of NSSI.
We found a negative relationship between NSSI severity and CCN RSFC. Although it focuses on a different network, this result falls in line with prior work showing lower amygdala-frontal RSFC (
) found that adolescents with NSSI had greater connectivity between DMN and the central executive network (similar to CCN in this study). Hence, these findings are likely to vary depending on the circuit being probed.
With respect to structure, the region with the strongest associations with NSSI severity was the ACC, with a lateralization effect within different subregions: higher NSSI severity was associated with thicker left dACC, but thinner right rACC. Furthermore, after controlling for covariates, higher NSSI severity became marginally associated with thicker right DLPFC. While there are still relatively few structural MRI studies considering correlates of NSSI, these findings contrast with the direction of findings from prior studies reporting that suicidal risk is associated with thinner cortices in various frontal areas (
). Interestingly, a gene transcription/neuroimaging study showed recently that NSSI-associated cortical thickness differences in youth with NSSI were linked to cellular component morphogenesis of astrocytes and excitatory neurons (
Cai S, Guo Z, Wang X, Huang K, Yuan K, Huang L (2022): Cortical thickness differences are associated with cellular component morphogenesis of astrocytes and excitatory neurons in nonsuicidal self-injuring youth. Cereb Cortex. https://doi.org/10.1093/cercor/bhac103
). Taken together, the current findings further highlight the value of multi-modal approaches to help understand how multiple systems may be operating in concert with each other in the context of complex behaviors like NSSI.
Limitations
First, despite the relatively substantial size of this clinically enriched sample, including multiple types of data per person introduced the risk of missing data across levels, reducing power for analyses requiring all data types. Second, although a standard regression approach was appropriate for this data, considering the limits of other penalized regression approaches (e.g. Ridge or bootstrapping), our data’s correlation structure, and the lack of a validation sample, this approach should not be considered as true predictive modeling. Third, generalizability is limited by the exclusion of adolescents who were assigned male sex at birth, and relatively low rates of racial and ethnic minorities in our sample. Fourth, the key outcome measure of NSSI severity (self-reported lifetime episodes) may have flaws such as recall bias and the fact that some adolescents tend to under-report while others over-report. Fifth, to optimize power by limiting the number of CC variables, we used an average within-network connectivity metric, which precludes interpretations regarding specific connections within the CCN. Furthermore, while the current research focused primarily on how CC (across multiple units of analysis) relates to NSSI and limited the analyses of connectivity to examining within CCN (without testing other networks), other work has found that connectivity between CCN and the default mode network was greater in adolescents with versus without NSSI (
), underscoring the potential value of looking beyond the CCN in future work. Sixth, although they are distinct phenomena, since they commonly co-occur, disentangling NSSI from depression in adolescents may be impossible to do completely (
Başgöze Z, Wiglesworth A, Carosella KA, Klimes-Dougan B, Cullen KR (2021): Depression, Non-Suicidal Self-Injury, and Suicidality in Adolescents: Common and Distinct Precursors, Correlates, and Outcomes. J Psychiatr Brain Sci 6. https://doi.org/10.20900/jpbs.20210018
). Seventh, CC has many facets which were not all considered here. For example, working memory performance, task-switching capabilities, and flexibility in goal-directed thoughts and behaviors would also be excellent CC candidates for future studies that would shed more light on NSSI. Finally, cross-sectional analyses limit interpretations with respect to development, and teasing out questions related to state versus trait. For example, it may be that some of the variables here are more reflective of NSSI only when measured in close temporal sequence with NSSI events (state) while others may be predictive of current (or past) NSSI regardless of measurement time.
Conclusion
These findings begin to paint a picture of the complex ways in which CCN structure and function maps onto NSSI severity, and how results from different units of analysis can help us piece together different parts of the story. Future work integrating results from other RDoC domains (e.g. Sustained Threat (
), Self-Knowledge), will allow a deeper understanding of how different neurobiological profiles may interact (concurrently and over time) to increase risk in different adolescents. This integrative research may suggest novel approaches for using profiles of biological data to characterize clinically relevant subgroups of adolescents which could potentially be useful in guiding treatment selection. Furthermore, while the RDoC domains are notably distinct, these investigations may suggest points of overlap (e.g., neural hubs of intersecting networks) which could be potential high-impact treatment targets for future clinical trials. Finally, longitudinal integrative research will be critical to understand dynamics (how clinical and biological trajectories of risk and adaptation evolve and interact over the course of adolescence) that will inform timing of neurobiologically-informed interventions for youth with NSSI.
Groschwitz RC, Plener PL (2012): The Neurobiology of Non-suicidal Self-injury (NSSI): A review. Suicidol Online. Retrieved February 23, 2022, from http://selfinjury-staging.com/perch/resources/groschwitz.pdf
Schedule for Affective Disorders and Schizophrenia for School-Age Children-Present and Lifetime Version (K-SADS-PL): initial reliability and validity data.
J Am Acad Child Adolesc Psychiatry.1997; 36: 980-988
Development of Three Web-Based Computerized Versions of the Kiddie Schedule for Affective Disorders and Schizophrenia Child Psychiatric Diagnostic Interview: Preliminary Validity Data.
J Am Acad Child Adolesc Psychiatry.2020; 59: 309-325
The study was funded by the National Institute of Mental Health R01MH107394, NIBIB P41 EB027061, and 1S10OD017974-01. Trainee support was provided to Lauren Demers by the University of Minnesota’s Institute of Child Development via a National Institute of Mental Health National Research Service Award (T32MH015755-39) and the University of Minnesota's Graduate School Doctoral Dissertation Fellowship. Additional support was provided by the Minnesota Supercomputing Institute. The data reported here was presented at the 2022 Annual meeting for the Society of Biological Psychiatry. The authors would like to express their sincere thanks to the adolescents and families who gave their time and energy to helping complete this study.
Disclosures
The authors report no biomedical financial interests or potential conflicts of interest.
Prevalence and function of non-suicidal self-injury (NSSI) in a community sample of adolescents, using suggested DSM-5 criteria for a potential NSSI disorder.
Klimes-Dougan B, Eberly LE, Schreiner MW (2014): Multilevel assessment of the neurobiological threat system in depressed adolescents: Interplay between the limbic system and hypothalamic–pituitary–adrenal axis. Development. Retrieved from https://www.cambridge.org/core/journals/development-and-psychopathology/article/multilevel-assessment-of-the-neurobiological-threat-system-in-depressed-adolescents-interplay-between-the-limbic-system-and-hypothalamicpituitaryadrenal-axis/BF2442694658D9483F8657B8978080AB
Coordination between frontolimbic resting state connectivity and hypothalamic-pituitary-adrenal axis functioning in adolescents with and without depression.
Development and Definitions of the RDoC Domains and constructs (n.d.): National Institute of Mental Health (NIMH). Retrieved February 21, 2023, from https://www.nimh.nih.gov/research/research-funded-by-nimh/rdoc/development-and-definitions-of-the-rdoc-domains-and-constructs
Claes L, Muehlenkamp JJ (2013): Non-Suicidal Self-Injury in Eating Disorders: Advancements in Etiology and Treatment. Springer Science & Business Media. Retrieved from https://play.google.com/store/books/details?id=Jty4BAAAQBAJ
Testing the relations between impulsivity-related traits, suicidality, and nonsuicidal self-injury: a test of the incremental validity of the UPPS model.
Validation of an emotional stop-signal task to probe individual differences in emotional response inhibition: Relationships with positive and negative urgency.
Groschwitz RC, Plener PL (2012): The Neurobiology of Non-suicidal Self-injury (NSSI): A review. Suicidol Online. Retrieved February 23, 2022, from http://selfinjury-staging.com/perch/resources/groschwitz.pdf
Schedule for Affective Disorders and Schizophrenia for School-Age Children-Present and Lifetime Version (K-SADS-PL): initial reliability and validity data.
J Am Acad Child Adolesc Psychiatry.1997; 36: 980-988
Development of Three Web-Based Computerized Versions of the Kiddie Schedule for Affective Disorders and Schizophrenia Child Psychiatric Diagnostic Interview: Preliminary Validity Data.
J Am Acad Child Adolesc Psychiatry.2020; 59: 309-325
Lang PJ, Bradley MM, Cuthbert BN (2005): International Affective Picture System (IAPS): Affective Ratings of Pictures and Instruction Manual. NIMH, Center for the Study of Emotion & Attention. Retrieved from https://play.google.com/store/books/details?id=VEW2PgAACAAJ
Defining biotypes for depression and anxiety based on large-scale circuit dysfunction: a theoretical review of the evidence and future directions for clinical translation.
Wickham H (2009): ggplot2: Elegant Graphics for Data Analysis. Springer Science & Business Media. Retrieved from https://play.google.com/store/books/details?id=bes-AAAAQBAJ
Cai S, Guo Z, Wang X, Huang K, Yuan K, Huang L (2022): Cortical thickness differences are associated with cellular component morphogenesis of astrocytes and excitatory neurons in nonsuicidal self-injuring youth. Cereb Cortex. https://doi.org/10.1093/cercor/bhac103
Başgöze Z, Wiglesworth A, Carosella KA, Klimes-Dougan B, Cullen KR (2021): Depression, Non-Suicidal Self-Injury, and Suicidality in Adolescents: Common and Distinct Precursors, Correlates, and Outcomes. J Psychiatr Brain Sci 6. https://doi.org/10.20900/jpbs.20210018
00044-7&id=gr1.jpg){kind=link}
00044-7&id=gr2.jpg){kind=link}
00044-7&id=gr3.jpg){kind=link}
00044-7&id=gr4.jpg){kind=link}