Salience and default mode networks in 9 and 10-year-olds with and without suicide ideation

• In Event: Functional brain predictors and correlates of suicidal ideation and non-suicidal self-injury in adolescents

Abstract

Over the past decade, suicidal thoughts and behaviors (STBs) have increased in prevalence in children and adolescents (CDC, 2017). Advancing our understanding of the developmental roots of suicide risk is critical for prevention efforts. While there is a growing body of work linking brain functioning and STBs in adolescents and adults, little is known about the neurobiological underpinnings of suicide risk in children. The current study aimed to characterize key neural network patterns in children who report suicide ideation (SI) compared to their non-ideating peers, while at rest and during tasks involving emotionally salient stimuli. We hypothesized that, in line with previous neuroimaging work in both adolescents and adults, SI would be associated with 1) lower resting-state functional connectivity (RSFC) of the default mode (DMN) and salience networks (SN) and 2) different activation patterns while viewing emotional stimuli, however we did not predict the direction of these patterns.

We analyzed data from the 2.0.1 ABCD study release (N=8,293 children ages 9-10; 48.53% female). Participants underwent fMRI, completing a resting-state scan and Emotional N-back task. We examined within-network RSFC and activation to emotional and neutral stimuli within the DMN and SN (defined by the Gordon and Destrieux atlases for resting-state and activation data respectively). SI (present or absent) and relevant clinical diagnoses were assessed with the Kiddie Schedule for Affective Disorders and Schizophrenia for DSM-5.

Linear mixed-effect regression models examined associations between SI and SN and DMN RSFC, and associations between SI and SN and DMN activation during emotional (positive; negative) versus neutral EN-back conditions. All models controlled for demographic factors, ABCD Study sampling design, and clinical diagnoses. We corrected for multiple comparisons using the Benjamini-Hochberg false discovery rate procedure and report adjusted p-values. To examine the replicability of our model results, repeated split-half reliability analyses (100-iterations) were conducted.

Compared to those without ideation (n=8,123), children with current SI (n=170, 2.1%) showed lower SN RSFC (b=-.046, p=.007), even when controlling for clinical diagnoses (b=-.042, p=.017). Compared to those without ideation, children with current SI also showed lower DMN RSFC (b=-.014, p=.045); this association, however, was no longer significant when controlling for clinical diagnoses (b=-.010, p=.138). Compared to those without ideation, children with current SI exhibited lower activation in regions of the DMN when viewing negative versus neutral stimuli (b=-.10, p=.039) but no differences when viewing positive versus neutral stimuli (b=-.08, p=.078); the significance of these associations were unchanged when controlling for clinical diagnoses. All models generally replicated in split-half analyses.

We report that SI is present even in children as young as 9 years old, thereby underscoring the importance of identifying the neural correlates of SI during later childhood and early adolescence. Consistent with findings in older adolescents, we found that current SI is associated with reduced intrinsic RSFC and activation to negatively valanced stimuli in regions comprising the DMN and SN. Our findings fill an important gap on neural correlates of SI and suicide risk in young adolescents by identifying potential targets for prevention or intervention.

Authors

• Andrea Wiglesworth, University of Minnesota - Twin Cities
  Presenting Author

• Conner A Falke, University of Minnesota - Twin Cities
  Non-Presenting Author

• Mark B. Fiecas, University of Minnesota
  Non-Presenting Author

• Monica Marie Luciana, University of Minnesota–Twin Cities
  Non-Presenting Author

• Kathryn Cullen, University of Minnesota
  Non-Presenting Author

• Bonnie Klimes-Dougan, University of Minnesota - Twin Cities
  Non-Presenting Author