Existing evidence suggests that the predisposition to suicidal and non-suicidal behaviors in adolescents is at least partly mediated by neurobiological factors. The altered cerebral function might play a key biological role in NSSI and SA of depressed adolescents [15–17, 36]. The present study was the first to examine both convergent and divergent neurobiological alternations that underpin adolescent NSSI with or without a history of SA. Compared to the clinical control group, we identified widespread aberrant intrinsic network FC patterns in the NSSI group. These alternations were not only between the three core neurocognitive intrinsic brain networks (the SalVAttn, Control, and Default networks) proposed as the triple network model of psychopathology [37], but also between these core neurocognitive intrinsic brain networks and other networks. These findings suggest that NSSI in depressed adolescents is associated with disruptions in the coordination and communication between different brain networks involved in cognitive and emotional processing. Referring to the alternations of FC in the NSSI group, the alternations of intra-/inter-network among the Control, SalVAtten, Limbic, and Default networks persist in depressed adolescents with a history of both NSSI and SA. Additionally, new FC alternations emerged, particularly including inter-network connections among the Limbic, DorsAttn, and Visual networks. These findings suggest that the combination of NSSI and SA may further influence the connectivity patterns within and between specific brain networks. Our results also expand on the link between emotional regulation deficits and suicidal behaviors by demonstrating the difference in resting-state FC between NSSI and NSSI + SA related to the Subcortical and Limbic networks. Besides, the severity of depressive symptoms only showed a significant correlation with altered FCs in Limbic-DorAttn and Limbic-Visual, strengthening the fact that these observed alternations could not all be explained by increased depression severity. We discuss potential interpretations of these findings below.
Our focus on networks is important, not only because individual brain regions do not function in isolation, but also because network maturation is a central feature of brain development during adolescence [38, 39]. Studies of normative development have shown that functional brain networks increasingly segregate through adolescence [40], making it a vulnerable period for emotional problems [41]. Importantly, the networks that support these processes continue to mature in adolescence [38, 39] and are aberrant in depressed adolescents [42]. Our account of NSSI in adolescence is supported by a substantial body of developmental cognitive and affective neuroscience [43]. According to existing evidence using predefined ROIs, the abnormal FC tended to be spatially distributed, highlighting the importance of considering whole-brain functional networks. Menon et al. critically reviewed the fMRI literature and proposed a triple-network model (SalVAttn, Control, and Default network) for neuropsychiatric disorders and these three networks are associated with emotional regulation and cognitive processing [37]. NSSI or SA is a maladaptive strategy that stems from greater difficulty with emotional awareness and regulation [44]. Specific patterns of network dysfunction of core neurocognitive networks may contribute to core deficits in cognitive and affective functioning that are believed to underlie these behaviors.
The Affective Network (AN) and the Ventral Attention Network (VAN), collectively referred to as the SalVAttn network, are key brain networks involved in emotion processing and monitoring salient events [18]. The SalVAttn network, primarily centered in the anterior insular and cingulate cortex, plays a crucial role in identifying biologically and cognitively salient events that are essential for directing attention and guiding goal-directed behaviors [45]. In the NSSI compared with the Control, we found decreased intra-modular FCs of the SalVAttn and inter-network FCs between the SalVattn and Control networks, and increased FCs between the SalVattn and Limbic networks. Increased sensitivity to socioaffective pain contributes to the onset of NSSI in adolescence and it is largely believed that it is a product of an imbalance between elevated response to affective information in the environment by the SalVAttn network [46, 47], and blunted regulation by the Control network associated with effortful emotion regulation and cognitive control [48]. Besides, the Salience network also plays a critical role in processing emotion in connection with the Limbic network [45]. Therefore, increased FC between the SalVAttn and Limbic networks could be explained by the excessive emotional responses to salient external stimuli (i.e., behaviorally relevant and potentially threatening), and SalVAttn-Control hypoconnectivity may represent difficulty in the regulation of negative affect in adolescents with NSSI.
Adolescents with NSSI also showed alternations in inter-network FCs of Limbic with the Control, Default, and DorsAttn networks. The limbic network has two main nodes including the orbitofrontal cortex (Limbic-OFC) and temporal pole (Limbic-TempPole) [18]. OFC is a key region for integrating emotional information and is associated with emotional regulation in emotional disorders. In NSSI participants, greater activation was found in the OFC during the emotional processing task [49–51]. As part of an extended limbic system, the temporal pole has some role in both social and emotional processes, and dysfunction of the temporal pole has been observed in clinical disorders with socioemotional regulation [52]. The Control network is involved in the top-down regulation of attention and emotion [53], and the Default [54] and DorsAttn [55] networks are involved in internally or externally oriented attention, respectively. Therefore, our results here are in congruent with the deficits in the socioemotional regulation associated with NSSI.
The relationship between NSSI and SA is complicated and both of them lie along a spectrum of self-destructive actions [28]. Initial findings support the view that NSSI and SA in adolescents are characterized by abnormal resting-state FCs, but inconsistent patterns of alterations were reported for adolescents with NSSI and SA before [16]. NSSI is a key risk factor for future SAs, and some of the same neurobiological systems have been implicated in both NSSI and SA. According to our results, both NSSI and NSSI + SA groups showed increased FC in Limbic-Control, and Limbic-SalVAttn, and decreased FC in SalVAttn and Limbic-Default. These observations emphasize theories about a potential continuum of self-destructiveness [56] and indicate that indeed there is a potential neurobiological underpinning of the continuum from self-injury to suicidal behavior. Furthermore, it has been suggested that individuals who engage in both NSSI and SA may present with more complex psychopathology [57]. In addition to the above-mentioned convergent alternations, the divergent alternations for adolescents who engage in both NSSI and SA were observed involving the dorsal Visual network, further emphasizing the complexity of their neural connectivity patterns. In the context of NSSI, adolescents show difficulties with social cue interpretation and deficits in interpersonal problem-solving, and they are prone to negatively interpreting social situations [58]. There have been consistent findings that adolescents with repeated NSSI who additionally reported a history of SAs presented greater severity of psychopathology such as a lower level of functioning compared to those with NSSI only [12]. During the maintenance of NSSI, heightened negative bias may contribute to the development of SA [57]. Hence, our findings of alternation FC associated with the Visual network in the NSSI + SA group could be interpreted from an “information processing” perspective [59, 60]. Vision is an important part of the selective attention process, and the dorsal visual network is associated with spatial awareness and guidance of actions [61]. Hence, the processing bias in adolescents with NSSI + SA may be initiated as a perceptual visual bias, and then eventually cause a series of cognitive and affective symptoms.
Furthermore, when comparing the NSSI group and the NSSI + SA group directly, both groups reported emotion regulation as their major motivation for NSSI behaviors, and the NSSI + SA group reported a higher rating for the attention-seeking factor. The stronger need for attention-seeking for adolescents with SA was in consistent with Joiner’s “interpersonal-psychological theory of suicidal behavior”, which states belongingness as one of the main predictors of suicidal behavior [56]. Emotion regulation was the most common reason for NSSI consistently suggested in many previous studies [62]. While a basic understanding of the relationship between NSSI and emotion regulation deficits has been established, one recently published study supported the possibility of unique associations between emotion regulation deficits with more frequent NSSI engagement and more suicidal attempts [63]. According to our results, the NSSI + SA group reported a higher frequency of self-injury behaviors, which means that suicidal behavior may gradually become another coping strategy due to the more pronounced habituation to pain and fear because of repeated self-injury behaviors when emotion dysregulation persists. Besides, the differences in resting-state FC between the NSSI group and the NSSI + SA group were mainly related to the basal ganglia of the Subcortical network. The basal ganglia traditionally have been assigned to roles within the motor domain, however, the basal ganglia have connections with a broad limbic network and yet recent research has proposed its contribution to a variety of limbic and cognitive processes including emotion recognition, reward- and decision-making [64]. Therefore, the current study expands on the link between emotional regulation deficits and suicidal behaviors by demonstrating the difference of resting-state FC related to the Subcortical and Limbic networks. Since the self-reported questionnaire for emotion regulation did not yield significant differences among groups, future studies could use a more objective assessment of emotion regulation function to validate the results observed here.
There are some limitations. First, the cross-sectional design limits our ability to establish the causal relationship or determine the development trajectory of the observed aberrant patterns. To fully comprehend the emergence and progression of these patterns, longitudinal studies that track individuals over time are needed. Secondly, the reliance on retrospective self-report to collect data on self-harm behaviors and depression symptoms introduces the potential reporting or recall biases, which may affect the accuracy and reliability of the obtained information. Future research could consider incorporating more objective measures to mitigate these biases. Thirdly, the sample of depressed adolescents limits the generalizability of the identified FC patterns to a broader population of adolescents engaging in NSSI and SA. As a related point, this study included both adolescent females and males (although mostly females). Given consistently observed sex differences in the prevalence of self-injurious behaviors [65] and brain function during adolescence [66], there may be sex differences in the neural mechanisms of NSSI and SA. To gain a more comprehensive understanding, future studies should investigate these neural circuits regarding different psychiatric conditions and potential sex differences. By addressing these limitations, we can further enhance our understanding of the neural mechanisms underlying NSSI and SA.