Nonsuicidal self-injury (NSSI; all abbreviations are listed in Table 1), defined as the deliberate infliction of harm to one’s body tissue without suicidal intent (1), predominantly manifests in adolescents and often co-occurs with depression, with the most common form being cutting (2–4). The global lifetime prevalence of NSSI among adolescents was estimated to be around 20% (5, 6), peaking in the age range of 15–17 years (7). Previous research indicates that females were more likely to engage in NSSI behaviors than males among adolescents in both community- (2, 8) and clinical- population (4). The emergence and persistence of NSSI during adolescence exert various adverse effects on individuals’ physical and mental health, including may heighten the risk for emotional dysregulation and even increased suicidal tendencies (9–11), thereby increasing the burden on social services and health resources.

Neuroimaging studies have significantly advanced our understanding of the neural underpinnings of NSSI in adolescents. Accumulating evidence suggests that NSSI is not merely a behavioral symptom, but a condition marked by measurable disruptions in brain connectivity—encompassing both functional networks and structural pathways. At the functional level, previous research has established that adolescents with NSSI exhibit widespread functional connectivity (FC) abnormalities, particularly in cortical-subcortical communication pathways. For example, some studies have identified aberrant resting-state FC between the amygdala and prefrontal areas and dorsal anterior cingulate cortex (12), as well as between the orbitofrontal cortex and insula (13), the latter predicting future self-injury frequency in depressed youth. Such findings extend to reward-related circuits, where dysregulated connectivity may underlie the addictive nature of some NSSI behaviors (14). Task-based studies further reveal deficits in reward processing and negative emotion regulation among adolescents with NSSI, linked to altered connectivity involving regions such as the orbitofrontal cortex (15) and the amygdala (12). Notably, FC patterns may be linked to the clinical prognosis of NSSI. For example, prefrontal-amygdala connectivity has been shown to predict responsiveness to both psychological and pharmacological interventions, suggesting its potential as a neural marker for forecasting treatment outcomes in NSSI (16, 17). On a broader scale, network-level disruptions are also evident. Adolescents with NSSI have been found to exhibit reduced coherence within major intrinsic networks such as the default mode and salience networks (18), suggesting widespread integration deficits in systems supporting self-referential and attentional processing. Structurally, adolescents with NSSI demonstrated widespread white matter microstructural deficits, with longer duration of self-injury associated with more severe impairment in white matter integrity (19). Together, neuroimaging evidence consistently points to extensive disruptions in brain connectivity among adolescents with NSSI, which may be closely linked to impairments in emotional processing, cognitive control, and related functions.

To foster further research into NSSI, the fifth edition of the Diagnostic and Statistical Manual of Mental Disorders (DSM-5) has incorporated NSSI into its recommended diagnoses, recognizing it as a distinct clinical disorder separate from other self-destructive behaviors. Among the proposed criteria, self-injurious behavior should occur on 5 or more days in the past year. The DSM-5 also outlines functional and intentional aspects relevant to diagnosis. However, both community-based and clinical studies indicate that many adolescents who self-injure do not meet the criteria for NSSI outlined in the DSM-5. Meta-analyzes suggest that among community-based adolescent populations, those who have engaged in self-injurious behaviors more than five times in their lifetime account for less than 30%, and only about 5% have engaged in such behaviors more than 10 times (2). Similarly, data from a public hospital in Mexico indicate that approximately 40% of adolescent patients presenting with self-injury do not satisfy the diagnostic criteria for NSSI (20).

Existing research has begun to differentiate between occasional and repetitive self-injury, which roughly align with subthreshold and suprathreshold frequencies relative to DSM-5 recommendations. For example, Brunner et al. (21) found that social factors were associated with occasional but not repetitive self-harm. Liu et al. (22) reported that impulsivity moderated the relationship between emotion dysregulation and NSSI only in the repetitive NSSI group. Additionally, Liu et al. (23) observed that both maladaptive cognitive schemas and emotion dysregulation mediated the link between stressful life events and addictive features of NSSI in adolescents with repetitive NSSI—a pattern not seen in those with occasional NSSI. Biologically, a recent study introduced the concept of “subthreshold NSSI” to describe individuals engaging in self-injury fewer than five times per year. This work revealed abnormalities in low−frequency oscillations in the left thalamus and posterior cingulate cortex, altered FC within the left thalamus, and reduced abundance of Prevotellaceae bacteria in the gut among subthreshold NSSI patients (24). While this study represents the first exploration of biological alterations in individuals who do not meet the DSM−5 diagnostic threshold for NSSI, it did not directly compare subthreshold and suprathreshold NSSI groups. Consequently, whether and how adolescents who meet the DSM-5 diagnostic threshold for NSSI differ biologically from those who do not still remains unclear.

As previously noted, NSSI is increasingly understood as a disorder characterized by disruptions in brain connectivity. However, the majority of earlier neuroimaging studies have examined brain alterations in NSSI from either a structural or a functional perspective in isolation. Researchers have established that structural and functional connectivity are highly interdependent—brain activity both emerges from and is constrained by underlying structural pathways (25, 26). Notably, a seminal finding revealed that the human brain exhibits a gradual structural-functional gradient, shifting from strong coupling in highly myelinated sensorimotor cortices toward greater decoupling in less myelinated associative regions (27). This organizational principle aligns with gradients observed in other biological modalities, such as gene expression (28, 29) and microstructural properties (30). These insights stem from structural-functional connectivity (SC-FC) coupling analysis based on magnetic resonance imaging (MRI). Conceptually, SC-FC coupling reflects the degree to which functional signaling depends on anatomical structure, often quantified as the smoothness of functional signals over the structural graph. Preti et al. (27) further demonstrated that brain regions supporting primary sensory and motor functions exhibit stronger SC-FC coupling, whereas areas involved in higher-order cognitive and affective processes—such as memory, reward, and emotion—show greater decoupling.

Importantly, altered SC-FC coupling has been implicated in the neuropathology of several psychiatric disorders, including bipolar disorder (31) and adolescent depression (32). Notably, Xu et al. (32) identified distinct SC-FC coupling profiles in limbic regions between depressed adolescents with and without NSSI behavior. Thus, SC-FC coupling analysis offers a novel integrative framework for investigating the neural substrates of NSSI, shifting the focus from isolated connectivity measures to the dynamic interplay between brain structure and function.

Therefore, from a structural-functional integration perspective, this study aims to investigate—through SC-FC coupling analysis—whether adolescents whose NSSI meets the DSM−5 diagnostic threshold exhibit distinct neurobiological substrates compared to those with subthreshold NSSI. Given the known heterogeneity of NSSI related to gender and comorbid psychiatric conditions, this work focuses specifically on female adolescents with depression. By leveraging multimodal neuroimaging, the study seeks to provide new insights into the neurobiological distinctions between clinically defined subtypes of NSSI in youth.

This experiment conformed to the requirements of the Declaration of Helsinki and was approved by the Ethics Committee of the Chengdu Fourth People’s Hospital. All participants and their guardians had thoroughly read and signed the informed consent form. All participants were recruited from December, 2021, to September, 2024.

This study represents a significant advancement in understanding the neural mechanisms underlying NSSI among depressive female adolescents. By classifying self-injurious participants into NSSI+ (meeting DSM-5 criteria) and NSSI- (without meeting DSM-5 criteria) groups, we aimed to elucidate distinct neural correlates associated with these subtypes. Our analysis focused on SC-FC coupling, providing novel insights into the neurobiological differences between these groups.

The primary finding of our study is the identification of specific brain regions where SC-FC coupling differs significantly between NSSI+ and NSSI- groups. These regions include the left dorsolateral SFG, right opercular IFG, and bilateral medial precuneus. Based on the Yeo 7-network framework, these areas are predominantly implicated in the brain’s dorsal attention network (DAN), ventral attention network (VAN), and frontoparietal network (FPN), all of which are typical task-positive networks and play critical roles in cognitive control, attentional processes, and goal-directed behaviors (43, 44).

The left dorsolateral SFG, overlapping with the dorsolateral prefrontal cortex (DLPFC), showed a more pronounced state of SC-FC decoupling in the NSSI+ group compared to the NSSI- group. The DLPFC is central to “cold” executive functions—higher-order cognitive operations that are relatively independent of emotional or reward-based input, such as working memory, goal-oriented planning, attentional control, and response inhibition (45–47). In contrast, “hot” executive functions involve affective or reward-related processing. Prior work has linked NSSI in depression to impairments in cognitive control, including reduced goal-directed behavior (48) and diminished cognitive flexibility (49), along with altered DLPFC activity during cognitively demanding tasks (50–52). The observed decoupling in the left dorsolateral SFG may therefore reflect gradations in cognitive-attentional and executive processing underlying different severity levels of NSSI in depressed adolescent females.

Similarly, the right opercular IFG, part of the VAN, exhibited increased SC-FC coupling in the NSSI+ group relative to the NSSI- group. The VAN is crucial for bottom-up attentional processes, such as attentional shifts (43). The IFG has reliable connections with both the supplementary motor area (SMA), supporting its crucial involvement in flexible cognitive-motor inhibition, such as the inhibition of premature or no longer appropriate motor responses (53, 54). Previous studies have reported reduced right IFG volume and abnormal error-related activation during Go/No Go task in individuals engaging in self-injurious behaviors (55, 56). Our results extend these observations by highlighting differences in SC-FC coupling properties of the right opercular IFG, which may indicate variations in cognitive motor inhibition functions among depressive female adolescents with different levels of self-injury severity. Furthermore, a negative correlation emerged between the extent of decoupling in this region and the baseline total score of NSSI functions within the NSSI+ group. We postulate that the functions or motivations underlying self-injurious behavior could cognitively intensify such behavior, resulting in subsequent difficulties in inhibiting self-harm. This process may be intricately linked to an aberrant over-coupling of structure and function in the right opercular IFG.

The precuneus was initially considered a brain region primarily associated with visual processing. Recent imaging studies highlight its more crucial functions in complex cognitive processes, including risk/reward assessment, working memory updating, and self-referential thinking, due to its extensive connections with large-scale brain networks such as the default mode network (DMN) and FPN (57). Our findings showed increased SC-FC coupling of the bilateral medial part of precuneus, which is intricately involved in the FPN (58, 59), in the NSSI+ group compared to the NSSI- group. Moreover, our study found a negative link between the decoupling degree of the right medial precuneus and the sensation-seeking factor of NSSI functions in the recent month within the NSSI- group. Sensation-seeking, a core trait in addictive disorders, has been linked to NSSI, which some view as a behavioral addiction (60, 61). Similarly, patients with substance addiction have shown to exhibit medial precuneus and FPN abnormalities (62, 63). Thus, it is plausible that the abnormal precuneus coupling might underlie vulnerability to sensation-seeking-induced self-injury behavior in less severe NSSI individuals. Further research is needed to clarify if this extends beyond pain perception. Interestingly, Baum et al. (64) reported that age-related increases in coupling were particularly prominent within brain regions of the DMN and FPN, including the precuneus. They emphasized that delayed development of coupling in these brain networks might provide a critical window for neural remodeling during childhood and adolescence. It is plausible that the excessive SC-FC coupling of the precuneus observed in NSSI+ adolescents in our study may reflect impaired neural plasticity, which could contribute to the escalated severity of self-injury behavior.

It has been summarized that neural structural and functional alterations in NSSI mainly involve the fronto-limbic-striatal system (65). Xu et al. (32) found increased SC-FC coupling in the right insula and left thalamus among depressed adolescents without NSSI behaviors compared to those with NSSI and HCs. In our study, we demonstrated that the depressive female adolescents with NSSI+ behavior showed widespread coupling abnormalities spanning the somatomotor network, FPN, DMN, and VAN versus the HC group, while those with NSSI- behavior only in the left dorsolateral SFG. These findings underscore subthreshold NSSI as a potential critical period for intervention, during which neurobiological alterations may still be reversible. Notably, we observed a dissociation in the left dorsolateral SFG: while the NSSI- group differed from HC in this region, the NSSI+ group did not. This paradoxical finding suggests a potential compensatory mechanism in which initial coupling disruptions (seen in NSSI- individuals) may normalize or shift to other regions as NSSI behaviors escalate.

Interestingly, correlational analyzes revealed several significant associations between decoupling indices and clinical features, especially in the right sgACC among the NSSI- participants, a key region which has recently been found to be able to enhance the precision and therapeutic efficacy of DLPFC targeting for depression treatment (66–68). Given the common comorbidity of depressive symptoms in NSSI adolescents and the empirical evidence supporting that the alleviation of depressive symptoms contributes to NSSI behavior improvement (69, 70), our findings highlight a promising approach, using dual-region modulation of superficial DLPFC and deeper sgACC via repetitive transcranial magnetic stimulation (rTMS), for early intervention of NSSI.

Certain limitations must be acknowledged in the present study. First, only female participants were included, and whether similar brain alterations exist in male self-injury populations remains unknown. Second, the majority of NSSI participants in this study had a background of depression. It is difficult to exclude the influence of affective disorders due to the absence of a non-NSSI depressive group. Third, while psychiatrists grouped participants according to the suggested diagnostic criteria for NSSI in DSM-5, there is currently no Structured Clinical Interview for DSM-5 specifically targeting NSSI diagnosis. Therefore, the subjective influence of psychiatrists cannot be entirely ruled out during group classification. Lastly, this study cannot determine whether the abnormal SC-FC coupling represents a state feature or a trait feature of different NSSI subtypes. Future research needs to delve deeper into the association between the brain SC-FC coupling characteristics and the dynamic changes in NSSI behaviors from a longitudinal perspective.

The present study demonstrates that SC-FC coupling analysis offers a sensitive, integrative framework for distinguishing between clinically relevant subtypes of NSSI in female adolescents with depression. By comparing individuals whose NSSI meets the DSM-5 diagnostic threshold with those exhibiting subthreshold symptoms, we identified subtype-specific alterations in SC-FC coupling within key cortical regions—including the left dorsolateral SFG, right opercular IFG, and bilateral medial precuneus. These regions are predominantly embedded in the task-positive networks, which collectively support executive control, attentional allocation, and goal-directed behavior. Importantly, the NSSI+ group exhibited more extensive and severe coupling abnormalities, whereas the NSSI- stage may represent a critical window for early intervention during which neural dynamics may still be modifiable. The right sgACC, which demonstrated significant correlations with several clinical features in the NSSI− group, emerges as a particularly promising neurobiological target for early intervention strategies.

To build upon the present findings, future research should adopt longitudinal designs to clarify whether SC-FC coupling alterations reflect state or trait characteristics of NSSI, and expand investigations to include males and non-depressed NSSI populations to improve generalizability. The development of standardized diagnostic tools for NSSI would enhance classification reliability, while studies linking baseline coupling profiles to treatment outcomes—particularly through targeted neuromodulation aimed at normalizing aberrant coupling patterns—could advance personalized intervention strategies. Further integration of multimodal data and translational applications may ultimately support the use of SC-FC coupling as a biomarker for risk stratification and treatment monitoring in youth with NSSI.