Impulsivity in adolescents with MDD and non-suicidal self-injury: a multimodal assessment using psychometric, behavioral, and neurophysiological measures

Abstract

Background

To investigate multidimensional impulsivity and its neuroelectrophysiological correlates in adolescents with depression and non-suicidal self-injury (NSSI).

Methods

A case–control study was conducted including three groups: depressed adolescents with non-suicidal self-injury (NSSI group), depressed adolescents without NSSI (non-NSSI group), and healthy controls (HCs group). The Ottawa Self-Injury Questionnaire was administered to assess NSSI behaviors in depressed participants and to quantify NSSI severity within the NSSI group. Impulsivity was evaluated using a multimodal approach, including the Barratt Impulsiveness Scale-11 (BIS-11), a go/no-go task, and event-related potentials (ERPs), with particular focus on the N2 and P3 components.

Results

Both depressed groups exhibited significantly higher BIS-11 total and subscale scores than HCs, particularly in motor impulsivity (p < 0.05). A significant overall group difference was observed in go trials accuracy among the three groups (p = 0.009). Post hoc analyses with Bonferroni correction showed that the NSSI group had significantly lower go-trial accuracy than the HCs group (p_Bonf < 0.05), whereas no significant differences were observed between the NSSI and non-NSSI groups (p_Bonf > 0.05) or between the non-NSSI and HCs groups (p_Bonf > 0.05). No significant group differences were found in N2 or P3 latency or amplitude across the three groups (all p > 0.05). Correlation analyses revealed that NSSI severity over the past 1, 6, and 12 months was positively correlated with BIS-11 total and motor impulsivity scores (r = 0.198–0.417, all p < 0.001) and negatively correlated with go-task accuracy (r =  − 0.248 to − 0.193, all p < 0.001).

Conclusion

Depressed adolescents with NSSI exhibited elevated self-reported impulsivity and impaired behavioral inhibitory control, both of which were associated with NSSI severity, whereas no significant neurophysiological differences were observed between depressed adolescents with NSSI and HCs.

Supplementary Information

The online version contains supplementary material available at 10.1186/s13034-026-01043-2.

Background

Non-suicidal self-injury (NSSI) refers to the deliberate, direct damage to body tissues without suicidal intent. This behavior encompasses a range of actions, including cutting, scratching, or burning the skin. It also includes indirect forms, such as forcefully striking hard surfaces (e.g., walls), which may result in injury to the skin or bones [1]. The prevalence of NSSI among Chinese students has become an increasing public health concern that warrants urgent attention. A comprehensive meta-analysis of 420 studies involving 160,348 Chinese students reported a pooled prevalence rate of NSSI of 22.37%, with a higher prevalence in females in (21.9%) than in males (20.6%) [2]. The rate escalates to 52% adolescents diagnosed with depression and other mood disorders [3]. Although NSSI typically occurs in the absence of suicidal intent, accumulating evidence indicates that it constitutes a significant risk factor for subsequent suicide attempts [4–6]. Moreover, NSSI is strongly associated with later suicidal behavior, underscoring its critical role in the progression from self-harm to more life-threatening actions attempts [4–6]. Recent studies have further demonstrated a time-dependent relationship between the onset of NSSI and future suicidal behavior, showing that an earlier initiation of NSSI is significantly associated with an increased risk of future suicide attempts [7]. Although not all individuals who engage in NSSI progress to suicidal behavior, many experience heightened emotional distress and persistent difficulties in emotion regulation [8]. Given the high prevalence of NSSI and its far-reaching adverse consequences, identifying the risk factors associated with NSSI engagement has become an urgent and paramount research priority.

Numerous factors [9, 10] have been implicated in the occurrence of NSSI, among which impulsivity stands out as a particularly critical contributor. Impulsivity is typically defined as a predisposition to respond rapidly and spontaneously to internal or external cues, often without adequate consideration of the potential negative consequences of one's actions, either for themselves or for others [11]. Impulsivity can be assessed at multiple levels. The most basic approach relies on self-report questionnaires, whereas more objective evaluations employ behavioral tasks. At the neurophysiological level, event-related potentials (ERPs) offer a direct and sensitive measure of capturing the neural processes underlying impulsivity-related behavior. Integrating these complementary measures enables a multidimensional assessment of impulsivity, providing a more complete understanding of its behavioral manifestations and underlying neural mechanisms in adolescents with depression and NSSI.

From a psychological perspective, Nock [12] emphasized that individuals with elevated impulsivity may be at increased risk of NSSI, particularly during emotionally challenging situations, as such acts can be carried out quickly and without substantial forethought or planning. This theoretical framework has prompted extensive investigation into the role of impulsivity in NSSI; however, empirical findings based on self-report and behavioral measures remain inconsistent [13–16]. Some studies have reported higher levels of self-reported and behaviorally assessed impulsivity among individuals with NSSI [16, 17], whereas others have failed to identify consistent group differences [13–15]. For instance, Janis et al. [16] reported that both adolescents and adults with NSSI exhibited higher levels of self-reported impulsivity, whereas no significant differences were observed in behavioral measures. In contrast, Claes et al. [13] found no significant differences in self-reported impulsivity between patients with eating disorders with and without NSSI, although those with comorbid NSSI demonstrated significantly greater behavioral impulsivity. Collectively, these mixed findings suggest that impulsivity in NSSI is a complex and multifaceted construct that may not be fully captured by self-report questionnaires or overt behavioral performance alone.

At the cognitive level, impulsivity is closely related to impairments in inhibitory control, defined as the ability to suppress irrelevant information or prepotent responses in order to achieve goal-directed behavior [18]. Event-related potentials (ERPs), with their high temporal resolution and non-invasive nature, are widely used to investigate the neural mechanisms underlying inhibitory control. Among ERP components, N2 and P3 are the most commonly examined indices of inhibitory regulation [19, 20]. In the go/no-go task, the N2 component is typically observed as a negative deflection occurring approximately 200–350 ms after stimulus onset. Within this paradigm, the N2 component is generally interpreted as reflecting early-stage conflict monitoring and the engagement of attentional control processes elicited by task-relevant stimuli, particularly when a dominant response must be withheld [19, 21]. In contrast, the P3 component, which usually emerges between 300 and 600 ms following stimulus onset, is associated with later-stage inhibitory processing. Specifically, the no-go P3 is thought to index the allocation of attentional resources required for successful response inhibition, as well as the updating of working memory representations related to task demands [20].Together, the N2 and P3 components reflect sequential stages of inhibitory control, with N2 indexing early control engagement and P3 reflecting the execution of inhibitory processes [22]. To date, research examining neurophysiologically assessed impulsivity in NSSI is limited. One ERP study in college students reported that individuals engaging in NSSI showed heightened impulsivity-related neural markers, including significantly larger N2 amplitudes in the prefrontal region and prolonged N2 latency during no-go trials, compared with controls [23]. Similarly, another study found that children with a history of NSSI exhibited a more negative feedback negativity difference (ΔFN; losses minus gains) than those without NSSI, a pattern interpreted as reflecting enhanced impulsivity and altered reward processing [24].

Taken together, the existing literature presents conflicting findings regarding the association between NSSI and impulsivity. While some studies suggest that individuals with NSSI exhibit elevated impulsivity, others have failed to replicate these results, highlighting ongoing debates about the nature and extent of this relationship across different dimensions. Moreover, neurophysiological evidence examining impulsivity in NSSI—particularly in clinical adolescent populations—remains scarce. Clarifying whether impulsivity-related ERP markers reliably characterize adolescents with NSSI may help reconcile inconsistencies in prior behavioral studies and advance understanding of the underlying neurocognitive mechanisms. Therefore, the present study aimed to comprehensively examine impulsivity across multiple dimensions in adolescents with NSSI, including self-reported, behavioral, and neurophysiological measures. We hypothesized that recurrent NSSI would be closely associated with heightened impulsivity, reflected by increased self-reported impulsivity, greater behavioral disinhibition, and distinct ERP patterns.

Material and methods

Participants

Between October 2020 and December 2021, adolescents diagnosed with major depressive disorder (MDD) were recruited from the Department of Child and Adolescent Psychiatry, the Second Xiangya Hospital, Central South University. All patients with MDD met the following inclusion criteria:(1) age 12–17 years; (2) meeting DSM-5 diagnostic criteria for MDD, with a Hamilton Rating Scale for Depression (HAMD-17) score ≥ 17 and a Young Mania Rating (YMRS) score ≤ 5; (3) right-handed, with normal uncorrected or corrected vision; and (4) ability to understand the informed consent form and study procedures. Participants in the NSSI group had engaged in NSSI at least five times within the past 12 months, with at least one episode occurring in the past month [25]. Those with no history of NSSI were assigned to the non-NSSI group. Patients with MDD were excluded if they: (1) had attempted suicide within the past month; (2) had organic mental disorders, schizophrenia, mental retardation, autism, tic disorder, personality disorder, substance or drug dependence and other mental disorders; (3) had physical diseases that could affect the assessment and testing; (4) had a history of craniocerebral trauma, epilepsy and other relevant neurological conditions. For comparison, healthy adolescents were recruited from local high schools in Changsha City as healthy controls (HCs). The inclusion criteria for HCs were (1) age 12–17 years; (2) no current or past history of mental disorders, as confirmed by the Mini International Neuropsychiatric Interview for Children and Adolescents, 5th edition (MINI-Kid 5.0); (3) No history of NSSI; (4) no family history of psychiatric disorders; (5) right-handed, with normal or corrected-to-normal vision; (6) ability to understand the informed consent form and study procedures. Exclusion criteria for HCs included: (1) physical illnesses that could affect assessments or testing; (2) a history of cranial trauma, epilepsy, or other relevant medical conditions; (3) any other conditions that might hinder completion of the study.

This study was approved by the Ethics Committee of the National Clinical Medical Research Center of Central South University in accordance with the Declaration of Helsinki. Written informed consent was obtained from all participants and their legal guardians prior to enrollment. All participants underwent clinical interviews and assessments conducted by two certified child psychiatrists, both of whom completed standardized training before the commencement of the study.

Methods

General background information and clinical assessment

General demographic and background information was collected for all the participants, including sex, age, family history of mental disorders, only child status, and years of education. For patients with MDD, additional variables were gathered, including the number of depressive episodes and the type of depression (recurrent vs. first episode). The Mini International Neuropsychiatric Interview for Children and Adolescents, 5th edition (MINI-Kid 5.0), was administered to exclude comorbid psychiatric disorders in the MDD groups and to confirm the absence of psychiatric diagnoses in healthy controls. The 17-item Hamilton Rating Scale for Depression (HAMD-17) [26] and Young Mania Rating Scale (YMRS) [27] were used as part of the inclusion criteria for the MDD group. The severity of NSSI was assessed using the Ottawa Self-Injury Questionnaire. For each participant, the frequency of NSSI behavior in the past month was rated on a 5-point scale: 0 (never), 1 (at least once), 2 (every week) and 3 (every day). The frequency of NSSI was assessed separately for two time frames using the Ottawa Self-Injury Questionnaire: (a) the past 6 months and (b) the past 12 months. For each period, frequency was rated on a 5-point ordinal scale: 0 (never), 1 (1–5 times), 2 (approximately once a month), 3 (every week), and 4 (every day or almost every day). Higher frequency of NSSI was considered an indicator of greater NSSI severity [28] Zhang et al. [29]. demonstrated the high reliability and validity of the scale in the Chinese adolescent with mental disorder. Self-reported impulsivity was evaluated using the Chinese version of the Barratt Impulsivity Scale (BIS-11), which comprises of three subscales: non-planning impulsivity, motor impulsivity, and attentional impulsivity. Each subscale contains 10 items, rated on a 4-point Likert scale ranging from 1 (rarely or never) to 4 (almost always) [30]. The BIS-11 has good reliability and validity among Chinese population [31]. The Ottawa Self-Injury Questionnaire and BIS-11 were finished by the participants.

Behavioral task and event-related potentials

1. Behavioral and EEG data recording

The go/no-go ERP task was programmed using E-prime 2.0 Basic Standard Edition. A total of 180 facial images (45 negative, 90 neutral, and 45 positive) were selected from the Chinese Facial Affective Picture System (CFAPS) [32]. The experiment consisted of one brief practice session followed by three formal blocks, with breaks in between each blocks. Each formal block contained 120 trials, comprising of 90 go trials and 30 no-go trials. Emotional (positive and negative) and neutral facial stimuli were presented in random order, with probabilities of 25% for negative, 25% for positive, and 50% for neutral stimuli. Each trial began with a 50 ms black screen, followed by the presentation of a facial image for 1000 ms. Participants were instructed to press the space bar for each image unless the current image was identical to the immediately preceding one, in which case they were required to withhold their response. The instruction provided to participants was as follows: "You will view various face pictures. Press the space bar each time you see a picture. If the current picture matches the previous one, no key press is needed." After the stimulus, a black screen was presented for approximately 1500 ms before the next trial. The entire task lasted approximately 15 min. This paradigm has been widely used in previous studies to assess impulsivity [33] (Fig 1).

Fig. 1.

Go/no-go experimental flow chart

ERP data was recorded using a 64-channel EEG system (Braincap), with electrodes placed according to the international 10–20 system and FCz as the reference electrode. Signals were acquired using Recorder software at a sampling rate of 5,000 Hz, with scalp impedance maintained below 5 kΩ. Behavioral and ERP data from all participants were collected by two trained researchers who were blinded to the participants’ identity information.

• (2)Behavioral and EEG data preprocessing

Behavioral indices, including go and no-go accuracy and reaction times for go trials were extracted from the behavioral data using E-Prime 2.0.1.0.8 for subsequent statistical analysis. EEG preprocessing was performed using the EEGLAB (2013b) implemented in Matlab2013b. The preprocessing pipeline included: lower the sample rate into 500 Hz; band-pass filtering (high-pass filter ≥ 1 Hz, low-pass filter ≤ 45 Hz) and notch filtering (48–52 Hz); epoch segmentation (from − 0.5 s to 1.0 s relative to stimulus onset); manual rejection of noisy epochs caused by head movement artifacts; spherical interpolation for bad channels [34]; removal of ocular and electrocardiographic artifacts using independent component analysis (ICA) [35], re-reference to the bilateral mastoids, and rejection of segments exceeding ± 100 μV. All preprocessing steps were conducted by one trained researcher.

Based on previous studies [36–38], the N2 component is typically maximal over prefrontal regions, while the P3 component is predominantly distributed over frontal and frontocentral regions in both adults and children. Accordingly, in the present study, the N2 component was analyzed at the Fz and Cz electrodes, and the P3 component at the Cz and Pz electrodes. Mean amplitudes were calculated with a time window defined as the mean latency ± one standard deviation. Only correct-response trials were included in the ERP averaging.

Statistical analysis

Statistical analyses were performed using SPSS 26.0 (IBM Corp., Armonk, NY, USA). Continuous variables are represented as mean (M) ± standard deviation (SD), and categorical variables as frequency (n). One-way analysis of variance (ANOVA) and Chi-square tests were used to compare demographic and clinical characteristics among groups, as appropriate. Post hoc comparisons were conducted using pairwise tests with Bonferroni corrections. To examine group differences in ERP components, repeated measures ANOVA with a group × condition × electrode design was applied, with group as the between-subjects factor and electrode and condition (go vs no-go) as within-subjects factors. Greenhouse–Geisser corrections were applied when the assumption of sphericity was violated. All statistical tests were two-tailed, with the significance level set at p < 0.05. Spearman's rank correlation analyses were performed to assess associations between the NSSI severity and impulsivity measures, with a significance threshold of α = 0.05.

Results

Comparison of general demographic and clinical data among the three groups

A total of 145 participants were enrolled and categorized into three groups: the NSSI group (n = 50; 12 males, 38 females), the non-NSSI group (n = 45; 16 males, 29 females), and the HCs group (n = 50; 25 males, 25 females). A significant difference in sex distribution was observed among the three groups (p = 0.026). Post hoc analyses indicated a significant difference between the HCs and NSSI groups (p_Bonf < 0.05), whereas no significant differences were found between the NSSI and non-NSSI groups (p_Bonf > 0.05), or between the HCs and non-NSSI groups (p_Bonf > 0.05) (Table 1). No significant differences were observed among the three groups in other demographic variables, including age, years of education, and only-child status (all p > 0.05). Furthermore, no significant differences were found between the non-NSSI and NSSI groups in clinical characteristics, including the number of depression episodes, type of depression (first episode vs. recurrent), family history of mental disorders, HAMD-17 score, YMRS score, and medication history (all p > 0.05) (Table 1).

Table 1.

Comparison of demographic and clinical data

Variables	NSSI (N = 50)	non-NSSI (N = 45)	HC (N = 50)	χ2/F/t	p	Contrast
	n/M ± SD	n/M ± SD	n/ M ± SD
Sex(male/female)	12/38	16/29	25/25	7.315	0.026a	NSSI > HC*
Only child (yes/no)	19/31	14/31	13/37	1.673	0.433
Age (years)	14.75 ± 1.41	14.77 ± 1.21	15.19 ± 1.54	1.555	0.215
Education time (year)	8.56 ± 1.36	8.56 ± 1.20	9.16 ± 1.77	2.754	0.067
Depression episodes	1.06 ± 0.31	1.04 ± 0.21	–	-0.281	0.779
Depression type (recurrent/first episode)	2/48	2/43	–	0.012	0.914
Family history (yes/no/missing value)	6/43/1	3/39/3	–	0.661	0.416
HAMD-17	21.84 ± 3.73	21.64 ± 3.89	–	− 0.259	0.796
YMRS	0.98 ± 1.44	1.25 ± 2.07	–	0.738	0.468
Medication or not (yes/no)	15/35	6/39	–	3.821	0.051

M: mean; SD: Standard deviation

^aNSSI group versus HC group

**: < 0.001; *: < 0.05

Comparison of self-reported impulsivity and behavioral impulsivity among the three groups

Adolescents in the NSSI group exhibited higher total BIS scores compared to those in the non-NSSI and HC groups(p = 0.001). Post hoc analysis revealed significant differences between the NSSI and HCs groups, as well as between the NSSI and non-NSSI groups (both p_Bonf < 0.05), whereas the difference between the HCs and non-NSSI groups was not significant (p_Bonf > 0.05). For the motor impulsivity subscale, adolescents in the NSSI group scored significantly higher than those in the non-NSSI group, who in turn scored higher than the HCs group. Pairwise comparisons indicated significant differences among all three groups: NSSI and HCs (p_Bonf < 0.001), non-NSSI and HCs (p_Bonf < 0.05), and NSSI and non-NSSI (p_Bonf < 0.05). No significant differences were found among the three groups on either the attentional impulsivity or non-planning impulsivity subscales (both p > 0.05) (Table 2). With respect to behavioral task performance, there was a significant overall group difference in the accuracy of go trials among the three groups (p = 0.009). Post hoc analyses with Bonferroni correction showed that the NSSI group had significantly lower go-trial accuracy than the HCs group (p_Bonf < 0.05), whereas no significant differences were observed between the NSSI and non-NSSI groups or between the non-NSSI and HC groups (both p_Bonf > 0.05). No significant differences were found among the three groups on response time of go trails or accuracy of the no-go trials (both p > 0.05) (Table 2).

Table 2.

Comparison of self-reported and behaviorally assessed impulsivity among the three groups

Variables	NSSI (N = 50)	non-NSSI (N = 45)	HC (N = 50)	F	p	Contrast
	M ± SD	M ± SD	M ± SD
Total score of BIS	68.28 ± 7.41	64.10 ± 7.90	62.32 ± 8.11	7.480	0.001a,c	NSSI > HC*; NSSI > non-NSSI*
Motor impulsivity subscale	25.20 ± 3.56	23.10 ± 3.66	20.43 ± 3.08	23.392	< 0.001a,b,c	NSSI > non-NSSI*; NSSI > HC**; Non-NSSI > HC*
Attentional impulsivity subscale	20.72 ± 3.52	19.60 ± 3.82	20.38 ± 3.79	1.089	0.339
Non-planning impulsivity subscale	22.36 ± 3.29	21.40 ± 3.32	21.51 ± 3.43	1.167	0.314
Go ACC	0.88 ± 0.10	0.91 ± 0.09	0.93 ± 0.05	4.914	0.009a	NSSI < HC*
No-go ACC	0.77 ± 0.12	0.76 ± 0.15	0.80 ± 0.10	1.123	0.329
Go RT	524.18 ± 65.62	511.32 ± 64.55	537.93 ± 51.63	1.959	0.145

M: mean; SD: Standard deviation; Go ACC: accuracy of go trials; No-go ACC: accuracy of no-go trials; Go RT: response time of go task

^aNSSI group versus HC group

^bNon-NSSI group versus HC group

^cNSSI group versus non-NSSI group (Bonferroni corrected p < 0.05)

**: < 0.001; *: < 0.05

Comparison of latency and amplitude of go/no-go trails among three groups

Repeated measurement ANOVA was conducted to compare the latencies and amplitudes of the N2 and P3 components across groups, task conditions, and electrodes. Detailed ERP latencies and amplitudes at different electrodes are presented in Table 3 and Supplementary Table 1.

Table 3.

The latency and amplitude of N2 and P3 among three groups in different electrodes

	N2				P3
	Latency		Amplitude		Latency		Amplitude
	F	P	F	P	F	P	F	P
Group(NSSI, non-NSSI, HC)	2.337	0.1	0.171	0.835	3.353	0.038	0.833	0.437
Conditions (go, nogo)	4.022	0.047	148.247	< 0.001	3.871	0.051	140.913	< 0.001
Electrodes(Fz, Cz/Cz, Pz)	2.607	0.109	197.4	< 0.001	21.588	< 0.001	194.435	< 0.001
Group(NSSI,non-NSSI, HC)*conditions	0.144	0.866	4.335	0.015	0.849	0.43	2.901	0.058
Group(NSSI,non-NSSI, HC)*Electrodes	0.999	0.371	2.036	0.134	1.467	0.234	0.781	0.460
Group*conditions*Electrodes	1.619	0.357	2.091	0.127	0.115	0.891	1.596	0.206

For N2 latency, the main effect of group was not significant (F = 2.337, η² = 0.032, p = 0.100). However, a significant main effect of condition was observed at electrode Cz, with longer N2 latencies during no-go trials than during go trials (F = 4.623, p = 0.033). A significant main effect of group was observed in P3 latency (F = 3.353, η² = 0.046, p = 0.038). Further analyses showed that at electrode Cz, P3 latency was significantly longer during no-go trials (536.14 ± 32.17 ms) than during go trials (527.72 ± 28.99 ms) (F = 8.251, p = 0.005). However, post hoc comparisons revealed no significant differences in P3 latency among the three groups (p > 0.05).

Regarding N2 amplitude, the main effect of group was not significant (F = 0.171 η² = 0.003, p = 0.835). In contrast, a significant main effect of condition was observed, with N2 amplitudes at electrodes Fz and Cz being significantly smaller during no-go trials than during go trials (p < 0.001). Similarly, the main effect of group on P3 amplitude was not significant (F = 0.833, η² = 0.012, p = 0.437). However, the P3 amplitude at electrodes Cz and Pz was significantly larger during no-go trials than during go trials (p < 0.001). No significant group differences in P3 amplitude were observed across the three groups (all p > 0.05). Although the NSSI group exhibited smaller P3 amplitudes at Cz and Pz during no-go trials compared to the non-NSSI and HC groups, these differences did not reach statistical significance (all p > 0.05) (Table 3 and Fig 2).   

Fig. 2.

The ERP of go/no-go task Fz, Cz, Pz electrode among the three groups

Correlation analysis between NSSI severity and impulsivity

The severity of NSSI over the past 1 month, 6 months and 12 months was positively correlated with motor impulsivity subscale and the total BIS score (r = 0.198–0.417, p < 0.001). In contrast, the severity of NSSI across the same time intervals was negatively correlated with the accuracy of the go task(r=-0.248 to -0.193, p < 0.05) (Table 4).

Table 4.

Correlation analysis between NSSI severity and impulsivity

NSSI	Past 1 month	Past 6 months	Past 12 months
Total score of BIS	0.283**	0.228**	0.198*
MIS	0.417**	0.372**	0.345**
Go ACC	− 0.225**	− 0.248**	− 0.193*

MIS: motor impulsivity subscale; Go ACC: accuracy of go task

**: < 0.001; *: < 0.05

Discussion

Depressed adolescent patients with NSSI reported higher impulsivity score

Our results align with prior research [39–42], demonstrating that adolescent MDD patients with NSSI exhibited higher scores on the BIS total scale and motor impulsivity subscale compared to both MDD patients without NSSI and healthy controls. Moreover, higher impulsivity scores—particularly motor impulsivity—were positively associated with the severity of past NSSI, in line with earlier report [43]. Notably, Cassels et al. [43] showed that impulsivity independently predicted the onset of new NSSI episodes, exceeding the predictive value of other established risk factors. As outlined in the introduction section [12], NSSI often occurs spontaneously and with minimal forethought. Accordingly, individuals with elevated motor impulsivity may be more likely to engage in NSSI rapidly, especially under intense emotional states or other triggering conditions. Previous research [43, 44] have further indicated that highly impulsive individuals frequently exhibit emotional dysregulation, attention deficits, and impairments in perseverance and planning. These characteristics—particularly emotional dysregulation—have been established as key risk factors for NSSI [45, 46]. In addition, impulsivity was found to play a moderating role in the relationship between emotion dysregulation and NSSI behaviors [47], indicating that individuals with higher levels of impulsivity are more likely to engage in NSSI.

Depressed adolescent patients with NSSI exhibit higher behavioral impulsivity

Adolescent MDD patients with NSSI exhibited the lowest accuracy rate on go trials (the highest rate of omission errors). This was followed by adolescent MDD patients without NSSI, and then HCs, with a significant difference observed specifically between adolescent MDD patients with NSSI and HCs. Moreover, the accuracy rate on go trials (omission errors) were negatively correlated with the severity of prior NSSI, which is in line with previous studies [48], suggesting that greater lapses in sustained attention were associated with more severe self-injurious behavior. In addition, adolescent MDD patients with NSSI showed a lower accuracy rate on no-go trials, reflecting increased commission errors; however, this difference among the three groups did not reach statistical significance. Behavioral performance indices provide insight into distinct cognitive processes: accuracy on go trials primarily reflects sustained attention, no-go accuracy indexes executive control and response inhibition, and reaction time reflects information-processing speed. In this context, the absence of significant group differences in no-go accuracy is consistent with a previous study [49], which reported no significant differences in behavioral impulsivity between adolescents with and without NSSI. Nevertheless, our findings differ from those of previous studies [23, 50], which reported higher rates of both omission and commission errors in individuals with NSSI compared with healthy controls. The present results suggest that adolescent MDD patients, irrespective of NSSI status, exhibit impaired sustained attention relative to healthy controls, with the most pronounced impairment observed in those with NSSI. This pattern may be partly explained by the fact that both MDD groups in the present study were diagnosed with moderate to severe depression, a clinical condition known to be associated with substantial cognitive deficits and severe attention impairments. Furthermore, previous study [51] has demonstrated that individuals with NSSI show greater deficits in cognitive flexibility independent of core psychopathology, which may further exacerbate difficulties in attentional control in this population.

No significant difference was observed in neurophysiologically assessed impulsivity

Our findings further indicated that, although no statistically significant differences in N2 or P3 amplitude were observed among adolescent MDD patients with NSSI, adolescent MDD patients without NSSI, and healthy controls, distinct trend-level patterns emerged during no-go trials. Specifically, the P3 amplitude tended to be smaller in adolescent MDD patients with NSSI compared with the other two groups, while the N2 amplitude showed a tendency toward being larger in the MDD patients with NSSI group. However, these group differences did not reach statistical significance. Our findings are in line with a previous study [52] that examined neurocognitive functioning in adolescents with NSSI compared with healthy controls. That study assessed 240 participants across multiple cognitive domains, including processing speed, attention, learning, working memory, and executive function, and reported no significant differences between the NSSI and control groups [52]. This pattern partially deviates from some previous findings [50]. For example, Yu et al. [23] reported a significantly larger no-go N2 amplitude in college students with NSSI compared with those without NSSI, a result that aligns with our a priori hypothesis. In contrast, their study did not identify significant group differences in no-go P3 amplitude between college students with and without NSSI. Yoon et al. [50] reported a significantly smaller no-go P3 amplitude in individuals with NSSI compared with health controls. Although statistically significant differences in N2 and P3 amplitudes between adolescents with NSSI and healthy controls were not observed in the present study, a consistent trend toward reduced N2 and P3 amplitudes during no-go trials was evident in adolescent MDD patients with NSSI.

Accumulating evidence suggests that the emergence of NSSI is shaped not only by deficits in inhibitory control but also by dysfunction within reward-processing systems. Prior research has shown that impaired inhibitory control under punitive conditions (i.e., following negative feedback) is associated with more frequent NSSI behaviors in adults [53]. Similarly, compared with adolescents without NSSI, those with repetitive NSSI demonstrate reduced inhibitory control in response to cues signaling punitive content, whereas inhibitory responses to reward-related cues appear relatively preserved [54]. These findings underscore the complexity of NSSI in adolescence, highlighting the involvement of multiple interacting neural systems, including cognitive control networks associated with the prefrontal cortex and socio-emotional or motivational systems linked to the striatum [55]. From a developmental perspective, the prefrontal cortex undergoes prolonged maturation and remains functionally immature during adolescence, while reactivity to salient and emotionally relevant stimuli peaks during this period [56, 57]. This developmental imbalance between regulatory control and reward-related sensitivity may increase vulnerability to risk-taking and maladaptive coping behaviors, including NSSI. In the present study, we focused exclusively on inhibitory control and did not include task conditions that explicitly assessed reward processing. This methodological constraint may partially explain why no statistically significant group differences were observed in the N2 and P3 components in the NSSI group. Given accumulating evidence that NSSI behaviors are influenced by the interaction between inhibitory control and reward-related processes, future studies incorporating paradigms that simultaneously manipulate punishment and reward contingencies may provide a more comprehensive understanding of the neural mechanisms underlying NSSI in adolescents.

Limitations

While our study examined impulsivity among adolescent MDD with NSSI from a multi-dimensional perspective, several limitations should be acknowledged. First, although ERPs provide excellent temporal resolution, their spatial resolution is limited. The integration of functional magnetic resonance imaging (fMRI) in future studies could help localize the neural generators of key ERP components, such as the N2 and P3. Second, the relatively small sample size underscores the need for larger, well-matched cohorts to improve the robustness and generalizability of our findings. Third, the cross-sectional design precludes causal inferences regarding the relationship between neural activity and self-harm behaviors. Longitudinal studies are therefore warranted to clarify how developmental changes in brain function contribute to the onset and maintenance of NSSI. Finally, the retrospective assessment of NSSI may be subject to recall bias. To address this limitation, a prospective longitudinal follow-up study is planned to track dynamic changes in impulsivity and their associations with the progression or remission of NSSI across adolescence.

Conclusions

Elevated motor impulsivity, as assessed by both self-report measures and behavioral tasks, was observed in adolescent MDD patients with NSSI, and these alterations were associated with NSSI severity. However, the present study did not identify robust neurocognitive deficits specific to NSSI at the neurophysiological level. From a developmental perspective, an imbalance between immature prefrontal control systems and heightened socio-emotional sensitivity during adolescence may increase vulnerability to NSSI. Future studies with larger samples, longitudinal designs, and reward-based paradigms are needed to clarify underlying mechanisms and inform early intervention strategies.

Author contributions

Yuqiong He collected, collated, and analyzed data and were key contributing authors of the manuscript. Tingyu Yang, Kun Zhang, and Shuxian Wu collected data and assisted with the revision of the manuscript. Chunxiang Huang, Xueping Gao and Xuerong Luo assisted in patient recruitment and drafting of the paper. Yaru Zhang and Xilong Cui managed and supervised the study, revised the manuscript, and performed clinical assessment of the severity of symptoms. All authors have contributed to and approved the final manuscript.

Funding

This work was supported by the National Key R&D Program of China (No. 2025YFC2511300), the Hunan Provincial Natural Science Foundation (2025JJ60569), the Natural Science Foundation of Changsha City (KQ2502193), the STI 2030 - Major Projects (2022ZD0209102), and the Scientific Research Launch Project for new employees of the Second Xiangya Hospital of Central South University.

Data availability

The data that support the findings of this study are available from the corresponding author upon reasonable request.

Declarations

Ethics approval and consent to participate

The study was approved by the Ethics Committee of the Second Xiangya Hospital of Central South University (Ethics Approval Number: 2021 K001), and all participants and their guardians provided written informed consent before enrollment in the study.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.