Cognitive–Affective Correlates of Adolescent Non-Suicidal Self-Injury: Executive Functioning, Social–Emotional and Interpersonal Cognition, and Emotional Processing

Round 1

Reviewer 1 Report

The manuscript addresses an important and timely topic and offers a coherent conceptual synthesis of cognitive-affective correlates of adolescent self-harm/NSSI. Its main strength lies in the organisation of a heterogeneous literature into three analytically distinguishable domains: executive functioning, emotional processing, and social cognition. The manuscript is generally well written, balanced, and appropriately cautious in several places, particularly when distinguishing between direct empirical evidence, clinically informed interpretation, and broader theoretical synthesis.

However, several issues should be addressed before publication.

1. The manuscript is explicitly presented as a focused narrative review rather than a systematic review or meta-analysis, which is acceptable. Nevertheless, the literature identification process should be described with greater precision. The authors mention the databases searched, broad search terms, prioritisation of studies from 2021 onwards, and the absence of formal risk-of-bias assessment. However, the manuscript would benefit from clearer information on the date of the searches, inclusion and exclusion criteria, handling of overlapping samples, and the number and type of studies ultimately included. A supplementary table summarising the included studies would considerably strengthen transparency without requiring the manuscript to become a full systematic review.
2. The authors correctly acknowledge that the literature does not always distinguish cleanly between NSSI, broader self-harm, and suicidality-related phenomena. This is one of the strengths of the manuscript. However, this distinction should be made more visible throughout the review. In several sections, evidence from NSSI-specific studies and broader self-harm/suicidality studies is discussed together. I recommend adding a short conceptual subsection or a summary table indicating whether each major body of evidence refers specifically to NSSI, broader self-harm, suicidal ideation/attempts, or mixed outcomes.
3. The manuscript repeatedly and appropriately notes that social cognition is the least directly assessed domain and that much of the evidence concerns adjacent constructs such as interpersonal sensitivity, family-relational interpretation, perceived invalidation, and emotional communication difficulties. This is a valuable distinction. However, because “social cognition” appears in the title and is treated as one of the three central domains, the authors should either strengthen this section with more direct evidence from theory of mind, social inference, emotion recognition, or social attribution tasks, or rename/reframe the domain more cautiously, for example as “social-emotional and interpersonal cognition”. This would avoid overstating the current empirical basis for a discrete social-cognitive impairment model.
4. The manuscript is conceptually coherent, but several points are repeated across the abstract, introduction, domain sections, discussion, table, and conclusion. In particular, the claims that emotional processing has the strongest evidence, executive functioning shows a selective rather than global pattern, and social cognition remains indirect but clinically relevant are reiterated many times. These are important claims, but the manuscript would be stronger if some repetition were reduced and the discussion focused more explicitly on what this review adds beyond existing narrative and systematic reviews on adolescent NSSI.
5. Table 1 is useful and supports the integrative aims of the manuscript. However, it is text-heavy and contains long phrases within narrow columns, which reduces readability. The table would be clearer if shortened, or if the clinical implications were separated from the evidence summary. The authors might also consider adding a figure representing the proposed developmental cognitive-affective model.

• The title is clear and informative. However, if the authors retain the current emphasis on NSSI rather than self-harm broadly, the title could be adjusted to reflect this more precisely.
• The abstract is well structured and accurately reflects the manuscript’s main argument. It may be useful to state explicitly in the abstract that this is a narrative review, not a systematic review.
• The introduction provides a strong rationale for the review and clearly defines the three domains. The distinction between emotional processing, executive functioning, and social cognition is helpful, although some overlap between these domains remains inevitable.
• The “Review Approach and Literature Identification” section is welcome, but it should be expanded slightly. At minimum, the authors should provide search dates, clearer inclusion/exclusion criteria, and a more explicit explanation of how studies were selected and weighted.
• The executive functioning section is balanced and avoids overclaiming. The distinction between performance-based measures, self-reported impulsivity, subjective cognitive dysfunction, and broader self-regulatory problems is particularly important and should be retained.
• The emotional processing section is the strongest part of the manuscript. The distinction between emotion regulation, cognitive emotion regulation, and alexithymia/emotional awareness is conceptually useful.
• The social cognition section is interesting but currently relies heavily on indirect evidence. The authors acknowledge this limitation, but the section would benefit from clearer differentiation between direct social-cognitive measures and broader interpersonal or relational constructs.
• The clinical implications are relevant and well formulated. However, some clinical interpretations go beyond the direct empirical evidence. The authors usually acknowledge this, but further qualification would strengthen the manuscript.
• The limitations section is unusually thorough and is one of the strengths of the paper. It appropriately addresses outcome heterogeneity, construct heterogeneity, measurement variability, cross-sectional evidence, clinical sample bias, and cultural/geographical concentration of the literature.
• The conclusion is supported by the preceding sections. It accurately reflects the central argument that adolescent self-harm is associated with heterogeneous cognitive-affective correlates rather than a single cognitive deficit.
• The English is generally clear and polished, but the manuscript would benefit from some stylistic tightening. Several sentences are long and abstract, and some paragraphs repeat similar formulations. A light language edit focused on concision would improve readability.

Author Response

Comment 1: The manuscript is explicitly presented as a focused narrative review rather than a systematic review or meta-analysis, which is acceptable. Nevertheless, the literature identification process should be described with greater precision. The authors mention the databases searched, broad search terms, prioritisation of studies from 2021 onwards, and the absence of formal risk-of-bias assessment. However, the manuscript would benefit from clearer information on the date of the searches, inclusion and exclusion criteria, handling of overlapping samples, and the number and type of studies ultimately included. A supplementary table summarising the included studies would considerably strengthen transparency without requiring the manuscript to become a full systematic review.

Response: Thank you for this helpful and constructive comment. We agree that the literature identification process required greater transparency, even though the manuscript is intentionally framed as a focused narrative review rather than as a systematic review or meta-analysis. We have therefore revised the former Section 1.1, now renumbered as Section 1.2, to provide more precise information on the search process, including the databases searched, the search period and final update date, the main search terms, the narrative inclusion and exclusion criteria, and the approach taken when potentially overlapping samples or closely related datasets were encountered.

We have also clarified that study selection was narrative and purposive rather than protocol-driven, and that greater interpretive weight was given to review-level evidence, longitudinal studies, meta-analytic findings, adolescent-focused studies, and findings replicated across more than one design or sample type. At the same time, we have preserved the distinction between a transparent narrative synthesis and a formal systematic review by explicitly noting that no PRISMA-style flow diagram, formal screening log, formal risk-of-bias assessment, or quantitative synthesis was undertaken.

In addition, we have added Supplementary Table S1 to improve transparency regarding the core publications used to organise the cognitive-affective framework. The table is not presented as a PRISMA-style list of all records identified or screened, but as a focused summary of the main sources that structured the synthesis across emotional processing, executive functioning, and emerging social-emotional/interpersonal cognition.

Comment 2: The authors correctly acknowledge that the literature does not always distinguish cleanly between NSSI, broader self-harm, and suicidality-related phenomena. This is one of the strengths of the manuscript. However, this distinction should be made more visible throughout the review. In several sections, evidence from NSSI-specific studies and broader self-harm/suicidality studies is discussed together. I recommend adding a short conceptual subsection or a summary table indicating whether each major body of evidence refers specifically to NSSI, broader self-harm, suicidal ideation/attempts, or mixed outcomes.

Response: Thank you for this important and constructive comment. We agree that the distinction between NSSI-specific evidence, broader self-harm evidence, suicidality-related evidence, and mixed or overlapping outcomes should be made more visible throughout the manuscript.

We have therefore added a new conceptual subsection in the Introduction, “Outcome Definitions and Evidential Scope”. This subsection clarifies how NSSI, broader self-harm, mixed self-injurious behaviours, suicidal ideation, and suicide attempts are treated in the review. We now state explicitly that NSSI-specific evidence is given the greatest interpretive weight when conclusions concern adolescent self-injury without suicidal intent, whereas broader self-harm and suicidality-related studies are used more cautiously and primarily to illuminate relevant cognitive-affective processes rather than to imply full equivalence of outcomes.

We have also added an “outcome scope” column to the newly added Supplementary Table S1. This table summarises the core publications used to organise the cognitive-affective framework and indicates whether each source primarily concerns NSSI-specific outcomes, broader self-harm or suicidality-related outcomes, mixed/overlapping outcomes, or broader youth evidence used cautiously. This allows readers to see more clearly which parts of the synthesis are grounded most directly in NSSI-specific evidence and which rely on related but non-equivalent outcome domains.

Comment 3: The manuscript repeatedly and appropriately notes that social cognition is the least directly assessed domain and that much of the evidence concerns adjacent constructs such as interpersonal sensitivity, family-relational interpretation, perceived invalidation, and emotional communication difficulties. This is a valuable distinction. However, because “social cognition” appears in the title and is treated as one of the three central domains, the authors should either strengthen this section with more direct evidence from theory of mind, social inference, emotion recognition, or social attribution tasks, or rename/reframe the domain more cautiously, for example as “social-emotional and interpersonal cognition”. This would avoid overstating the current empirical basis for a discrete social-cognitive impairment model.

Response: Thank you for this important and helpful suggestion. We agree that, given the limited direct evidence from theory-of-mind, social-inference, emotion-recognition, or social-attribution tasks in adolescent self-harm, retaining the unqualified term “social cognition” could overstate the empirical basis for a discrete social-cognitive impairment model.

Rather than adding a small number of indirect or only partly relevant studies in order to strengthen the section artificially, we have chosen to reframe the domain more cautiously as “social-emotional and interpersonal cognition”. This change has been made in the title, abstract, keywords, relevant parts of the Introduction, the Section 3 heading and text, the Discussion, Figure 1/caption, Limitations, and Conclusions. The revised terminology better reflects the evidence actually reviewed, which concerns interpersonal sensitivity, family-relational interpretation, perceived invalidation, emotional communication, and the communicability of distress more than direct performance-based social-cognitive impairment. We have also revised Section 3 to state more explicitly that direct evidence on core social-cognitive processes remains limited and that the domain should be read as exploratory and indirectly supported. The section now distinguishes more clearly between narrow experimental social-cognitive constructs, such as theory of mind, social inference, social attribution, and emotion-recognition accuracy, and adjacent social-emotional/interpersonal constructs used in the current literature. Future research directions have also been adjusted to call for more direct assessment using performance-based theory-of-mind, social-inference, attributional, and emotion-recognition paradigms.

Comment 4: The manuscript is conceptually coherent, but several points are repeated across the abstract, introduction, domain sections, discussion, table, and conclusion. In particular, the claims that emotional processing has the strongest evidence, executive functioning shows a selective rather than global pattern, and social cognition remains indirect but clinically relevant are reiterated many times. These are important claims, but the manuscript would be stronger if some repetition were reduced and the discussion focused more explicitly on what this review adds beyond existing narrative and systematic reviews on adolescent NSSI.

Response: Thank you for this helpful comment. We agree that the previous version repeated several central claims too frequently, particularly the relative evidential status of emotional processing, executive functioning, and social-emotional/interpersonal cognition. We have therefore revised the manuscript to reduce repetition and to make the contribution of the review more explicit.

Specifically, we have shortened the framing paragraph in the Introduction so that it no longer repeats the full evidential hierarchy later developed in the Discussion. We have also revised the Integrative Synthesis section to focus more directly on the added value of the review. The revised Discussion now presents the review as a formulation-oriented synthesis that organises cognitive-affective findings according to their conceptual function within a developmental framework: emotional processing as the identification, elaboration, regulation, and symbolisation of distress; executive functioning as action selection and inhibition under affective pressure; and social-emotional/interpersonal cognition as the interpretation and communicability of distress within relational contexts. We have also streamlined the Conclusions to avoid restating the same hierarchy of evidence in the same terms. The revised conclusion now focuses more directly on the integrative contribution of the framework and its implications for conceptual precision, clinical formulation, and future mechanism-oriented research.

Comment 5: Table 1 is useful and supports the integrative aims of the manuscript. However, it is text-heavy and contains long phrases within narrow columns, which reduces readability. The table would be clearer if shortened, or if the clinical implications were separated from the evidence summary. The authors might also consider adding a figure representing the proposed developmental cognitive-affective model.

Response: Thank you for this helpful suggestion. We agree that the original Table 1, although intended to support the integrative aims of the manuscript, was text-heavy and partly repetitive with the surrounding text. To improve readability and strengthen the conceptual presentation of the review, we have removed the original Table 1 from the main manuscript. In its place, we have added Figure 1, which presents a conceptual model of the proposed developmental cognitive-affective framework. The figure illustrates how emotional processing, executive functioning, and emerging social-emotional/interpersonal cognition may interact under conditions of adolescent distress, while also showing that these pathways are developmentally and culturally embedded. The figure is intended to convey the integrative structure of the argument more clearly than the previous table. We have also kept the evidence summary and methodological transparency separate from the main conceptual figure. The core publications used to organise the framework are now summarised in Supplementary Table S1, which includes the outcome scope of each source. This allows the main manuscript to remain more readable while preserving transparency about the evidence base.

Detailed comment 1: The title is clear and informative. However, if the authors retain the current emphasis on NSSI rather than self-harm broadly, the title could be adjusted to reflect this more precisely.

Response: Thank you for this helpful suggestion. We agree that, although the manuscript discusses broader self-harm and suicidality-related evidence where relevant, its primary focus is adolescent non-suicidal self-injury. To reflect this more precisely, we have revised the title to foreground NSSI explicitly. We have also changed “Cognitive Correlates” to “Cognitive-Affective Correlates”, as this better captures the scope of the review, which includes executive functioning, emotional processing, and social-emotional/interpersonal cognition.

Detailed comment 2: The abstract is well structured and accurately reflects the manuscript’s main argument. It may be useful to state explicitly in the abstract that this is a narrative review, not a systematic review.

Response: Thank you for this helpful suggestion. We agree that the review type should be made explicit in the Abstract. We have therefore revised the Abstract to state that the article is a focused, non-systematic narrative review. We have also clarified that the review is centred on adolescent non-suicidal self-injury, while drawing on broader self-harm and suicidality-related evidence only where relevant to the cognitive-affective formulation.

Detailed comment 3: The introduction provides a strong rationale for the review and clearly defines the three domains. The distinction between emotional processing, executive functioning, and social cognition is helpful, although some overlap between these domains remains inevitable.

Response: Thank you for this positive and helpful comment. We agree that some overlap between the domains is inevitable, particularly because emotional regulation, behavioural control, and interpersonal meaning-making often interact in real-world episodes of adolescent distress. We have therefore revised the Introduction to make clearer that the three domains are intended as analytically distinguishable but clinically interrelated areas, rather than as independent or mutually exclusive mechanisms.

Detailed comment 4: The “Review Approach and Literature Identification” section is welcome, but it should be expanded slightly. At minimum, the authors should provide search dates, clearer inclusion/exclusion criteria, and a more explicit explanation of how studies were selected and weighted.

Response: Thank you for this helpful comment. We agree that the review approach required greater methodological transparency, even though the manuscript is intentionally framed as a focused, non-systematic narrative review rather than as a systematic review or meta-analysis. We have expanded the “Review Approach and Literature Identification” section to provide the requested information. The revised section now specifies the databases searched, the search period and final update date, the main search terms, and the use of backward citation tracking. We have also clarified the narrative inclusion and exclusion criteria, including the prioritisation of adolescent or youth self-harm/NSSI studies addressing cognitive-affective or interpersonal constructs relevant to the review. In addition, we have made more explicit how studies were selected and weighted. The revised section now states that study selection was narrative and purposive rather than protocol-driven, and that greater interpretive weight was given to review-level evidence, longitudinal studies, meta-analytic findings, adolescent-focused studies, and findings replicated across more than one design, sample type, or methodological approach. We have also clarified that studies drawing on overlapping populations were not treated as independent replications unless they contributed distinct information.

Detailed comment 5: The executive functioning section is balanced and avoids overclaiming. The distinction between performance-based measures, self-reported impulsivity, subjective cognitive dysfunction, and broader self-regulatory problems is particularly important and should be retained.

Response: Thank you for this positive and helpful comment. We have retained the distinction between performance-based executive measures, self-reported impulsivity, subjective cognitive dysfunction, and broader self-regulatory difficulties in the Executive Functioning section. We agree that this distinction is important for avoiding overclaiming and for clarifying why the evidence supports a selective rather than global executive-impairment account. Minor wording revisions made elsewhere in the manuscript were undertaken without weakening this distinction.

Detailed comment 6: The emotional processing section is the strongest part of the manuscript. The distinction between emotion regulation, cognitive emotion regulation, and alexithymia/emotional awareness is conceptually useful.

Response: Thank you for this positive comment. We are pleased that the distinction between emotion regulation, cognitive emotion regulation, and alexithymia/emotional awareness was found useful. We have retained this structure in the Emotional Processing section, as it is central to the manuscript’s conceptual organisation and helps clarify why emotional processing has the strongest evidential support among the domains reviewed.

Detailed comment 7: The social cognition section is interesting but currently relies heavily on indirect evidence. The authors acknowledge this limitation, but the section would benefit from clearer differentiation between direct social-cognitive measures and broader interpersonal or relational constructs.

Response: Thank you for this important and constructive comment. We agree that the previous version needed to distinguish more clearly between direct social-cognitive measures and broader interpersonal or relational constructs.

In response, we have reframed the domain more cautiously as “social-emotional and interpersonal cognition” throughout the manuscript. Section 3 has been revised to state explicitly that direct evidence from core social-cognitive measures, such as theory of mind, social inference, social attribution, and emotion-recognition accuracy, remains limited in adolescent self-harm/NSSI. We now distinguish these narrower performance-based social-cognitive constructs from adjacent interpersonal and relational constructs, including interpersonal sensitivity, family-relational interpretation, perceived invalidation, emotional communication, cognitive insight, family functioning, and the communicability of distress. We have also revised the section to clarify that the evidence reviewed should be read primarily as indirect and clinically suggestive rather than as demonstrating a discrete social-cognitive impairment model. This distinction is also reflected in the title, Abstract, keywords, Discussion, Limitations, Conclusions, and Figure 1/caption.

Detailed comment 8: The clinical implications are relevant and well formulated. However, some clinical interpretations go beyond the direct empirical evidence. The authors usually acknowledge this, but further qualification would strengthen the manuscript.

Response: Thank you for this helpful comment. We agree that some of the clinical implications necessarily involve formulation-level interpretation rather than direct empirical demonstration, particularly given the heterogeneity of the literature and the frequent use of cross-sectional designs or adjacent constructs. We have therefore added further qualification at the beginning of the Clinical Implications section. The revised text now states explicitly that the clinical implications should be read as formulation-level implications rather than as direct treatment recommendations derived from intervention trials. We also clarify that the assessment suggestions are intended to indicate clinically relevant areas for enquiry, not to imply established causal mechanisms or domain-specific treatment targets.

Detailed comment 9: The limitations section is unusually thorough and is one of the strengths of the paper. It appropriately addresses outcome heterogeneity, construct heterogeneity, measurement variability, cross-sectional evidence, clinical sample bias, and cultural/geographical concentration of the literature.

Response: Thank you for this positive comment. We are pleased that the Limitations section was considered thorough and appropriately balanced. We have retained this section and ensured that the revisions made elsewhere in the manuscript remain consistent with the limitations already acknowledged, including outcome heterogeneity, construct heterogeneity, measurement variability, cross-sectional evidence, clinical sample bias, and cultural/geographical concentration of the literature.

Detailed comment 10: The conclusion is supported by the preceding sections. It accurately reflects the central argument that adolescent self-harm is associated with heterogeneous cognitive-affective correlates rather than a single cognitive deficit.

Response: Thank you for this positive comment. We are pleased that the Conclusion was considered well supported by the preceding sections and consistent with the central argument of the manuscript. We have retained this core conclusion, while making minor wording revisions to reduce repetition and to emphasise more clearly the integrative contribution of the proposed developmental cognitive-affective framework.

Detailed comment 11: The English is generally clear and polished, but the manuscript would benefit from some stylistic tightening. Several sentences are long and abstract, and some paragraphs repeat similar formulations. A light language edit focused on concision would improve readability.

Response: Thank you for this helpful comment. We have undertaken a light language edit focused on concision, readability, and stylistic consistency. In particular, we have shortened several long or abstract sentences, reduced repeated formulations across the Introduction, Discussion, and Conclusions, and revised passages where similar claims were restated in closely overlapping terms. We have also streamlined the framing of the manuscript’s central argument so that key distinctions are retained without unnecessary repetition. These edits were intended to improve fluency and readability while preserving the conceptual precision and cautious tone of the review.

Reviewer 2 Report

The manuscript is a well-written and clinically well-argued narrative review that offers a genuine contribution by organizing a fragmented body of literature into three analytically distinguishable domains. The measured tone, the clear distinction between empirical findings and clinical inferences, as well as the honest acknowledgment of the heterogeneity of empirical support across the different domains, are all commendable aspects.
Major point
Identification and selection of the literature.
Section 1.1 describes a literature search process without reporting how many records were identified, screened, and included. If a search procedure is described, a minimum level of transparency regarding its outcomes is expected. The authors should therefore either provide this information or explicitly define the review as non-systematic and justify study selection narratively, avoiding the appearance of a systematic protocol.

Geographic concentration.
The limitation concerning the predominance of Chinese adolescent samples is acknowledged only briefly. However, this issue is more substantial than a passing remark suggests. Cultural context influences the meaning of self-harm, family functioning, norms regarding emotional expression, and interpersonal constructs, thereby affecting the generalizability of the proposed cognitive-affective framework. This aspect should be addressed more directly in the Discussion section, not only in the Limitations section.

Social cognition section.
This domain is the one least directly supported by evidence, as the authors themselves acknowledge. However, the section does not always consistently maintain the distinction between core social cognition constructs (theory of mind, social inference) and adjacent interpersonal constructs (interpersonal sensitivity, family functioning). In some passages this boundary is carefully respected, whereas in others it becomes blurred. The section would benefit from being reframed more explicitly as an emerging or exploratory area, rather than as a domain with empirical standing equivalent to the other two.

 

The term clinically proximal is repeatedly used without ever being operationally defined. Table 1 is well organized but largely redundant with the text; replacing it with a conceptual figure illustrating interactions among the domains could be considered. The Clinical Implications section is the most readable part of the paper and would benefit from at least a brief connection to evidence-based interventions for adolescent NSSI (e.g., DBT-A, MBT-A).

Author Response

Major point 1: Identification and selection of the literature

Comment: Section 1.1 describes a literature search process without reporting how many records were identified, screened, and included. If a search procedure is described, a minimum level of transparency regarding its outcomes is expected. The authors should therefore either provide this information or explicitly define the review as non-systematic and justify study selection narratively, avoiding the appearance of a systematic protocol.

Response: Thank you for this helpful comment. We agree that the previous wording of Section 1.1 could have created the impression of a more formal systematic search procedure than was intended. The section has therefore been revised to define the article more explicitly as a focused, non-systematic narrative review and conceptual synthesis rather than as a systematic review or meta-analysis. We have clarified that no PRISMA-style flow diagram, formal screening log, risk-of-bias assessment, or quantitative synthesis was undertaken. We have also expanded the description of the narrative and purposive approach to literature selection, including the rationale for prioritising recent studies, adolescent samples, NSSI-specific evidence, and review-level, longitudinal, meta-analytic, or replicated findings where available.

Major point 2: Geographic concentration

Comment: The limitation concerning the predominance of Chinese adolescent samples is acknowledged only briefly. However, this issue is more substantial than a passing remark suggests. Cultural context influences the meaning of self-harm, family functioning, norms regarding emotional expression, and interpersonal constructs, thereby affecting the generalizability of the proposed cognitive-affective framework. This aspect should be addressed more directly in the Discussion section, not only in the Limitations section.

Response: Thank you for this important observation. We agree that the geographic and cultural concentration of much of the recent evidence has implications not only for generalisability, but also for the interpretation of the proposed cognitive-affective framework itself. The Discussion has therefore been revised to address this issue more directly. We now explicitly note that several constructs central to the review, including family functioning, interpersonal sensitivity, emotional communication, perceived invalidation, and help-seeking expectations, are culturally embedded and may not carry identical meanings across settings. We have also clarified that the proposed framework should be read as a provisional organising model requiring further cross-cultural testing rather than as a culturally neutral account of adolescent self-harm.

Major point 3: Social cognition section

Comment: This domain is the one least directly supported by evidence, as the authors themselves acknowledge. However, the section does not always consistently maintain the distinction between core social cognition constructs (theory of mind, social inference) and adjacent interpersonal constructs (interpersonal sensitivity, family functioning). In some passages this boundary is carefully respected, whereas in others it becomes blurred. The section would benefit from being reframed more explicitly as an emerging or exploratory area, rather than as a domain with empirical standing equivalent to the other two.

Response: Thank you for this important and constructive comment. We agree that the previous version did not always maintain a sufficiently clear boundary between core social-cognitive constructs and adjacent interpersonal or relational constructs. We also agree that this domain should not be presented as having empirical standing equivalent to emotional processing or executive functioning. In response, we have reframed the domain more cautiously throughout the manuscript as “social-emotional and interpersonal cognition”. This terminology better reflects the evidence actually reviewed, which concerns interpersonal sensitivity, family-relational interpretation, perceived invalidation, emotional communication, cognitive insight, family functioning, and the communicability of distress more than direct performance-based measures of theory of mind, social inference, social attribution, or emotion-recognition accuracy. We have also revised Section 3 to distinguish more clearly between core social-cognitive processes, such as theory of mind, social inference, social attribution, and emotion-recognition accuracy, and adjacent social-emotional/interpersonal constructs. Passages that could have implied that adjacent interpersonal constructs provide direct evidence of a discrete social-cognitive impairment model have been revised. To ensure consistency, this reframing has not been limited to Section 3. The title, Abstract, keywords, Introduction, Discussion, Clinical Implications, Figure 1/caption, Limitations, and Conclusions have also been updated. The manuscript now states more explicitly that emotional processing has the strongest evidential basis, executive functioning is selectively implicated, and social-emotional/interpersonal cognition remains a clinically relevant but exploratory and indirectly supported extension of the framework.

Detailed comments: The term clinically proximal is repeatedly used without ever being operationally defined. Table 1 is well organized but largely redundant with the text; replacing it with a conceptual figure illustrating interactions among the domains could be considered. The Clinical Implications section is the most readable part of the paper and would benefit from at least a brief connection to evidence-based interventions for adolescent NSSI (e.g., DBT-A, MBT-A).

Response: Thank you for these helpful and constructive suggestions. We have revised the manuscript in response to all three points.

First, we agree that the phrase “clinically proximal” was insufficiently defined and could introduce unnecessary ambiguity. Rather than adding a new operational definition for a term that was not essential to the argument, we have removed the phrase throughout the manuscript and replaced it with more precise wording, such as “closely linked to self-harm in clinical formulation”, “closely related to NSSI in clinically relevant samples”, and “clinically relevant and closely associated with NSSI”. This revision preserves the intended meaning while avoiding an undefined technical expression.

Second, we agree that the previous Table 1, although useful as a summary, partly repeated material already presented in the text. We have therefore replaced it with Figure 1, a conceptual figure illustrating the proposed interactions among emotional processing, executive functioning, and emerging social-emotional/interpersonal cognition. The figure is intended to show the relative evidential status of the three areas and their hypothesised interactions under conditions of adolescent distress. It also makes explicit that the framework is developmentally and culturally embedded, and that the arrows represent hypothesised interactions rather than established causal pathways. Third, we have added a brief paragraph to the Clinical Implications section connecting the formulation proposed in the review to evidence-based interventions for adolescent self-harm, particularly Dialectical Behavior Therapy for Adolescents (DBT-A) and Mentalization-Based Treatment for Adolescents (MBT-A). This addition clarifies that DBT-A is especially relevant to emotion dysregulation, distress tolerance, impulsive action, and crisis-related behavioural control, whereas MBT-A is relevant to mentalization, attachment-related affect, interpersonal meaning, and the communicability of distress. We have also added relevant references to randomized trials and review-level evidence.

Round 2

Reviewer 2 Report

The authors have addressed the major concerns raised in the previous review in a substantial and satisfactory manner.

In particular, the manuscript now clearly defines itself as a focused non-systematic narrative review rather than implying a systematic-review methodology, and the rationale for literature selection has been described with appropriate transparency.
The distinction between NSSI-specific evidence and broader self-harm/suicidality outcomes has also been clarified effectively.The revision of the social cognition section is especially improved.
The Discussion section now addresses the cultural and geographic concentration of the literature in a more theoretically meaningful way, rather than limiting this issue to a brief methodological caveat. The addition of the conceptual figure and the integration of DBT-A and MBT-A into the Clinical Implications section further improve the manuscript’s clarity and clinical relevance.