Enhanced reactive inhibition in adolescents with non‐suicidal self‐injury disorder

To investigate whether the core of the pathophysiology underlying non‐suicidal self‐injury (NSSI) relates to poor impulse control due to impaired motor inhibition (i.e. the ability to inhibit a preplanned motor response).

predisposition for quick reactions to internal or external stimuli without proper evaluation of their consequences, 9 it is not a unitary construct.At the behavioural level, at least two components of impulsivity can be distinguished: 9 (1) decisional impulsivity (characterized by making hasty decisions without fully evaluating the potential costs and benefits or weighing the available alternatives); and (2) motor impulsivity (characterized by the poor ability to avoid premature responses or stopping already initiated actions).Frequently, studies on NSSI have used unidimensional assessments of impulsivity, lacking a well-defined conceptual framework for understanding this construct.Understanding the multidimensional aspect of impulsivity is crucial when it comes to comprehending the complex nature of NSSI behaviour.
Second, impulsivity has been assessed using two sets of tools: (1) self-report questionnaires; and (2) task-based measures. 5However, these instruments capture different aspects of impulsivity.While self-report questionnaires capture trait impulsivity, task-based measures indicate state impulsivity. 6otably, trait and state impulsivity measures usually have a poor correlation. 10In fact, Hamza et al. 5 found that when considering studies based on self-report questionnaires, NSSI engagers seem to have higher impulsivity than other people.However, experimental measures of impulsivity assessing inhibitory control proficiency were not associated with NSSI behaviour.A possible explanation is that self-report questionnaires are very subjective; thus, their accuracy strongly depends on the respondents' insight into their behavioural tendencies.
Furthermore, self-report measures indicate the general risk of self-injurious behaviours and not the short-term risk.Therefore, task-based measures are more reliable in assessing the risk for NSSI engagement. 5According to this reasoning, Liu et al. 6 evaluated the association between motor impulsivity with NSSI and suicidal intent or behaviours in a meta-analysis focusing only on state-sensitive indices of impulsivity (measured using the go/no-go task and the stop-signal task [SST]).This time, they found that NSSI was associated with task-based measures of motor impulsivity. 6cHugh et al. 7 confirmed the association between such measures, NSSI, and suicidal behaviour.However, single studies included in both meta-analyses failed to find significant differences between task-based measures of motor impulsivity in typical versus pathological individuals, indicating that the effect was modest.Furthermore, such an effect was obtained by considering studies involving people engaging in both NSSI and suicidal behaviours, that is, two different clinical syndromes.To date, there is no clear evidence of whether NSSI behaviour is associated with deficits in motor impulsivity.
Inhibitory control is not a unitary executive function but encompasses at least a cognitive and a motor domain. 11he ability to inhibit a preplanned motor response, that is, motor inhibitory control, is the more reasonable domain to be studied in people performing out-of-context movements, such as cleaning a surface even when it is already clean, or involuntarily jerking their head to the side while sitting in a library, for example.This domain could be further divided into two different subdomains: (1) reactive inhibition, that is, the ability to cancel an ongoing response at the presentation of a stop-signal; and (2) proactive inhibition, that is, the ability to modulate inhibitory control in advance according to the current context and one's goals.Mirabella 11 reviewed the studies on five neurodevelopmental disorders, in which both reactive and proactive inhibition were assessed using the SST.He found that the clinical phenotypes showed striking differences regarding the impairment of motor inhibition.Thus, the inability to control urges cannot be simply ascribed to a general impairment of this executive function.
Given these premises, we aimed to reassess the hypothesis that a deficit in motor inhibition can feature NSSI behaviours.We conducted a systematic search in PubMed and Scopus, according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines, 12 to identify articles using SST to study NSSI (Figure S1).We found seven publications (Table 1).Three of them [13][14][15] did not provide the measure of reactive inhibition, i.e. the stop-signal reaction time (SSRT), which represents the time it takes to suppress an ongoing action. 168][19][20] However, after splitting the NSSI group into low-severity and high-severity subgroups, Fikke et al. 18 found that the latter had a shorter SSRT, that is, a more efficient reactive inhibitory control, than the former.Critically, no previous studies measured proactive inhibition and none included people diagnosed with NSSID.
To overcome the limitations of previous research, we compared reactive and proactive inhibition in a relatively large cohort of adolescents consisting of drug-naive adolescents with NSSID with those of age-matched and sexmatched typically developing adolescents by administering the reaching arm version of the SST. 21Following the results of the two most recent meta-analyses, 6,7 we hypothesized

What this paper adds
• Adolescents with non-suicidal self-injury disorder (NSSID) showed enhanced reactive inhibition versus controls.• Proactive inhibition was similar in adolescents with NSSID and controls.• Impulsivity assessed using the Barratt Impulsiveness Scale Version 11 did not differ in adolescents with NSSID versus controls.• Adolescents with NSSID showed higher impulsivity than controls in the lack of perseverance subscale of the UPPS-P Impulsive Behavior Scale.that deficits in inhibitory control would be the core feature of NSSID.

Participants
We recruited 30 adolescents diagnosed with NSSID from the inpatient psychiatric emergency Adolescent Unit of the Human Neurosciences Department of Sapienza University, Rome.Adolescents were screened using the Deliberate Self-Harm Inventory and the Repetitive-Not Suicidal SelfInjury Questionnaire.They were included if they were right-handed, 22 drug-naive, had not received cognitive behavioural treatments designed to address NSSI, had a normal cognitive profile (fullscale IQ ≥ 70, measured using the Wechsler Intelligence Scale for Children), and had a normal or corrected-to-normal vision.
Adolescents were excluded if they presented organic diseases, schizophrenia, or autism spectrum disorder. 23Impulsivity was evaluated using (1) the UPPS-P Impulsive Behavior Scale and (2) the Barratt Impulsiveness Scale Version 11.The Multi-Attitude Suicide Tendency Scale and Difficulties in Emotion Regulation Scale were used to assess suicidal tendencies and emotion regulation disorder, respectively.Lastly, comorbidities were evaluated using the Kiddie Schedule for Affective Disorders and Schizophrenia for School-Aged Children-Lifetime Version.The psychometric properties of all scales are described in Appendix S1.We excluded two adolescents because they could not complete the test.
To collect normative values of inhibitory control from a secondary school in Brescia, we recruited and tested 28 righthanded, typically developing adolescents (three males and 25 females; mean age [SD] 15 years 8 months [1 year 5 months], range: 11 years 4 months-17 years 8 months) without a history of neurological or psychiatric disorders and with normal or corrected-to-normal vision.The age of adolescents with NSSID and typically developing adolescents was not significantly different (t-test, t [54] = 0.06; p = 0.95).The demographic information and scores of clinical scales for the 28 adolescents with NSSID (three males and 25 females), along with the corresponding average values for typically developing participants, are shown in Table 2.
All participants were unaware of the study's purpose and completed the experiment in one session after the clinical assessment.We computed the sample size using a power analysis. 24We hypothesized that adolescents with NSSID could have an SSRT 20 ms longer than typically developing controls.We used a univariate approach to repeated measures with Greenhouse-Geisser correction and applied previous data to assess the SD and mean SSRT values. 21With a target power of 0.85 and a variability scale factor of 2, the computed total sample size was 56.
The Ethical Committee of Sapienza University, Rome (protocol no.278/16 RIF.CE: 4007) approved all procedures and the study was conducted according to the Declaration of Helsinki.We obtained written informed consent from the participants' parents.

Behavioural tasks
We gave adolescents with NSSID and controls two tasks (Figure 1) in a counterbalanced fashion: (1) the go-only task; and (2) the reaching version of the SST. 21In the go-only task, participants were instructed to execute a reaching movement towards a peripheral target.This task provided a measure of reaction times (i.e., the time between the appearance of the go-signal and the moment in which the index finger was moved away from the screen) and movement times (i.e., the time between the index finger moving away and the moment in which the participant touched the peripheral target) in a context in which participants knew that they would never be required to stop an ongoing action.The SST consisted of two types of trials, no-stop (67% of the total trials) and stop (33% of the total trials) trials, intermixed pseudo-randomly.No-stop trials were identical to the go-only trials.In the stop trials, a stop-signal (given by the presentation of the central stimulus) was shown at a variable delay (stop-signal delay) after the presentation of the go-signal.To succeed, participants had to keep the index finger on the stop-signal for 300 to 400 ms, cancelling the pending movement towards the peripheral target.The stop-signal delay was initially set at 119.7 ms (9 refresh rates) but was changed adaptively using a staircase procedure. 21This procedure aimed to keep the success rate of the stop-signal trials to around 50%.Participants had to complete two blocks of 200 trials of the SST and one block of 100 trials of the go-only task.If requested, short breaks were interposed between the execution of blocks or tasks.
During the SST, a common strategy unconsciously adopted by participants to increase the probability of successfully stopping a response consisted of slowing movement initiation during the no-stop trials. 16To discourage such a waiting strategy, we verbally urged participants to respond as quickly as possible to the go-signal and informed them that the probability of successfully stopping would approximate 50% in any case.For the same reason, we also set a reaction time limit for the no-stop trials.Whenever the reaction time length exceeded 500 ms, no-stop trials were considered errors.Nevertheless, to avoid cutting off the right tail of the reaction time distribution, 25 we gave participants 100 ms extra before aborting the trial.Thus, if participants removed the finger between 500 and 600 ms, reaction times were recorded but the trial was signalled as a failure.Over time, these trials were named reaching trials and were included in the analyses.Reaching trials accounted for 13.8% of the no-stop trials in typically developing adolescents and 13.2% in adolescents with NSSID.

Behavioural measures and analysis
We used reaction times, movement times, and SSRT as the behavioural parameters.SSRT was estimated based on the theoretical framework of the horse race model 26 (Figure 2a) using the integration method (Figure 2b) because this provides the most reliable estimate. 16o-only and no-stop trials with 3SD from the mean reaction time were discarded (1.36% for adolescents with NSSID and 0.83% for typically developing peers).We assessed proactive inhibition by comparing the reaction and movement times of the no-stop and go-only trials. 27This approach differs from previous methods because it leverages the reciprocal relationship between these two behavioural measures.Our findings demonstrate that during a no-stop trial, reaction time is lengthened; correspondingly, movement time is shortened compared to the go-only trials.This phenomenon, named the 'context effect' , 27 represents an unconscious optimization of motor strategy in two different contexts.In the SST, awareness of a stop-signal leads to increased reaction times in the no-stop trials, facilitating better encoding of movement parameters.Conversely, the shorter reaction times in the go-only trials do not allow for this optimization, necessitating completion of the movement plan during execution of the motor response, thereby increasing movement times.
The normality assumption of the distributions of all parameters was assessed using the Shapiro-Wilk test.This assumption was always verified except in three cases: the movement times of the no-stop trials for participants with NSSID (Shapiro-Wilk test W = 0.909, p = 0.019, skewness = 1.3; kurtosis = 5.29) and typically developing peers ) for each ANOVA and (2) Cohen's d for each t-test.We computed the likelihood of the null hypothesis by computing the Bayes factors (BF 10 ), setting the prior odds to 0.707, which is compatible with differences in the range of (hypothesized) medium-to-large effect sizes, where less emphasis is placed on the alternative hypothesis than on the null hypothesis, as suggested by Rouder et al. 32 BF 10 values smaller than 0.33 provide moderate support for a null hypothesis, while BF 10 values smaller than 0.10 provide robust support.By contrast, BF 10 values greater than 3 provide moderate support for the alternative hypothesis, while BF 10 values greater than 10 provide strong support. 33All statistical analyses were performed using the R Stats package. 34

R E SU LTS
First, we checked that our estimates of the SSRT were accurate.The horse race model assumes that the go-processes and stopprocesses are stochastically independent; that is, the occurrence of the former is not affected by the latter, and vice versa (Figure 2a).We tested such an assumption by comparing the reaction times of no-stop and stop-failure trials (instances in which adolescents moved although they should not) because, in both cases, movements were performed, but stop-failure trials were initiated because the go-process finished before the stopprocess (Figure 2b).Thus, it follows that the responses escaping inhibition should be those corresponding to movements with reaction times shorter than the sum of the stop-signal delay and SSRT.Therefore, the mean reaction times of the stop-failure trials should be shorter than those of the no-stop trials. 16The twoway, mixed-design ANOVA (between-individual factor: group [adolescents with NSSID, typically developing peers]; withinindividual factor: trial type [no-stop trials, stop-failure trials]) confirmed the assumption by showing a significant main effect of trial type (Table 3), as the reaction times of the stop-failure trials were shorter than the no-stop trials in both groups.No other significant effect was found.
We also assessed whether the staircase algorithm worked as expected.Both adolescents with NSSID and typically developing adolescents had a success rate of stop-signal trials around 50% (Table 4), with no significant difference (p [failure]; two-tailed t-test t [54] = 1.68, p = 0.099).Thus, we concluded that our estimates of reactive inhibition were accurate.

Proactive inhibition in adolescents with NSSID and controls is similar
We evaluated proactive inhibition, exploiting the context effect phenomenon, in three different ways. 27,35As the outcome was consistent regardless of the method used, we illustrate one approach here while including the other two in Appendix S1.We assessed the context effect at the population level by comparing the means of the reaction and movement times with two separate two-way, mixed-design ANOVAs.The ANOVA on the reaction times revealed a main effect of trial type because participants had shorter reaction times during the goonly trials (mean = 220.2ms, 95% CI = 214.1-226.3)than the no-stop trials (mean = 382.1 ms, 95% CI = 372.3-391.9;Table 5 and Figure 3b).The ANOVA on movement times did not show any significant main effects or interaction (Table 6 and Figure 3c).Consequently, we did not detect any difference in proactive inhibitory control between adolescents with NSSID and typically developing peers.

DISCUS SION
We assessed reactive and proactive inhibition in a relatively large cohort of drug-naive adolescents with NSSID; unexpectedly, they exhibited a selective improvement of reactive inhibitory control with regard to typically developing peers.This result indicates that a core feature of NSSI behaviour involves heightened action control rather than decreased inhibitory control. 36We hypothesized that enhanced reactive inhibition would enable adolescents with NSSID to suppress instinctive actions that would otherwise protect the physical well-being of their bodies.To the best of our knowledge, this is the first study showing that a F I G U R E 1 Depiction of the sequence of events of the go-only and stop-signal tasks.Visual stimuli consisted of red circles (2.434 cd/m 2 luminance, 4-degree diameter of the visual angle [dva]).The background was black and with uniform luminance (<0.01 cd/m 2 ).Go-only task: adolescents had to reach and hold the central red circle for 500 to 800 ms.Then, after the central stimulus vanished, a peripheral circle was presented at 18.6dva (13.5 cm) on the horizontal plane on the right side (go-signal).Adolescents were instructed to reach and hold it for 300 to 400 ms (in 50-ms increments).Stop-signal task: no-stop trials were the same as those of the go-only task.In the stop trials (33% of the total trials), after presentation of the go-signal, the stop-signal (given by the reappearance of the central circle) was shown at variable delays (stop-signal delays).Adolescents were instructed to withhold the pending movement without moving the finger away from the touchscreen (MicroTouch, sampling rate 200 Hz; 3 M Touch Systems) and to hold the central stimulus for 300 to 400 ms.The stop trial was aborted whenever the finger was moved away (stop-failure trial).Corrected responses were signaled by an auditory feedback (represented by a musical note).The dotted circles surrounding the circles represent the tolerance area for the touches (5-dva diameter) and were not visible to the individual.CORTEX, a non-commercial software package developed at the National Institute of Health (Bethesda, USA) to control stimulus presentation and behavioral response recording (https:// www.nimh.nih.gov/ resea rch/ resea rch-condu cted-at-nimh/ resea rch-areas/ clini cs-and-labs/ ln/ shn/ softw are-projects).Abbreviations: SSD, stop-signal delay; SSRT, stop-signal reaction time.psychiatric disorder is associated with better functioning of one executive function rather than its impairment.

Reactive inhibition in NSSID
Our evidence contrasts with previous research using the SST on NSSI, showing no differences in reactive inhibitory control between typical controls and NSSI engagers.Several methodological features can explain the differences.First, all previous studies recruited people from community samples, where the pathological group consisted of individuals performing at least one NSSI act in their lifetime, 15,17,19,20 or in one study, at least two NSSI acts in the current year. 18By contrast, we enrolled inpatients with NSSID, and our behavioural assessment was temporally very proximal to the last episode of self-injury.In summary, our patients had a much more severe form of NSSI than in all other research.Second, all previous studies but one 18 enrolled young adults and not adolescents with NSSI (Table 1).As NSSI onset typically occurs between 12 and 14 years, 1 studying young adults is unlikely to be the best choice.The only study in which adolescents with NSSI and typically developing peers were tested found a result compatible with ours, showing that adolescents with high-severity NSSI had a shorter SSRT than adolescents with low-severity NSSI. 18However, none Results of the two-way, mixed-design analysis of variance (ANOVA) of reaction types in the no-stop and stop-failure trials.Trial type: no-stop trial, stop-failure trial; group: adolescents with non-suicidal self-injury disorder, typically developing peers.Statistically significant results are reported in bold.Bayes factors (BF 10 ) report the ratio of the likelihood of the alternative hypothesis to the likelihood of the null hypothesis; partial η p 2 was used to measure the size effects for the ANOVA; differences in the estimated marginal means (M diff ) are reported along with their 95% confidence intervals (CIs).

F I G U R E 2
Computations of the reactive inhibition measure, that is, the stop-signal reaction time (SSRT).(a) Schematic representation of the horse race model.Performance in the stop-signal task is modelled as a race towards a threshold (broken horizontal line) between two stochastically independent processes, that is, the go-process (green line), triggered by the go-signal, and the stop-process (red line), elicited by the stop-signal.If the stop-process finishes before the go-process, the movement is cancelled (top panel  Mean (SD) behavioural values of the stop-signal and go-only tasks for adolescents with non-suicidal self-injury disorder (NSSID) and typically developing adolescents.Accuracy was computed as the ratio between correct go-only trials and the total number of go-only trials (as the sum of correct trials plus trials in which participants missed the target, remained still on the central stimulus for more than 2 seconds, or did not hold the central stimulus/target for the requested amount of time).had significant differences from typical controls.Given that the difference in the SSRT between adolescents with NSSID and controls is characterized by a large effect size and a high Bayes factor value, we conclude that our result is very robust.

Proactive inhibition in NSSID
Proactive inhibition consists of a preparatory process modulating the effectiveness of reactive inhibition in the future according to environmental constraints. 37The absence of differences between adolescents with NSSID and controls indicates that the disorder has a selective effect on inhibitory control.Interestingly, tasks impinging different aspects of proactive control, such as those dealing with delay and probabilistic discounting, never show an impact of NSSI. 38his evidence suggests that proactive control in individuals with NSSID is comparable to that of typical individuals.

NSSID and impulsivity and inhibitory control
We failed to find significant differences between adolescents with NSSID and typical peers on the Barratt Impulsiveness Scale Version 11.Even though this might appear odd, we must underline that we tested adolescents and not young adults, as most other studies did.On self-reports, adolescents tend to be more impulsive than young adults.In fact, our mean (SD) value of the total Barratt Impulsiveness Scale Version 11 score of 65.7 (7.1) is close to that reported by Fossati et al., 39 that is, 66.5 (9.9).By contrast, we found that adolescents with NSSID were more impulsive than controls in one subscale of the UPPS-P questionnaire.The data on typically developing adolescents should be taken cautiously because they were obtained just after the end of the COVID-19 pandemic in Italy, which had a negative impact on mental health. 40As expected, [5][6][7] self-report measures of impulsivity in adolescents with NSSID did not correlate with the SSRT, confirming that reactive inhibition is disentangled from such measures.Besides this, a recent review 11 showed that individuals might show motor impulsivity in the absence of motor inhibition deficits or have selective impairment of one subdomain of motor inhibition.Thus, a simple equation cannot be made between poor urge control and a deficit in motor inhibition.

NSSI behaviour is linked to reinforcement processes
A completely different view of NSSI behaviour is that it serves as an emotion regulation strategy or a way to communicate feelings to others. 8From an intrapersonal perspective, selfinjuries can either move attention away from aversive affective or cognitive states (negative reinforcement) or generate pleasurable stimulations (positive reinforcement).From an interpersonal perspective, NSSI behaviours can be used as social signals of great intensity when other communication strategies fail, either to avoid undesirable social situations (negative reinforcement) or to request help (positive reinforcement).In all these instances, individuals deliberately and purposefully engage in self-harm behaviours not because they cannot withhold the action at the sight of a sharp object but because self-injuries elicit a rewarding process by decreasing negative affective states.The coupling between the rewards associated with self-injuries and the heightening Two-way, mixed-design analysis of variance (ANOVA) of the reaction times in the no-stop and go-only trials: trial type (no-stop, go-only); group (adolescents with nonsuicidal self-injury disorder, typically developing peers).Statistically significant results are reported in bold.Bayes factors (BF 10 ) report the ratio of the likelihood of the alternative hypothesis to the likelihood of the null hypothesis.Partial η p 2 was used to measure the size effects of the ANOVA.Differences in the estimated marginal means (M diff ) are reported along with their 95% confidence intervals (CIs).
of reactive inhibition would allow individuals to override their self-preservation instinct.As rewards enhance reactive inhibition by decreasing SSRT, 41 we hypothesized that individuals with NSSID unconsciously improve their reactive inhibition through repeated engagement in deliberate self-harm acts.

Limitations
A key limitation of our study was the absence of an assessment of decisional impulsivity and reinforcement processes within our sample.Thus, certain conclusions are speculative and would benefit from validation in future studies.

Conclusions
Our findings revealed that adolescents with NSSID exhibited a selective improvement in reactive inhibition compared to controls, indicating that the core feature of the NSSI pathophysiology does not lie in an increase in motor impulsivity.Instead, we propose that NSSI behaviour is sustained by the association between reinforcement processes linked to selfinjuries and enhancement in reactive inhibition proficiency.This research has clinical implications.First, consistent with recent suggestions, [5][6][7] our study highlighted the divergent results obtained from laboratory tests compared to self-report questionnaires when assessing motor impulsivity in adolescents with NSSID.This underscores the need to combine laboratory tests and questionnaires for a more precise psychodiagnostic evaluation.Second, in the context of NSSI, current pharmacological interventions may include antipsychotics to address motor impulsivity. 42However, based on our findings, this approach does not yield fruitful results and should be reconsidered after obtaining consistent outcomes.

AC K NO W L E D GE M E N T S
We thank Dr. Roberto Chiappini and Dr. Davide Manfredi of the Canossa Campus for allowing us to collect data from typically developing adolescents.We also thank Dr. Michele Grassi for providing statistical advice.GM conceptualized the study.GM and CM wrote the initial draft of the manuscript.GM wrote all the subsequent versions.CM and GM provided the statistical analyses.AT, SP, and DG recruited the participants, and read and edited all drafts of the manuscript.This work was supported by the Antonio Meneghetti Award 2019 (GM) and by the Erasmus+ project (GM, grant no.610134-EPP-1-2019-1-JO-EPPKA2-CBHE-JP).

C ON F L IC T OF I N T E R E S T S TAT E M E N T
The authors declare that no competing interests exist.

DATA AVA I L A BI L I T Y S TAT E M E N T
Raw data and R scripts can be obtained via the Open Science Framework (OSF) at https:// osf.io/ dfsn4/ .

F I G U R E 3
) and vice versa (bottom panel).(b) Estimating the SSRT with the integration method.The SSRT is obtained by subtracting the starting time of the stop-process, the mean stop-signal delay, from its finishing time, which is calculated by 'integrating' the no-stop trial reaction time (RT) distribution from the onset of the go-signal until the nth reaction times, with n = the number of reaction times in the reaction time distribution of no-stop trials multiplied by the proportion of stop-failure trials, that is, p (failure).For instance, when there are 360 no-stop trials and p (failure) is 0.5, the nth reaction time is the 180th fastest no-stop reaction time.The horse race model relies on the following assumption for the way it has been conceived.Considering the hypothetical distribution of no-stop reaction times during stop trials, the responses that escape inhibition (stop-failure trials) should correspond to reaching movements with reaction times less than the sum between the stop-signal delay and the SSRT.Abbreviations: SSD, stop-signal delay.Violin plots of the average values of the stop-signal reaction times (SSRTs) of adolescents with non-suicidal self-injury disorder (NSSID) (red) and typically developing (TD) adolescents (sky blue).(a) Mean SSRT, (b) mean reaction times (RT), and (c) mean movement times (MT) for each group in the no-stop trials during the stop-signal task (no-stop; adolescents with NSSID shown in red and typically developing peers shown in sky blue) and the go-only task (go-only; adolescents with NSSID shown in pink and typically developing peers shown in turquoise).The violin plots depict regular box plots enclosed in a kernel probability density (the area's width represents the relative data frequency).The box boundary closest to zero indicates the first quartile.A black line within the box marks the median.The box boundary farthest from zero indicates the third quartile.The whiskers indicate values 1.5 times the interquartile range below the first quartile and above the third quartile.Outliers are shown as black dots.
Summary of the studies identified in the systematic literature research.
T A B L E 1 18 Seven studies were included.The stop-signal reaction time (SSRT) was not computed in two of them.The number of adolescents and age are reported for non-suicidal self-injury (NSSI) engagers and controls.Age values are reported with the corresponding SD (where possible) and are reported in decimal years as in the original studies.Differences in SSRT length between adolescents with NSSI disorder (NSSID) and typically developing adolescents and the corresponding p-values are reported (where possible).Negative differences indicate those instances where adolescents with NSSID were faster than typically developing adolescents.Abbreviation: SST, stop-signal task.
Clinical features and questionnaire scores of adolescents with NSSID and controls.Clinical scales and questionnaire scores of adolescents with non-suicidal self-injury disorder (NSSID).The Multi-Attitude Suicide Tendency Scale (MAST) purports to measure attitudes to suicide on four factors: attraction to life (AL); repulsion to life (RL); attraction to death (AD); and repulsion to death (RD).The Impulsive Behavior Scale (UPPS-P) assesses negative urgency (the tendency to act rashly under extreme negative emotions), lack of premeditation (tendency to act without thinking), lack of perseverance (inability to remain focused on a task), sensation seeking (tendency to seek out new and thrilling experiences), and positive urgency (tendency to act rashly under extreme positive emotions).The total score of the Difficulties in Emotion Regulation Scale (DERS) and subscales are non-acceptance of emotional responses, difficulty engaging in goal-directed behaviours, impulse control difficulties, lack of emotional awareness, limited access to emotion regulation strategies, and lack of emotional clarity.The total score of the Barratt Impulsiveness Scale Version 11 with first-order factor scores is shown in italics (attention, cognitive instability, motor, perseverance, self-control, cognitive complexity); second-order factor scores are shown in bold (attentional, motor, non-planning).Mean (SD) values for age and questionnaire scores are shown for adolescents with NSSID and typically developing controls.Significant p-values of the t-tests between groups are shown in bold.Abbreviations: F, female; M, male; R-NSSI-Q, Repetitive Non-Suicidal Self-Injury Questionnaire.
T A B L E 2
T A B L E 3 Average values of the stop-signal and go-only tasks for adolescents with NSSID and controls.
Two-way, mixed-design ANOVA on movement times for the no-stop and go-only trials.-way, mixed-design analysis of variance (ANOVA) of movement times in the no-stop and go-only trials: trial type (no-stop, go-only); group (adolescents with nonsuicidal self-injury disorder, typically developing peers).Bayes factors (BF 10 ) report the ratio of the likelihood of the alternative hypothesis to the likelihood of the null hypothesis.Partial η p 2 was used to measure the size effects for the ANOVA.Differences in the estimated marginal means (M diff ) are reported along with their 95% confidence intervals (CIs).Two-way, mixed-design ANOVA on the reaction times for no-stop and go-only trials.
T A B L E 6TwoT A B L E 5