Structural neuroimaging of skin-picking disorder

Background: Skin-picking disorder (SPD) is conceptualized as an obsessive-compulsive and related disorder (OCRD). Patients with SPD excessively manipulate their skin, which leads to skin lesions, psychological distress, and functional impairment. The neuroanatomical facets of this disorder are still poorly understood. Methods: A total of 220 participants (123 patients with a primary diagnosis of SPD and 97 healthy controls; mean age = 30 years, 80% female) were recruited for a voxel-based morphometry (VBM) study. VBM data were compared between patients and controls, and between three SPD subgroups, each characterized by a distinct age of symptom onset (before puberty, during puberty, adulthood). Results: Relative to the healthy comparison group, patients with SPD had significantly less grey matter volume (GMV) in regions of interest (ROIs: insula, orbitofrontal cortex, pallidum, cerebellum, supramarginal gyrus) and in the frontal pole and occipital regions (whole-brain findings). Early onset of symptoms (before puberty) was associated with elevated levels of focused skin-picking, in addition to less GMV in specific ROIs (insula, orbitofrontal cortex) as well as in paracingulate/ superior temporal regions (whole-brain findings). Conclusions: SPD-related reductions in GMV were identified in brain regions involved in interoception, emotion regulation, and motor control. This partially aligns with findings for OCD. The detection of different age-of-onset groups based on clinical as well as morphometric data points to the heterogeneity of the disorder and warrants further investigation.


INTRODUCTION
Skin-picking disorder (SPD) is conceptualized as a condition related to obsessivecompulsive disorders (OCRDs) according to the Diagnostic and Statistical Manual of Mental Disorders (DSM-5; American Psychiatric Association, 2013) and the International Classification of Diseases (ICD-11, World Health Organization, 2022).
Core symptoms include 'compulsive' picking, scratching, and/or squeezing of the skin, which causes skin lesions, clinical distress, and impairment in important areas of functioning.SPD is a common mental disorder that affects approximately 2-3 % of the general population (APA, 2013).
Structural neuroimaging is a useful method for detecting neuroanatomical alterations associated with a diagnosis of SPD.This approach can assist in understanding important neurobiological facets of this disorder, which in turn may assist in the development of more effective treatments (Grant et al., 2023).However, investigations that have aimed at identifying neuroanatomical correlates of pathological skin-picking are still scarce and have mainly included small sample sizes.Further, the findings obtained from those studies have been inconsistent concerning the brain regions involved.
For example, Roos et al. (2015) compared neuroimaging data from 17 patients with SPD, 15 control participants, and 17 patients with trichotillomania ('compulsive hair pulling').Skin picking was associated with increased grey matter volume (GMV) in the basal ganglia, reduced cortical thickness in frontal areas, and increased thickness of the cuneus.In a study by Harris et al. (2017), a positive correlation was observed between impulsiveness and cortical thickness in the left insula in 20 patients diagnosed with SPD.Further, skin-picking severity correlated negatively with cortical thickness in the supramarginal gyrus (SMG).In a study conducted by Schienle et al. (2018a), SPD patients (n = 35) showed reduced GMV in the J o u r n a l P r e -p r o o f Journal Pre-proof orbitofrontal cortex (OFC) and insula compared to healthy controls (n = 35).A study by Blum et al. (2018) identified a negative correlation between cortical thickness of the insula and impulsive motor behavior of SPD patients (n = 15).In a study by Wabnegger et al. (2019), SPD patients (n = 30) showed structural alterations in specific subregions of the cerebellum related to motor and affective-cognitive functions, relative to healthy controls (n = 31).Another study by the same research group (Schienle & Wabnegger, 2020) focused on two types of SPD and compared GMV between patients who described their skin-picking as pleasant (SPD+,n = 21) or unpleasant (SPD-, n = 30).Relative to the healthy control group (n = 25), insular volume was found to be reduced in both SPD groups.Further, the SPD-group was shown to have less insula volume than the SPD+ group.
The largest structural neuroimaging study conducted so far included 193 patients diagnosed with trichotillomania and/or SPD, as well as 58 healthy controls (Grant et al., 2023).In that study, no brain region was found to differ in GMV between the SPD group and the control group.Moreover, two subtypes of skin-picking (emotional/reward picking; n = 99; a group showing a strong urge to pick and little control; vs. functional picking; n = 42; a group with a lower urge to pick), were not found to be associated with specific brain-structural characteristics.Critically, it should be noted that this study included patients with a combined diagnosis of SPD and trichotillomania.This limits the possibility of detecting disorder-specific GMV correlates.
In sum, as has been outlined above, it is difficult to draw firm conclusions from the studies that have attempted to investigate the neuroanatomical correlates of pathological skin-picking, due to inconsistent results, small sample sizes, and other methodological issues.

Journal Pre-proof
For the current analysis, voxel-based morphometry was used to compare structural fMRI data between individuals with a primary diagnosis of SPD (n = 123; no comorbid trichotillomania) and healthy controls (n = 97).GMV in regions of interest (ROIs: insula, orbitofrontal cortex (OFC), basal ganglia, supramarginal gyrus (SMG), cerebellum) was compared between the two groups; the ROIs were identified based on findings from previous research (e.g., Harris et al., 2017;Roos et al., 2015;Schienle et al., 2018a;Wabnegger et al., 2019).
All participants completed questionnaires for the assessment of skin-picking symptoms.Furthermore, board-certified clinical psychologists conducted a structured clinical interview with the patients.Based on the collected data, a subgrouping was performed relating to the presence/absence of comorbid mental disorders (no comorbidity: 56% vs. comorbidity: 44%) and reported age of onset of SPD symptoms (before puberty, during puberty, and adulthood).The clinical and neuroanatomical data of these groups were compared with each other.A comparable age-related distinction has been established for patients with OCD symptoms, which either can have an early onset (~11 years) or a later onset (~19 years) according to large epidemiological studies (e.g., Ruscio et al., 2010).Finally, it should be noted that the subgroups 'reward skin-picking' vs. 'functional skin-picking', as previously specified (Grant et al., 2023) were unable to be compared to one another since the majority of patients (85%) in the present study was found to carry out 'emotional/reward picking'.

Sample
A total of 123 patients (23 male/ 100 female; all Caucasian) had been diagnosed with primary SPD by a board-certified clinical psychologist according to DSM-5 criteria (APA, 2013).The control group consisted of 97 participants (20 male/ 77 female; all J o u r n a l P r e -p r o o f Journal Pre-proof Caucasian).The majority of the participants (87%) had at least a high school diploma (Table 1).Of the patients, 7% received psychopharmacological treatment (selective serotonin reuptake inhibitors).
Patients with a diagnosis of a dermatological disease (e.g., scabies, psoriasis), and/or a comorbid diagnosis of a mental disorder associated with non-suicidal selfinjury (e.g., borderline personality disorder) were not included in the sample.
Moreover, patients with previous/current substance abuse, psychotic symptoms, and those with severe symptoms of depression were excluded.
The clinical psychologist diagnosed current comorbid mental disorders in 44% (n = 54) of the patients (mainly from the anxiety disorder spectrum: generalized anxiety disorder, panic disorder, specific phobia, 33%).Additional diagnoses included depression with mild to moderate symptoms in 8% of the patients, eating disorders (7%), somatoform disorders (2%), adjustment disorders (5%) and obsessivecompulsive disorders (2%).None of the patients was diagnosed with comorbid trichotillomania.Diagnoses were determined through a structured clinical interview, utilizing the short version of the diagnostic interview for mental disorders (Margraf & Cwik, 2017).Following this interview, specific questions related to SPD were posed to each patient.Patients were asked to describe their skin-picking symptoms in an open-answer format, along with detailing the age at which symptoms first manifested.Furthermore, each patient underwent targeted questioning regarding the presence of additional body-focused repetitive behaviors, such as hair-pulling.Based on the reports, three onset groups were created: early (n = 42; before puberty; < 13 years of age), mid (n = 46; onset during puberty; 13-17 years), and late-onset (n = 34; onset ≥ 18 years).One participant did not provide information concerning the onset of symptoms.

J o u r n a l P r e -p r o o f
Journal Pre-proof Control participants reported no diagnoses of mental disorders and use of psychotropic medication.The study complied with all relevant ethical guidelines and regulations involving human participants (Code of Ethics of the World Medical Association; Declaration of Helsinki) and was approved by the ethics committee of the University (GZ 39/29/26 ex 2018/19).

Questionnaires a)
Skin Picking Scale, revised (SPS-R, German version, Gallinat et al., 2016) The SPS-R gathers information about the severity of skin-picking symptoms.The scale contains eight items covering the following domains: (1) frequency of urge to pick, (2) intensity of urge to pick, (3) time spent picking, (4) control over picking, ( 5) interference due to skin picking, (6) emotional distress, (7) avoidance behavior due to picking, (8) skin damage due to skin-picking.Each item is rated on a 5-point scale from 0 (none) to 4 (extreme).The item scores can be summed to a total score (range: 0 -32; McDonald's omega = .96)as well as two subscale scores (symptom severity and impairment).Higher scores indicate greater symptom severity.The two SPS_R subscales were substantially correlated in the present sample (r = .85).
Therefore, the analysis was restricted to the total SPS_R.Total SPS_R scores ≥ 7 index pathological skin-picking (clinical cut-off score).

b)
Milwaukee Inventory for the Dimensions of Adult Skin-picking (MIDAS, Walther et al., 2009).
The MIDAS is a self-report measure for the assessment of focused and automatic skin-picking.The two subscales comprise six items each (focused; omega = .Journal Pre-proof rated on a five-point scale ranging from 1 (= "not true for any of my skin-picking") to 5 (= "true for all of my skin-picking").Higher scores indicate greater symptom severity.
This questionnaire was only completed by the patients (since it has been developed with a clinical sample of patients diagnosed with SPD).

MRI recording and VBM analysis
Structural Gaussian kernel with a full width at half maximum (FWHM) of 6 mm.Finally, the smoothed images were harmonized to reduce heterogeneity between both datasets using 'neuroHarmonize' (Pomponio et al., 2020).
For the VBM analyses, comparisons of gray matter volume (GMV) between the control group and patient group, patients with and without diagnosed comorbid mental disorders as well as the three symptom-onset groups (early, mid, and late) were conducted using analyses of covariance (ANCOVAS; covariates: total intracranial volume (TIV), age, and participants' sex).
The selection of ROIs (insula, OFC, basal ganglia, SMG, cerebellum) was based on previous findings from structural neuroimaging research (e.g., Harris et al., 2017;Roos et al., 2015;Schienle et al., 2018a;Wabnegger et al., 2019).The clusterbuilding threshold was set at 0.005 uncorrected.Results were considered statistically significant when family-wise error-corrected p-values for voxel-peaks were below 0.05.All ROI results were small-volume-corrected.Additionally, for the ANCOVA analyses, results were masked with the F-contrast for the main effect (threshold: 0.05).

Statistical analyses of self-report data
To compare questionnaire scores and demographic data between the groups, Welch's t-test for independent samples, and Welch's analyses of variance were used (JAMOVI: 2.3.28).Results were considered statistically significant when p-values were below .05.To minimize the risk of false positives when comparing the three age-of-onset groups, we applied the Bonferroni correction (i.e.alpha/number of comparisons; new critical p-value = .017).

J o u r n a l P r e -p r o o f
Journal Pre-proof 3.1 Self-report data

Comparison of patients and controls
Patients with SPD obtained higher scores on the SPS_R than healthy controls (Table 1).Patients' mean scores on both subscales of the MIDAS were higher (p ≤ .001)than those reported by Walther et al. (2009)   Note: SPS_R: Skin-picking scale revised; MIDAS: Milwaukee Inventory for the Dimensions of Adult Skin-Picking; M = mean, SD = standard deviation; CI = confidence intervals

Comparison of patients with and without comorbid mental disorders
Patients with current comorbidities reported more intense focused skin-picking and had a higher total score on the Skin-Picking Scale_revised (Table 3).

Comparison of patients and controls
The patients displayed less GMV in the selected ROIs compared to the control group (Table 4).Moreover, the whole-brain analysis indicated reduced GMV in the right fusiform gyrus, the left frontal pole, and the right lateral occipital gyrus in individuals diagnosed with SPD.

Comparison of patients with early/mid/late onset of symptoms
Patients with earlier onset of symptoms (before/ during puberty) had less volume in the right OFC than those with late-onset symptoms (Table 5).The analysis also revealed three statistically significant whole-brain findings.Early-onset symptoms were associated with less GMV in the insula and paracingulate gyrus (relative to midonset), and less GMV in the superior temporal gyrus (relative to late-onset).

J o u r n a l P r e -p r o o f
Journal Pre-proof

Comparison of patients with and without comorbid mental disorders
The analysis revealed a statistically significant whole-brain finding.Patients without comorbidity showed greater GMV in the right supplementary motor area compared to patients with comorbidity (MNI coordinates, x,y,z: 6,-3,62, t = 5.20, p(FWE) = .025,number of voxels: 1029).There were no significant ROI findings.

J o u r n a l P r e -p r o o f
Journal Pre-proof In the present study, participants with a diagnosis of skin-picking disorder (SPD) were characterized by reduced grey matter volume (GMV) in brain regions of interest: the insula, the orbitofrontal cortex (OFC), the pallidum, the supramarginal gyrus (SMG), and the cerebellum.
In the following section, associations between these regions of interest and SPD are discussed.Firstly, the insula is involved in affective processing (decoding of stimulus salience/ valence) and interoception (awareness of internal bodily sensations; e.g., Uddin, 2015).These processes are dysfunctional in patients with SPD.For example, excessive skin-picking is associated with difficulties in emotion regulation (e.g., Snorasson et al., 2010, Chesivoir et al., 2023), enhanced emotional reactivity to negative images (e.g., Schienle et al., 2018b, Wabnegger et al., 2018), and deficits concerning the perception of bodily sensations of low intensity (e.g., Houghton et al., 2018, Schienle et al., 2018c).
The insula is tightly interconnected with basal ganglia regions (e.g., pallidum) and the OFC.Like the insula, the OFC plays a role in encoding the affective value and salience of stimuli (Rudebeck et al., 2018).Moreover, the OFC encodes expected outcomes and is involved in flexible, outcome-guided behavior, including inhibitory control and impulse regulation (e.g., Liu et al., 2020).Similar functions have been attributed to the pallidum and also to the cerebellum, which help to control and regulate the initiation and termination of voluntary movements (e.g., Turner & Desmurget, 2010).The relation of these functional areas with SPD becomes clear, considering that insufficient control of motor behaviors (skin manipulation with fingers) despite negative consequences (skin lesions) is the core symptom of SPD.
Furthermore, before the classification of SPD as an OCRD, it had been recognized as an impulse control disorder.Research findings also point to elevated emotion-J o u r n a l P r e -p r o o f Journal Pre-proof based impulsivity in patients with SPD, who however do not appear to be impulsive in other ways (e.g., Snorrason et al., 2011).
In addition, SPD-related volume reduction was found in the supramarginal gyrus (SMG) with functions such as integration of sensory information (body awareness) and social cognition (e.g., understanding other people's emotions and perspectives; see Silani et al., 2014).The exploratory correlation analysis demonstrated that patients who reported more automatic skin-picking showed more pronounced SMG volume reduction.In sum, as an interim conclusion, it can be stated that patients with SPD showed reduced grey matter in brain regions concerned with evaluating and integrating emotionally relevant information, as well as regions involved in subsequent behavioral control.These findings align nicely with the ABC model by Stein et al. (2006) which outlines three components contributing to dysfunctional skin-picking (and other habit disorders).These components include deficits in affect regulation (A), characterized by difficulties in managing negative emotions accompanied by hypo-or hyperarousal; the immediate rewarding nature of skin-picking, representing a behavioral addiction component (B); and reduced cognitive control (C).The authors suggest that the neuronal correlates of these components can be traced to fronto-striatal-cerebellar regions.
Some of the identified brain areas with reduced GMV in SPD overlap with those relevant for obsessive-compulsive disorders (OCD).For example, the VBM mega-analysis by de Wit et al. (2014) detected reduced insula volume in OCD patients (total n = 780).Moreover, Pujol et al. (2004) identified reduced GMV in specific prefrontal cortex regions (including the OFC) and the insula in 77 patients with OCD, relative to healthy controls (n = 72).However, another meta-analysis of VBM studies (928 OCD patients and 942 HCs) found that an OCD diagnosis was J o u r n a l P r e -p r o o f Journal Pre-proof associated with increased volume in the insula, basal ganglia regions, and the cerebellum relative to control subjects (Carlisi et al., 2017).
These discrepant findings may be driven by OCD subgroups, which are differentiated according to different symptom constellations and associated neuroanatomical characteristics.For example, research on OCDs points to the existence of subgroups associated with the age of onset of symptoms.Patients with symptom manifestation during childhood typically demonstrate higher levels of OCD severity (e.g., Anhold et al., 2014;Taylor et al., 2011).The subgrouping in the current study also indicated that earlier onset SPD (before/during puberty) was associated with higher severity of focused skin-picking and less GMV in the OFC and insula compared to late-onset.Since early symptom manifestation was positively correlated with symptom duration in the present study, it remains to be seen whether the observed reductions in regional brain volume reflect alterations in developmental or aging trajectories: On the one hand, reduced insular/OFC volumes could already be present in children with SPD.This, however, has never been investigated before.On the other hand, volume reduction may primarily be a result of persistent SPD symptoms (on average, symptom duration was almost 17 years in the present sample).Thus, longitudinal studies (with prospective analyses of brain scans) are necessary.
It also needs to be noted that the current VBM analysis, comparing patients and healthy controls, obtained significant whole-brain findings.SPD was associated with reduced GMV in the fusiform gyrus, frontal pole, and lateral occipital gyrus.
While the precise functionality of the fusiform gyrus remains incompletely understood, it has been associated with a range of functions related to recognition, high-level vision, and reward processing (Owens et al., 2017).The frontal pole, on the other hand, plays a crucial role in motor planning behavior.For instance, in a study by J o u r n a l P r e -p r o o f Journal Pre-proof Soon et al. (2008), frontopolar activation was observed during a motor decision task in which participants were free to press a button when they felt the urge to do so.
Moreover, the subgroup analysis concerning the age of onset of symptoms revealed significant whole-brain findings: Patients with an early onset displayed less GMV in the paracingulate gyrus, which is implicated in reality monitoring (discriminating between real and imagined information, as well as predictions relating to future intentional social interactions; see Walter et al., 2004), and in the middle temporal gyrus (MTG).The MTG is implicated in reward processing (e.g., delayed reward discounting, which refers to a person's preferences for smaller immediate rewards versus larger delayed rewards; Owens et al., 2017) Finally, comorbidity (the presence of an additional mental disorder) was associated with higher symptom severity, particularly for focused skin-picking and reduced volume in the supplementary motor area (SMA), which is involved in voluntary action (self-initiated movements; Nacev et al., 2008).A study on restingstate functional connectivity in patients with SPD found that higher symptom severity was associated with weaker coupling between the SMA and clusters within the OFC (Huggins et al., 2020).The authors speculated that the OFC, particularly during states of strong urges to pick, may not effectively plan goal-directed behavior and consequently fails to exert top-down regulation of skin-picking behavior.
Several limitations should be considered concerning the present study.The sample was predominantly female, Caucasian, and was characterized by a high educational level.Thus, the present findings cannot be generalized to other sociodemographic groups.Further, the subgrouping (age of symptom onset) was based on patients' self-reports, which may have introduced a small level of imprecision.However, it is important to note that patients were able to recall specific incidences (e.g., a teacher in elementary school who repeatedly asked one patient to stop picking, or puberty-related acne that triggered skin-picking) that substantiated their reports.
Several assets of the present study should be considered as well.The clinical group was characterized by moderate to severe skin-picking symptoms, a low rate of psychotropic medication, and the absence of other diagnoses related to bodyfocused repetitive behaviors (trichotillomania).Finally, the diagnoses of SPD were not only based on self-reports but substantiated by a board-certified clinical psychologist.

Conclusion
This VBM study identified reductions in gray matter volume associated with skinpicking disorder in regions of the brain that are relevant to interoception, emotion regulation, and motor control.The detection of different age-of-onset groups based on clinical as well as morphometric data warrants further investigation.

CRediT authorship contribution statement
Anne Schienle: Conceptualization, Investigation, Writingoriginal draft, Funding acquisition Albert Wabnegger: Methodology, Formal analysis, Writingreview & editing 82): e.g."I pick my skin when I am experiencing a negative emotion such as stress, anger, frustration, or sadness.";automatic (omega = .70):e.g."I am usually not aware of picking my skin during the picking episode.").Each item of the MIDAS is J o u r n a l P r e -p r o o f for patients with SPD (n = 92): automatic picking: M = 17.09(SD = 4.60); focused picking: M = 18.93 (SD = 5.15).

Table 1 : Comparison of patients and controls
Note: SPS_R: Skin-picking scale revised; MIDAS: Milwaukee Inventory for the Dimensions of Adult Skin-Picking; M = mean, SD = standard deviation; CI = confidence intervals 3.
J o u r n a l P r e -p r o o f