P300 Source Localization Contrasts in Body-Focused Repetitive Behaviors and Tic Disorders

Tic disorders (TD) and body-focused repetitive behaviors (BFRB) have similar phenotypes that can be challenging to distinguish in clinical settings. Both disorders show high rates of comorbid psychiatric conditions, dysfunctional basal ganglia activity, atypical cortical functioning in the prefrontal and motor cortical regions, and cognitive deficits. Clinicians frequently confound the two disorders and it is important to find reliable objective methods to discriminate TD and BFRB. Neuropsychological tests and event-related potential (ERP) studies have yielded inconsistent results regarding a possible context updating deficit in TD and BFRB patients. However, most previous studies did not control for the presence of comorbid psychiatric condition and medication status, which might have confounded the findings reported to date. Hence, we aimed to investigate the psychophysiology of working memory using ERP in carefully screened TD and BFRB patients excluding those with psychiatric comorbidity and those taking psychoactive medication. The current study compared 12 TD patients, 12 BRFB patients, and 15 healthy control participants using a motor oddball task (button press). The P300 component was analyzed as an index of working memory functioning. Results showed that BFRB patients had decreased P300 oddball effect amplitudes over the right hemisphere compared to the TD and control groups. Clinical groups presented different scalp distributions compared to controls, which could represent a potential endophenotype candidate of BFRB and TD.


Introduction
Tic disorders (TD) and body-focused repetitive behaviors (BFRB) are two motor-related neuropsychiatric afflictions that can be difficult to distinguish in clinical settings because they often co-occur and because their symptoms can easily be confounded [1,2]. Indeed, large proportions  control participants (n = 15) was recruited via local media to match patients on age, sex, and nonverbal intelligence (evaluated with Raven's matrices [66]). Control participants were screened for psychiatric or neurological diseases and were not taking psychoactive medication. All data presented here were collected prior to the beginning of the therapy for which tic and BFRB patients were initially recruited [63]. The Tourette Syndrome Global Scale [TSGS; 68] and the Yale Global Tic Severity Scale [YGTSS; 69] were used to uniformly quantify tics and BFRB. The TSGS combines a clinical evaluation with reports made by patients and parents to assess tics (simple/complex and motor/phonic) and social functioning in the past week. Global scores vary from 0 (no symptoms) to 100 (worst Tourette possible). Inter-rater reliability, construct validity, and concurrent validity were all found to be good [67]. The YGTSS is a symptom checklist for tic severity (number of tics, frequency, intensity, complexity and interference) and functioning impairments (distress experienced in interpersonal, academic and occupational spheres) over the previous week. The YGTSS global score (0-100) is the sum of subscales evaluating the severity of tics (0-50) and daily impairment (0-50). The YGTSS has demonstrated good internal consistency, test-retest stability, inter-rater agreement and convergent validity (with the TSGS) [68][69][70].
Both the TSGS and the YGTSS were used "as is" in the TD group and adapted to assess the presence of habits in the BFRB group. In these adapted versions, the word "tic" was replaced by the word "habit". The correlations between the adapted scales and appropriate instruments measuring BFRB frequency and severity were reported to be good [1,61]. The Massachusetts General Hospital Hair Pulling Scale (MGH-HPS; [71]) was additionally administered to assess BFRB severity. The MGH-HPS is a seven-point inventory measuring the severity of trichotillomania symptoms. Again, an adaptation of this scale was proposed to assess onychophagia, skin picking, and skin scratching. This scale was thus used to quantify the severity of the principal BFRB of each patient.
Since the focus was to investigate patients unaffected by comorbid disorders, a thorough examination of the most frequent comorbid afflictions was made using the SCID-I/P [72]. Levels of anxiety (Beck Anxiety Inventory (BAI); [73]), depression (Beck Depression Inventory (Beck Depression Inventory (BDI); [74]), and obsessive-compulsive (Vancouver Obsessional Compulsive Inventory (VOCI); [75]) were also assessed ( Table 1). All participants were right-handed (Edinburgh Handedness Inventory (Edinburgh Handedness Inventory; [76]), had normal or corrected to normal visual acuity (Snellen notation system) and gave their written informed consent before participating. The study was approved by the local ethic committee of the Institut universitaire en santé mentale de Montréal.  control participants (n = 15) was recruited via local media to match patients on age, sex, and nonverbal intelligence (evaluated with Raven's matrices [66]). Control participants were screened for psychiatric or neurological diseases and were not taking psychoactive medication. All data presented here were collected prior to the beginning of the therapy for which tic and BFRB patients were initially recruited [63]. The Tourette Syndrome Global Scale [TSGS; 68] and the Yale Global Tic Severity Scale [YGTSS; 69] were used to uniformly quantify tics and BFRB. The TSGS combines a clinical evaluation with reports made by patients and parents to assess tics (simple/complex and motor/phonic) and social functioning in the past week. Global scores vary from 0 (no symptoms) to 100 (worst Tourette possible). Inter-rater reliability, construct validity, and concurrent validity were all found to be good [67]. The YGTSS is a symptom checklist for tic severity (number of tics, frequency, intensity, complexity and interference) and functioning impairments (distress experienced in interpersonal, academic and occupational spheres) over the previous week. The YGTSS global score (0-100) is the sum of subscales evaluating the severity of tics (0-50) and daily impairment (0-50). The YGTSS has demonstrated good internal consistency, test-retest stability, inter-rater agreement and convergent validity (with the TSGS) [68][69][70].
Both the TSGS and the YGTSS were used "as is" in the TD group and adapted to assess the presence of habits in the BFRB group. In these adapted versions, the word "tic" was replaced by the word "habit". The correlations between the adapted scales and appropriate instruments measuring BFRB frequency and severity were reported to be good [1,61]. The Massachusetts General Hospital Hair Pulling Scale (MGH-HPS; [71]) was additionally administered to assess BFRB severity. The MGH-HPS is a seven-point inventory measuring the severity of trichotillomania symptoms. Again, an adaptation of this scale was proposed to assess onychophagia, skin picking, and skin scratching. This scale was thus used to quantify the severity of the principal BFRB of each patient.
Since the focus was to investigate patients unaffected by comorbid disorders, a thorough examination of the most frequent comorbid afflictions was made using the SCID-I/P [72]. Levels of anxiety (Beck Anxiety Inventory (BAI); [73]), depression (Beck Depression Inventory (Beck Depression Inventory (BDI); [74]), and obsessive-compulsive (Vancouver Obsessional Compulsive Inventory (VOCI); [75]) were also assessed ( Table 1). All participants were right-handed (Edinburgh Handedness Inventory (Edinburgh Handedness Inventory; [76]), had normal or corrected to normal visual acuity (Snellen notation system) and gave their written informed consent before participating. The study was approved by the local ethic committee of the Institut universitaire en santé mentale de Montréal. control participants (n = 15) was recruited via local media to match patients on age, sex, and nonverbal intelligence (evaluated with Raven's matrices [66]). Control participants were screened for psychiatric or neurological diseases and were not taking psychoactive medication. All data presented here were collected prior to the beginning of the therapy for which tic and BFRB patients were initially recruited [63]. The Tourette Syndrome Global Scale [TSGS; 68] and the Yale Global Tic Severity Scale [YGTSS; 69] were used to uniformly quantify tics and BFRB. The TSGS combines a clinical evaluation with reports made by patients and parents to assess tics (simple/complex and motor/phonic) and social functioning in the past week. Global scores vary from 0 (no symptoms) to 100 (worst Tourette possible). Inter-rater reliability, construct validity, and concurrent validity were all found to be good [67]. The YGTSS is a symptom checklist for tic severity (number of tics, frequency, intensity, complexity and interference) and functioning impairments (distress experienced in interpersonal, academic and occupational spheres) over the previous week. The YGTSS global score (0-100) is the sum of subscales evaluating the severity of tics (0-50) and daily impairment (0-50). The YGTSS has demonstrated good internal consistency, test-retest stability, inter-rater agreement and convergent validity (with the TSGS) [68][69][70].
Both the TSGS and the YGTSS were used "as is" in the TD group and adapted to assess the presence of habits in the BFRB group. In these adapted versions, the word "tic" was replaced by the word "habit". The correlations between the adapted scales and appropriate instruments measuring BFRB frequency and severity were reported to be good [1,61]. The Massachusetts General Hospital Hair Pulling Scale (MGH-HPS; [71]) was additionally administered to assess BFRB severity. The MGH-HPS is a seven-point inventory measuring the severity of trichotillomania symptoms. Again, an adaptation of this scale was proposed to assess onychophagia, skin picking, and skin scratching. This scale was thus used to quantify the severity of the principal BFRB of each patient.
Since the focus was to investigate patients unaffected by comorbid disorders, a thorough examination of the most frequent comorbid afflictions was made using the SCID-I/P [72]. Levels of anxiety (Beck Anxiety Inventory (BAI); [73]), depression (Beck Depression Inventory (Beck Depression Inventory (BDI); [74]), and obsessive-compulsive (Vancouver Obsessional Compulsive Inventory (VOCI); [75]) were also assessed ( Table 1). All participants were right-handed (Edinburgh Handedness Inventory (Edinburgh Handedness Inventory; [76]), had normal or corrected to normal visual acuity (Snellen notation system) and gave their written informed consent before participating. The study was approved by the local ethic committee of the Institut universitaire en santé mentale de Montréal. control participants (n = 15) was recruited via local media to match patients on age, sex, and nonverbal intelligence (evaluated with Raven's matrices [66]). Control participants were screened for psychiatric or neurological diseases and were not taking psychoactive medication. All data presented here were collected prior to the beginning of the therapy for which tic and BFRB patients were initially recruited [63]. The Tourette Syndrome Global Scale [TSGS; 68] and the Yale Global Tic Severity Scale [YGTSS; 69] were used to uniformly quantify tics and BFRB. The TSGS combines a clinical evaluation with reports made by patients and parents to assess tics (simple/complex and motor/phonic) and social functioning in the past week. Global scores vary from 0 (no symptoms) to 100 (worst Tourette possible). Inter-rater reliability, construct validity, and concurrent validity were all found to be good [67]. The YGTSS is a symptom checklist for tic severity (number of tics, frequency, intensity, complexity and interference) and functioning impairments (distress experienced in interpersonal, academic and occupational spheres) over the previous week. The YGTSS global score (0-100) is the sum of subscales evaluating the severity of tics (0-50) and daily impairment (0-50). The YGTSS has demonstrated good internal consistency, test-retest stability, inter-rater agreement and convergent validity (with the TSGS) [68][69][70].
Both the TSGS and the YGTSS were used "as is" in the TD group and adapted to assess the presence of habits in the BFRB group. In these adapted versions, the word "tic" was replaced by the word "habit". The correlations between the adapted scales and appropriate instruments measuring BFRB frequency and severity were reported to be good [1,61]. The Massachusetts General Hospital Hair Pulling Scale (MGH-HPS; [71]) was additionally administered to assess BFRB severity. The MGH-HPS is a seven-point inventory measuring the severity of trichotillomania symptoms. Again, an adaptation of this scale was proposed to assess onychophagia, skin picking, and skin scratching. This scale was thus used to quantify the severity of the principal BFRB of each patient.
Since the focus was to investigate patients unaffected by comorbid disorders, a thorough examination of the most frequent comorbid afflictions was made using the SCID-I/P [72]. Levels of anxiety (Beck Anxiety Inventory (BAI); [73]), depression (Beck Depression Inventory (Beck Depression Inventory (BDI); [74]), and obsessive-compulsive (Vancouver Obsessional Compulsive Inventory (VOCI); [75]) were also assessed ( Table 1). All participants were right-handed (Edinburgh Handedness Inventory (Edinburgh Handedness Inventory; [76]), had normal or corrected to normal visual acuity (Snellen notation system) and gave their written informed consent before participating. The study was approved by the local ethic committee of the Institut universitaire en santé mentale de Montréal.  control participants (n = 15) was recruited via local media to match patients on age, sex, and nonverbal intelligence (evaluated with Raven's matrices [66]). Control participants were screened for psychiatric or neurological diseases and were not taking psychoactive medication. All data presented here were collected prior to the beginning of the therapy for which tic and BFRB patients were initially recruited [63]. The Tourette Syndrome Global Scale [TSGS; 68] and the Yale Global Tic Severity Scale [YGTSS; 69] were used to uniformly quantify tics and BFRB. The TSGS combines a clinical evaluation with reports made by patients and parents to assess tics (simple/complex and motor/phonic) and social functioning in the past week. Global scores vary from 0 (no symptoms) to 100 (worst Tourette possible). Inter-rater reliability, construct validity, and concurrent validity were all found to be good [67]. The YGTSS is a symptom checklist for tic severity (number of tics, frequency, intensity, complexity and interference) and functioning impairments (distress experienced in interpersonal, academic and occupational spheres) over the previous week. The YGTSS global score (0-100) is the sum of subscales evaluating the severity of tics (0-50) and daily impairment (0-50). The YGTSS has demonstrated good internal consistency, test-retest stability, inter-rater agreement and convergent validity (with the TSGS) [68][69][70].
Both the TSGS and the YGTSS were used "as is" in the TD group and adapted to assess the presence of habits in the BFRB group. In these adapted versions, the word "tic" was replaced by the word "habit". The correlations between the adapted scales and appropriate instruments measuring BFRB frequency and severity were reported to be good [1,61]. The Massachusetts General Hospital Hair Pulling Scale (MGH-HPS; [71]) was additionally administered to assess BFRB severity. The MGH-HPS is a seven-point inventory measuring the severity of trichotillomania symptoms. Again, an adaptation of this scale was proposed to assess onychophagia, skin picking, and skin scratching. This scale was thus used to quantify the severity of the principal BFRB of each patient.
Since the focus was to investigate patients unaffected by comorbid disorders, a thorough examination of the most frequent comorbid afflictions was made using the SCID-I/P [72]. Levels of anxiety (Beck Anxiety Inventory (BAI); [73]), depression (Beck Depression Inventory (Beck Depression Inventory (BDI); [74]), and obsessive-compulsive (Vancouver Obsessional Compulsive Inventory (VOCI); [75]) were also assessed ( Table 1). All participants were right-handed (Edinburgh Handedness Inventory (Edinburgh Handedness Inventory; [76]), had normal or corrected to normal visual acuity (Snellen notation system) and gave their written informed consent before participating. The study was approved by the local ethic committee of the Institut universitaire en santé mentale de Montréal. control participants (n = 15) was recruited via local media to match patients on age, sex, and nonverbal intelligence (evaluated with Raven's matrices [66]). Control participants were screened for psychiatric or neurological diseases and were not taking psychoactive medication. All data presented here were collected prior to the beginning of the therapy for which tic and BFRB patients were initially recruited [63]. The Tourette Syndrome Global Scale [TSGS; 68] and the Yale Global Tic Severity Scale [YGTSS; 69] were used to uniformly quantify tics and BFRB. The TSGS combines a clinical evaluation with reports made by patients and parents to assess tics (simple/complex and motor/phonic) and social functioning in the past week. Global scores vary from 0 (no symptoms) to 100 (worst Tourette possible). Inter-rater reliability, construct validity, and concurrent validity were all found to be good [67]. The YGTSS is a symptom checklist for tic severity (number of tics, frequency, intensity, complexity and interference) and functioning impairments (distress experienced in interpersonal, academic and occupational spheres) over the previous week. The YGTSS global score (0-100) is the sum of subscales evaluating the severity of tics (0-50) and daily impairment (0-50). The YGTSS has demonstrated good internal consistency, test-retest stability, inter-rater agreement and convergent validity (with the TSGS) [68][69][70].
Both the TSGS and the YGTSS were used "as is" in the TD group and adapted to assess the presence of habits in the BFRB group. In these adapted versions, the word "tic" was replaced by the word "habit". The correlations between the adapted scales and appropriate instruments measuring BFRB frequency and severity were reported to be good [1,61]. The Massachusetts General Hospital Hair Pulling Scale (MGH-HPS; [71]) was additionally administered to assess BFRB severity. The MGH-HPS is a seven-point inventory measuring the severity of trichotillomania symptoms. Again, an adaptation of this scale was proposed to assess onychophagia, skin picking, and skin scratching. This scale was thus used to quantify the severity of the principal BFRB of each patient.
Since the focus was to investigate patients unaffected by comorbid disorders, a thorough examination of the most frequent comorbid afflictions was made using the SCID-I/P [72]. Levels of anxiety (Beck Anxiety Inventory (BAI); [73]), depression (Beck Depression Inventory (Beck Depression Inventory (BDI); [74]), and obsessive-compulsive (Vancouver Obsessional Compulsive Inventory (VOCI); [75]) were also assessed ( Table 1). All participants were right-handed (Edinburgh Handedness Inventory (Edinburgh Handedness Inventory; [76]), had normal or corrected to normal visual acuity (Snellen notation system) and gave their written informed consent before participating. The study was approved by the local ethic committee of the Institut universitaire en santé mentale de Montréal.

Experimental Task: The Motor Oddball Paradigm
A classical oddball task was administered to all participants. Stimuli were presented using the Presentation software (Neurobehavioural Systems, Albany, CA, USA) and consisted of 200 black letters (Xs and Os) in Arial font (size = 48) on a white background displayed in a pseudo-random order. The letter "O" was the frequent stimulus (probability = 0.80) and the letter "X" was the rare stimulus (probability = 0.20). Stimuli were presented during 100 ms each at the center of a monitor screen (Viewsonic SVGA 17 inches flat screen monitor) with an inter-stimulus interval randomly ranging from 1700 to 2200 ms. Participants were seated in a dimly lit room with their head positioned 90 cm from the monitor. Before the beginning of the task, they were instructed to press the left arrow of the keyboard with their left hand when they perceived a rare (letter X) stimulus and the right arrow with their right hand for frequent (letter O) stimuli.

ERP Extraction
Ocular artifacts that contaminated the EEG signals were corrected offline using the Gratton algorithm [78]. Subsequently, raw EEG signals were averaged offline, time-locked to the stimuli onset, in a time window of 100 ms prior to stimulus onset until 900 ms after stimulus onset. The averaging procedure was completed separately for the two conditions, corresponding to rare and frequent categories. Averaged signals were digitally filtered offline with low-and high-pass filters set at 0.3 Hz and 30 Hz, respectively. Remaining signals exceeding 100 µV and clippings due to amplifier saturation were eliminated during the averaging procedure. Epochs containing fewer than 20 trials for each category were eliminated. The P300 component was scored as the maximum peak during the 300-550 ms post-stimulus interval.

Statistical Analyses
Sociodemographic, clinical, and behavioral data were analyzed with one-way ANOVAs when a three groups comparison was involved, while independent-sample t-tests (bilateral α = 0.05) or Kruskall-Wallis tests (for non-parametric data) were used when the comparisons involved only the two clinical groups. Tukey post hoc tests were used when variances were homogeneous (according to Levene's test), while Games-Howell tests were utilized for heterogeneous variances. The electrophysiological data were analyzed using repeated measures ANOVAs and the Greenhouse-Geisser correction was applied when the sphericity assumption was not met. One between-subject factor with three levels (Group: TD . Additionally, analyses were conducted with midline electrodes with the same between-subject factor and the following within-subject factors: Condition (Rare, frequent) and Electrode sites (6 levels; AFZ, FZ, FCZ, CZ, CPZ, PZ).

P300 Source Localization
In a subsequent analysis, neural generators of the averaged ERPs were analyzed with sLORETA software (University Hospital of Psychiatry, Zürich, Switzerland) in the P300 time interval (300-550 ms post-stimulus) showing significant differences between experimental conditions (rare vs. frequent). sLORETA uses a distributed source localization algorithm to solve the inverse problem of brain electric activity [79], regardless of the final number of neural generators [79,80]. The sLORETA algorithm calculates the current density values (unit: amperes per square meter; A/m 2 ) of 6.239 gray matter voxels belonging to the brain compartment with a spatial resolution of 5 mm × 5 mm × 5 mm each. The whole three-dimensional brain compartment comprises cortical gray matter and the hippocampus only and does not contain any deep brain structures such as the thalamus or the cerebellum. Anatomical regions are labeled according to (1) the probabilistic MNI-152 template made digitally available by the Brain Imaging Center of the Montreal Neurological Institute [81] and (2)  To display differences for the P300 time window between the two experimental conditions (e.g., "Rare deviant" > "frequent standard" for the oddball paradigm), statistical non-parametric mapping (SnPM) as introduced by Nichols and Holmes [84] was used to compute the averaged intracerebral current density distribution at time intervals showing significant differences based on non-parametric voxel-by-voxel one-tailored paired samples log of f-ratio of averages (with 5000 permutations) on the three-dimensional sLORETA images. Statistical significance was assessed by defining critical thresholds (critical t) corrected for multiple comparisons (p < 0.01 and p < 0.05, respectively) for all tested voxels. The same procedure was done to display differences to compare the three group differences on the P300 time oddball effect (e.g., "tic disorders" > "control"; "tic disorders" > "BFRB"; "BFRB" > "control").
The null hypotheses equaled the assumption that there were no differences among experimental conditions. Current density values at each voxel were computed in the solution space as a linear and weighted sum of the scalp electric potentials. Activation of a given voxel was based on the smoothness assumption, meaning that neighboring voxels show a highly synchronous activity [85]. Support comes from electrophysiological studies showing that neighboring neural populations show a highly correlated electrical activity [85,86]. As proposed by Friston [87], activated voxels exceeding critical t values were considered as being regions of cortical activation. Finally, statistical analysis resulted in an averaged corresponding three-dimensional intracerebral current density distribution and obtained cortical regions were classified about their corresponding BA [88] and normalized coordinates (Talairach and MNI, respectively).

Socio-Demographic and Clinical Description
No significant group difference was found regarding age, education level, and non-verbal intelligence (Table 1). Gender ratio was significantly different between TD and BFRB groups (H(2) = 6.55, p = 0.04) (followed by post hoc Mann-Whitney Test U = 36, p = 0.04, r = −0.52). All groups were comparable regarding depression levels, but BFRB patients showed significantly higher anxiety scores than the control group (F(2,36) = 3.92, p = 0.03, r = 0.30). Both clinical groups were comparable regarding obsessive-compulsive symptoms.  = 0.78) revealed faster responses in the frequent (421 ms) than in the rare (491 ms) condition. There was no group difference regarding reaction times. Additionally, all participants showed excellent performance accuracy in rare (93% average accuracy) and frequent (97% average accuracy) conditions, with no group difference at that level.

Amplitude of the P300 Component
The P300 component peaked at an average latency of 396 ms after stimulus onset. The omnibus ANOVA revealed significant Condition (F(1,36) = 149.71, p < 0.001, partial ntory (BDI); [74]), and obsessive-compulsive (Vancouver Obsessional Compulsive I); [75]) were also assessed ( Table 1). All participants were right-handed (Edinburgh entory (Edinburgh Handedness Inventory; [76]), had normal or corrected to normal ellen notation system) and gave their written informed consent before participating. pproved by the local ethic committee of the Institut universitaire en santé mentale de e 1. Sociodemographic and clinical characteristics. Means (standard deviations).  (3) --13 (6) --12 (5) --12 (5) --15 (5) -tric data (sex); BAI: Beck Anxiety Inventory (max. score 63); BFRB: Body-focused repetitive behaviors; Pulling Scale; ns: non-significant VOCI: Vancouver core 220); OCS: Obsessive-Compulsive Symptoms; e 100). Significant results are in bold. quent comorbid afflictions was made using the SCID-I/P [72]. Levels of ventory (BAI); [73]), depression (Beck Depression Inventory (Beck ); [74]), and obsessive-compulsive (Vancouver Obsessional Compulsive e also assessed ( Table 1). All participants were right-handed (Edinburgh nburgh Handedness Inventory; [76]), had normal or corrected to normal on system) and gave their written informed consent before participating. y the local ethic committee of the Institut universitaire en santé mentale de ographic and clinical characteristics. Means (standard deviations). = 0.15). This four-way interaction was decomposed by condition. No group interaction was significant in the frequent condition. However, the Hemisphere by Electrode by Group (F(8,144) = 3.97, p < 0.001, partial e focus was to investigate patients unaffected by comorbid disorders, a thorough of the most frequent comorbid afflictions was made using the SCID-I/P [72]. Levels of ck Anxiety Inventory (BAI); [73]), depression (Beck Depression Inventory (Beck Inventory (BDI); [74]), and obsessive-compulsive (Vancouver Obsessional Compulsive OCI); [75]) were also assessed ( Table 1). All participants were right-handed (Edinburgh Inventory (Edinburgh Handedness Inventory; [76]), had normal or corrected to normal (Snellen notation system) and gave their written informed consent before participating. as approved by the local ethic committee of the Institut universitaire en santé mentale de s used to quantify the severity of the principal BFRB of each patient. cus was to investigate patients unaffected by comorbid disorders, a thorough e most frequent comorbid afflictions was made using the SCID-I/P [72]. Levels of nxiety Inventory (BAI); [73]), depression (Beck Depression Inventory (Beck tory (BDI); [74]), and obsessive-compulsive (Vancouver Obsessional Compulsive ; [75]) were also assessed ( Table 1). All participants were right-handed (Edinburgh ntory (Edinburgh Handedness Inventory; [76]), had normal or corrected to normal llen notation system) and gave their written informed consent before participating. proved by the local ethic committee of the Institut universitaire en santé mentale de = 0.10) interactions reached the significance threshold in the Rare condition, meaning that group differences were mainly found in the Rare condition.
The four-way interaction of Region by Hemisphere by Electrode by Group was decomposed by hemisphere. There was no significant group interaction within the left hemisphere. A significant Electrode by Group (F(8,144) = 2.73, p = 0.02, partial 1]) was additionally administered to assess BFRB severity. The ory measuring the severity of trichotillomania symptoms. Again, posed to assess onychophagia, skin picking, and skin scratching. y the severity of the principal BFRB of each patient. stigate patients unaffected by comorbid disorders, a thorough comorbid afflictions was made using the SCID-I/P [72]. Levels of (BAI); [73]), depression (Beck Depression Inventory (Beck and obsessive-compulsive (Vancouver Obsessional Compulsive ssessed ( Table 1). All participants were right-handed (Edinburgh Handedness Inventory; [76]), had normal or corrected to normal em) and gave their written informed consent before participating. cal ethic committee of the Institut universitaire en santé mentale de = 0.13) interaction was found in the right hemisphere, which indicates between-group differences over the right hemisphere, in response to the rare condition. The latter interaction of Electrode by Group was decomposed according to the following electrode lines (i.e., 1st AF4/FC4/CP6; 2nd AF2/FC2/CP2; 3rd F6/C6/P6; 4th F4/C4/P4; 5th F2/C2/P2). BFRB patients showed a significantly reduced P300 amplitude compared to controls, over the 1st line AF4/FC4/CP6 (F(2,36) = 5.49, p = 0.008, partial e YGTSS were used "as is" in the TD group and adapted to assess the RB group. In these adapted versions, the word "tic" was replaced by the ons between the adapted scales and appropriate instruments measuring ty were reported to be good [1,61]. The Massachusetts General Hospital HPS; [71]) was additionally administered to assess BFRB severity. The inventory measuring the severity of trichotillomania symptoms. Again, was proposed to assess onychophagia, skin picking, and skin scratching. quantify the severity of the principal BFRB of each patient. to investigate patients unaffected by comorbid disorders, a thorough equent comorbid afflictions was made using the SCID-I/P [72]. Levels of ventory (BAI); [73]), depression (Beck Depression Inventory (Beck ); [74]), and obsessive-compulsive (Vancouver Obsessional Compulsive re also assessed ( Table 1). All participants were right-handed (Edinburgh inburgh Handedness Inventory; [76]), had normal or corrected to normal ion system) and gave their written informed consent before participating. ency, test-retest stability, inter-rater agreement and convergent were used "as is" in the TD group and adapted to assess the p. In these adapted versions, the word "tic" was replaced by the een the adapted scales and appropriate instruments measuring reported to be good [1,61]. The Massachusetts General Hospital ]) was additionally administered to assess BFRB severity. The ry measuring the severity of trichotillomania symptoms. Again, osed to assess onychophagia, skin picking, and skin scratching. the severity of the principal BFRB of each patient. igate patients unaffected by comorbid disorders, a thorough omorbid afflictions was made using the SCID-I/P [72]. Levels of (BAI); [73]), depression (  is" in the TD group and adapted to assess the ted versions, the word "tic" was replaced by the d scales and appropriate instruments measuring ood [1,61]. The Massachusetts General Hospital ally administered to assess BFRB severity. The e severity of trichotillomania symptoms. Again, nychophagia, skin picking, and skin scratching. the principal BFRB of each patient. unaffected by comorbid disorders, a thorough ons was made using the SCID-I/P [72]. Levels of epression (Beck Depression Inventory (Beck ompulsive (Vancouver Obsessional Compulsive ). All participants were right-handed (Edinburgh entory; [76]), had normal or corrected to normal ir written informed consent before participating. tee of the Institut universitaire en santé mentale de racteristics. Means (standard deviations). inter-rater agreement and convergent TD group and adapted to assess the s, the word "tic" was replaced by the d appropriate instruments measuring . The Massachusetts General Hospital istered to assess BFRB severity. The of trichotillomania symptoms. Again, gia, skin picking, and skin scratching. al BFRB of each patient. by comorbid disorders, a thorough ade using the SCID-I/P [72]. Levels of (Beck Depression Inventory (Beck (Vancouver Obsessional Compulsive ipants were right-handed (Edinburgh ]), had normal or corrected to normal informed consent before participating.

Institut universitaire en santé mentale de
Means (standard deviations).  Figure 1). Furthermore, BFRB patients also showed a significantly reduced P300 compared to TD patients over the 3rd electrode line F6/C6/P6 (p = 0.02; Bonferroni: p < 0.05). Despite a general tendency toward a P300 reduction, TD patients showed amplitudes that were statistically comparable to control participants. The aforementioned electrophysiological differences between groups were not explained by BFRB patients' higher scores on the BAI since the Condition by Hemisphere by Electrode by Group interaction from the omnibus test remained significant after covarying for BAI scores (F(8,140) = 2.33, p = 0.03).

sLORETA Source Localization-The Rare-Frequent Oddball Contrast
The statistical comparison between experimental conditions "frequent" and "rare" in the peak activation within a time window between 300 and 550 ms (corresponding to the P300 oddball effect) was done across the three groups separately. Results of Table 2 showed that in the TD group, a significantly (p between 0.0012 and 0.0018) larger cortical activation of the oddball effect at 518 ms post-stimulus was expressed in the superior and middle frontal gyri which corresponds to BA6 and BA10 (Figure 2). The Table 3 showed that in the BFRB group a larger oddball effect activation at 402 ms post-stimulus was significant (p between 0.0014 and 0.0001) in the right parietal lobe, mainly within the precentral (BA43) and postcentral (BAs 1/2/3) gyri as well as the right inferior parietal gyrus (BA 40). Other right hemisphere structures are also involved, including mainly the right superior temporal gyrus (BAs 42/22), the inferior frontal gyrus (BAs 9/44/45), and the frontal precentral gyrus (BA 4/6) ( Figure 3). The Table 4 showed that in the control group (Figure 4)  ERPs in the rare condition over the right hemisphere according to a division by electrode line. The x-axis represents the amplitude (μV) and the y-axis represents the time in milliseconds. The ERPs were time-locked to the stimulus onset. The P300 was measured from baseline-to-peak between 300-550 ms post-stimulus onset. Arrows show significantly smaller P300 voltages for BFRB patients compared to control participants over the first, third, fourth and fifth electrode lines.

The Group Comparison Contrast
The statistical comparison between groups ("control"; "tic" and "BFRB") in the peak activation within a time window between 300 and 550 ms was done on the rare-frequent subtraction. Results of Table 5 showed that in the control/BRFB comparison, a significantly larger difference in cortical activation of the oddball effect peaked at 366 ms post-stimulus significant (p between 0.006 and 0.012) in the right hemisphere superior frontal gyrus (BA6) and middle temporal gyrus (BA39) ( Figure 5) confirming that BFRB patients showed a significantly reduced P300 oddball effect compared to the controls, mainly over the right superior frontal gyrus. Table 6 showed that, in the tic/BFRB comparison, a significantly larger difference in cortical activation of the oddball effect peaked at 324 ms post-stimulus significant (p between 0.0004 and 0.0040) in the right hemisphere middle temporal gyrus (BA22) and middle temporal gyrus (BA39) (see Figure 6) confirming that BFRB patients showed a significantly reduced P300 oddball effect compared to the TD patients, mainly over the right middle temporal gyrus. In the control/tic group comparison (Figure 7) a larger difference in cortical activation of the oddball effect peaked at 356 ms post-stimulus, but there is no significant difference between groups (p between 0.1708 and 0.2942).
the probabilistic MNI-152 template [81]. Activated voxels are indicated by yellowish and reddish colors (after correction for multiple comparisons (p < 0.01 and p < 0.05, respectively)). (A) The peak of highest cortical activity has been found in parts of the superior frontal gyrus (BA10), the medial frontal gyrus (BA6), the paracentral lobule (BA31), and the cingulate gyrus (BA24). (B) A shifted transverse and sagittal view of the left and the right hemisphere, showing cortical activations on the threedimensionally rendered Colin27 template [89] L, left; R, right; A, anterior; P, posterior; MNI, Montreal Neurological Institute; X, Y, Z, corresponding MNI coordinates; BA, Brodmann Area. Figure 3. Results of the Body-focused Repetitive Behaviors patients based on the standardized lowresolution brain electrotomography (sLORETA) current source density localization of the oddball effect (contrast: "Rare"> Frequent") in the peak activation of the P300 component within a 300-550 ms time window. Images have been obtained after statistical non-parametric mapping (SnPM) and coregistration to the stereotaxic Talairach space based on the Co-Planar Stereotaxic Atlas of the Human Brain [83] and the probabilistic MNI-152 template [81] Activated voxels are indicated by yellowish and reddish colors (after correction for multiple comparisons (p < 0.01 and p < 0.05, respectively)). (A) The peak of highest cortical activity has been found in parts of the inferior parietal lobule (BA40), the postcentral gyrus (BA43/3/1) and the superior temporal gyrus (B42/22). (B) A shifted lateral view of the right hemisphere, showing cortical activations on the three-dimensionally rendered Colin27 template [89]. A, anterior; P, posterior; MNI, Montreal Neurological Institute; X, Y, Z, corresponding MNI coordinates; BA, Brodmann area. Figure 3. Results of the Body-focused Repetitive Behaviors patients based on the standardized low-resolution brain electrotomography (sLORETA) current source density localization of the oddball effect (contrast: "Rare"> Frequent") in the peak activation of the P300 component within a 300-550 ms time window. Images have been obtained after statistical non-parametric mapping (SnPM) and co-registration to the stereotaxic Talairach space based on the Co-Planar Stereotaxic Atlas of the Human Brain [83] and the probabilistic MNI-152 template [81] Activated voxels are indicated by yellowish and reddish colors (after correction for multiple comparisons (p < 0.01 and p < 0.05, respectively)).  Images have been obtained after statistical non-parametric mapping (SnPM) and co-registration to the stereotaxic Talairach space based on the Co-Planar Stereotaxic Atlas of the Human Brain [83] and the probabilistic MNI-152 template [81]. Activated voxels are indicated by yellowish and reddish colors (after correction for multiple comparisons (p < 0.01 and p < 0.05, respectively)).  Images have been obtained after statistical non-parametric mapping (SnPM) and co-registration to the stereotaxic Talairach space based on the Co-Planar Stereotaxic Atlas of the Human Brain [83] and the probabilistic MNI-152 template [81]. Activated voxels are indicated by yellowish and reddish colors (after correction for multiple comparisons (p < 0.01 and p < 0.05, respectively)).      the stereotaxic Talairach space based on the Co-Planar Stereotaxic Atlas of the Human Brain [83] and the probabilistic MNI-152 template [81]. Activated voxels are indicated by yellowish and reddish colors (after correction for multiple comparisons (p < 0.01 and p < 0.05, respectively)). (A) The peak of highest cortical activity has been found in parts of the cuneus (BA7/18) and the precuneus (BA19). (B) A shifted view of the right hemisphere and the posterior cortex, showing cortical activations on the three-dimensionally rendered Colin27 template [89]. L, left; R, right; A, anterior; P, posterior; MNI, Montreal Neurological Institute; X, Y, Z, corresponding MNI coordinates; BA, Brodmann area.   Figure 6. Comparison of the oddball effect between Tic and BFRB patients based on the standardized low-resolution brain electrotomography (sLORETA) current source density localization ofthe group effect (contrast: "BFRB'' < ''tic") in the peak activation of the P300 component within a 300-550 ms time window. Images have been obtained after statistical non-parametric mapping (SnPM) and coregistration to the stereotaxic Talairach space based on the Co-Planar Stereotaxic Atlas of the Human Brain [83] and the probabilistic MNI-152 template [81]. Activated voxels are indicated by blue colors (after correction for multiple comparisons (p < 0.05)). (A) The peak of highest cortical activity of group difference has been found in parts of the middle temporal gyrus (BA22) and the middle temporal gyrus (BA39). (B) A shifted transverse and coronal view of the right hemisphere and the occipital region, showing cortical activations on the three-dimensionally rendered Colin27 template [89]. A, anterior; P, posterior; L, Left; R, Right. MNI, Montreal Neurological Institute; X, Y, Z, corresponding MNI coordinates; BA, Brodmann area. Figure 6. Comparison of the oddball effect between Tic and BFRB patients based on the standardized low-resolution brain electrotomography (sLORETA) current source density localization ofthe group effect (contrast: "BFRB" < "tic") in the peak activation of the P300 component within a 300-550 ms time window. Images have been obtained after statistical non-parametric mapping (SnPM) and co-registration to the stereotaxic Talairach space based on the Co-Planar Stereotaxic Atlas of the Human Brain [83] and the probabilistic MNI-152 template [81]. Activated voxels are indicated by blue colors (after correction for multiple comparisons (p < 0.05)). (A) The peak of highest cortical activity of group difference has been found in parts of the middle temporal gyrus (BA22) and the middle temporal gyrus (BA39 (after correction for multiple comparisons (p < 0.05)). (A) The peak of highest cortical activity of group difference has been found in parts of the middle temporal gyrus (BA22) and the middle temporal gyrus (BA39). (B) A shifted transverse and coronal view of the right hemisphere and the occipital region, showing cortical activations on the three-dimensionally rendered Colin27 template [89]. A, anterior; P, posterior; L, Left; R, Right. MNI, Montreal Neurological Institute; X, Y, Z, corresponding MNI coordinates; BA, Brodmann area. Figure 7. Results of the tic disorder patients based on the standardized low-resolution brain electrotomography (sLORETA) current source density localization of the oddball effect (contrast:"Ctrl" > "Tic") in the mean activation of the P300 component within a 300-550 ms time window. Images have been obtained after statistical non-parametric mapping (SnPM) and coregistration to the stereotaxic Talairach space based on the Co-Planar Stereotaxic Atlas of the Human Brain [83] and the probabilistic MNI-152 template [81]. There is no significant activated voxels (after correction for multiple comparisons (p < 0.05)). (A) Although non-significant, the peak of highest cortical activity has been found in parts of the temporal gyrus (BA22), supramarginal gyrus (BA40). (B) A shifted transverse and sagittal view of the left and the right hemisphere, showing cortical activations on the three-dimensionally rendered Colin27 template [89]. A, anterior; P, posterior; MNI, Montreal Neurological Institute; X, Y, Z, corresponding MNI coordinates; BA, Brodmann area. Figure 7. Results of the tic disorder patients based on the standardized low-resolution brain electrotomography (sLORETA) current source density localization of the oddball effect (contrast:"Ctrl" > "Tic") in the mean activation of the P300 component within a 300-550 ms time window. Images have been obtained after statistical non-parametric mapping (SnPM) and co-registration to the stereotaxic Talairach space based on the Co-Planar Stereotaxic Atlas of the Human Brain [83] and the probabilistic MNI-152 template [81]. There is no significant activated voxels (after correction for multiple comparisons (p < 0.05)). (A) Although non-significant, the peak of highest cortical activity has been found in parts of the temporal gyrus (BA22), supramarginal gyrus (BA40

Discussion
The current study aimed at investigating P300 processes during a motor oddball task, in carefully selected unmedicated patients affected solely by TD or BFRB, without comorbid diagnosis. We initially hypothesized that both patient groups would show reduced P300 amplitude compared to controls, and that TD should have more prominent P300 scalp topography over motor regions, while BFRB should activate a more widespread activation of the same component suggesting qualitatively different patterns of electrocortical activity related to P300 context updating process. ERP analysis revealed that patients with TD showed no significant P300 effect, while the BFRB group presented a significant P300 amplitude reduction, shifted over the right temporal region, in comparison to both TD and control groups. sLORETA analysis further confirmed that the two clinical groups exhibited different scalp distributions of the oddball effect, compared to the control group. The contrasts between groups also revealed significant differences between the control and BFRB, and between BFRB and TD, but no current source topography differences between the control and the TD group.
The P300 is considered to be influenced, in part, by context updating in working memory [90], as well as stimulus evaluation and categorization [91,92]. The P300 amplitude also indexes processing related to the maintenance of the surrounding environment's exact representation in the brain [93]. It has been proposed that the P300 component would index brain activities underlying revision of the mental representation induced by incoming stimuli [94,95]. Concerning the scalp topography, the P300 has previously shown a robust temporo-posterior distribution in healthy subjects. By combining ERPs and fMRI, Li et al. [96] observed that the P300 was generated in the parietal lobules, frontal and central gyri, and in the anterior cingulate cortex. Earlier research involving healthy participants with the motor oddball task showed localized activation of the supplementary motor area, the cerebellum, the thalamus, and the parietal cortex. An additional activation of the middle frontal gyrus was observed, along with the central opercular cortex and the parietal operculum, which reflect integration of sensory and motor responses [97]. Some of these structures are affected in BFRB patients [21], which show increased gray matter density in the cingulate and the parietal regions. Other results also proposed, that TD patients show compensatory activity, notably within the prefrontal and cingulate cortical regions [98]. Our results showed that while all groups presented similar RTs and response accuracy, only patients with BFRB had a significantly smaller P300 oddball effect compared to the other two groups. In addition, the source localization of the oddball effect clearly reveals a right hemisphere shift within the inferior parietal lobule, the postcentral and the superior temporal gyri. Hence, the increased gray matter density in the cingulate and parietal regions, previously reported in BFRB patients, might underlie an increased mobilization of available resources to perform that task at the same level as the controls. Correspondingly, the P300 oddball effect decrease in BFRB patients could reflect a failure to shift their attentional focus toward the actual context, since they are coping with thoughts patterns and feelings driving them to perform their BFRB [99]. This could partially explain the difference in P300 amplitude and the specific right hemisphere scalp topography between BFRB patients and healthy controls. A recent review underlined the prominent role for the right hemisphere-particularly the right inferior parietal lobule-involved in representing regularities and in distinguishing when new information is congruent or discrepant with respect to the predictions generated by a currently held mental model [100]. Another way to approach that question is to integrate a cognitive, behavioral and physiological scheme to functionally link these symptom clusters with neurocognitive processes. For instance, according to the cognitive-behavioral and psychophysiological model [99], context updating processes are important in both TD and BFRB populations. Within that perspective, patients have a specific style of action planning, which drives them to update the production and reinforcement of their tics and BFRB continuously (i.e., the context). If we assume that the right inferior parietal lobule is involved in the processing of salient and surprising information with respect to a current internal model, this mean that the BFRB would overactivate the right hemisphere regions (including the inferior parietal lobule) in order to reinforce new information that conforms to model-generated expectations [100]. Indeed, these patients have the perfectionist desire to always stay ahead of themselves. Therefore, this inclination leads to set their attentional focus toward unrealistic goals, and not toward the "here and now" or the actual model-relevant context [99].
Interestingly, recent studies reported enhanced frontal P300 amplitudes in unmedicated OCD patients [101][102][103][104][105][106][107], which suggests that BFRB and OCD patients may be distinct disorders with similar behavioral symptoms, while showing dissimilar P300 pattern. During a countermanding motor task, TD and BFRB patients both failed to adjust their hand responses to automated or controlled movements [48]. To perform that motor task, TD patients had the most severe divergence in the synchrony between their electrocortical activity (readiness potentials) and their actual motor responses, followed by the BFRB and the control group. Their findings gave support to a dimensional model of classification with BFRB falling closer to TD than to the control group along a continuum of motor arousal. Altogether, these results further showed that BFRB, TD, and OCD correspond to a clinical spectrum that could share some commonalities and discrepancies in their cerebral motor networks.
Earlier findings suggest that comorbidity or medication status may have influenced the interpretation of previously reported ERP recordings of the P300 that often showed a reduction in amplitude in TD [51,108]. Therefore, the presence of comorbid psychiatric disorders as well as adrenergic medications may both impact on the P300, which would lead to atypical electrocortical activity, not entirely related to tic disorders. The presence of attention deficit hyperactive disorder-a frequently comorbid disorder with TD-is also associated with irregular release of norepinephrine in frontal areas [109], which could also impact on the P300 amplitude. To that extent, pharmacological treatment for tics, such as Clonidine and Guanfacine (noradrenergic receptor antagonists [110]), may have involved the cingulate cortex, a region previously identified as one of multiple P300 generators [96,111]. Given that norepinephrine is released in the cingulate cortex among other regions, the reduction in P300 amplitude may then be related to medication status rather than TD per se. Also, we cannot rule out the possibility of an overlap of movement-related potentials with the P300 component [112], which could have affected P300 amplitude in TD patients. This is especially true since increased lateralized readiness potential amplitude has been reported in TD patients [113].

Conclusions
To sum up, our study showed decreased P300 amplitudes and different distribution of cortical activity in unmedicated patients solely affected by BFRB (i.e., without comorbid psychiatric condition). This pattern of electrocortical activation represents a potential endophenotype candidate of BFRB where the P300 amplitude scalp topography may reveal important clues of subtle alterations, probably in the gray matter density of the cingulate and parietal regions. While our study has some limitations, as the small sample size, the exclusion of medicated and comorbid TS patients represents a strength. In this vein, future studies are encouraged to control for medication and comorbidities such as depression, anxiety or obsessive-compulsive disorders given their important influence on ERP recordings. Moreover, our results support previous findings of an unaltered P300, measured by a motor oddball task, in unmedicated patients with TD as their sole psychiatric affliction [50,52]. Future research could include patients suffering from obsessive-compulsive disorders, when comparing ERP patterns of unmedicated patients suffering from TD and BFRB, given the high comorbidity between those three disorders. According to the DSM-5 [114], trichotillomania and skin picking are classified in the obsessive-compulsive and related disorders category, while onychophagia and dermatophagia are mentioned as "other specified obsessive-compulsive and related disorders". Although they are all classified in the same category and broadly conceptualized as BFRB, their clinical presentation can be rather heterogeneous, which could confound our results. Yet, given the recent classification of these disorders, in the category of obsessive-compulsive and related symptoms, the inclusion of a group of OCD patients might deepen our understanding of the neurocognitive chronometry specific to BFRB.

Clinical Implications
Unmedicated TD and BFRB patients with no psychiatric comorbidity showed distinct P300 scalp topography patterns during a motor oddball task. This could constitute a promising electrocortical marker distinguishing BFRB from TD.

Limitations
Future studies could include a group of patients afflicted with obsessive-compulsive disorder given their high comorbitity rates with TD and BFRB to further refine distinctions. These results require replications with larger sample size.