教育学院教师刘洋在NEUROSCIENCE AND BIOBEHAVIORAL REVIEWS期刊上发表最新研究成果.pdf 
教育学院教师刘洋在NEUROSCIENCE AND BIOBEHAVIORAL REVIEWS期刊上发表最新研究成果.pdf
Neuroscience and Biobehavioral Reviews 148 (2023) 105127 Contents lists available at ScienceDirect Neuroscience and Biobehavioral Reviews journal homepage: www.elsevier.com/locate/neubiorev Addiction as a brain disease? A meta-regression comparison of error-related brain potentials between addiction and neurological diseases Yang Liu a, *, Fabio Masina b, K. Richard Ridderinkhof c, 1, Rachele Pezzetta b, 1 a Department of Psychology, School of Education, Shanghai Normal University, Shanghai, China IRCCS San Camillo Hospital, Venice, Italy c Department of Psychology, University of Amsterdam, Amsterdam, the Netherlands b A R T I C L E I N F O A B S T R A C T Keywords: Brain-disease model Addiction Neurological disorders ERN Pe Meta-analysis The notion that addiction is a “brain disorder” is widespread. However, there is a lack of evidence on the degree of disorder in terms of error processing in addiction. The present meta-analysis aimed at shedding light on this by comparing error-processes with populations with well-recognized brain disorders. We included 17 addiction and 32 neurological disorder studies that compared error-related negativity (ERN) or error positivity (Pe) ampli tudes/latencies between experimental and healthy-control groups. Meta-regression analyses were performed for the intergroup comparison and other moderators. Both diagnoses were accompanied by a diminished ERN amplitude, although the degree of impairment was marginally larger in neurological disorders. Neurological disorders presented shorter ERN latencies than addiction when compared with controls. The two groups did not differ in Pe amplitude/latency. Except for a reduced ERN amplitude found along with aging, no other moderator contributed significantly to divergent findings about these four ERP indexes. The results support the brain disease model of addiction, while stressing the importance of quantifying the degrees of brain dysfunctions as a next step. 1. Introduction possibly without treatment, which stands in sharp contrast to the causes and the possibilities for recovery associated with diseases that incur neurological damage (Satel and Lilienfeld, 2014). Accordingly, the neural dysfunction characteristic of addiction is insufficient for perma nent impairment, thus countering the notion that addiction is best un derstood as a brain disease (Levy, 2013). 1.1. The debate about addiction as a brain disease The question of whether addiction is a brain disease, first debated in Leshner’s (1997) influential Science paper, has triggered considerable dispute ever since. The brain-disease model was built upon countless neuroimaging findings that indicated structural and functional brain differences between people with addictive disorders and healthy con trols (Xiao et al., 2015). Accordingly, the term of ‘brain disease’ can be used interchangeably with ‘brain disorder’ in this case (not referring to examples of forms due to infection). For instance, drug or alcohol abuse was associated with brain structure changes, ranging from minor dam age to neurons to severe brain injuries similar to neurological diseases such as traumatic brain injury (TBI), stroke, and multiple sclerosis (Bechara, 2005; Bechara et al., 2000; Goldstein et al., 2009; Volkow et al., 2004). This suggests that both addiction and neurological disor ders are associated with brain disorders. The main arguments of those who oppose this view are that the choice of using a substance was made by oneself, and one can recover 1.2. Analogy of addiction and neurological disorder in error-related processes To shed new light on this debate, our present aim is to provide unique and robust evidence on the analogies and dissimilarities of brain dysfunction between addiction and other types of neurological disease, focusing in particular on the neural activities underlying error process ing through a meta-analysis. The rationale for comparing addiction and neurological disease regarding error-processing related neural activities includes but is not limited to: 1) brain regions involved in reinforcement learning (which capitalizes on calculating prediction error and using it to learn) were more or less impaired in addiction as well as neurological disease, such * Correspondence to: Department of Psychology, School of Education, Shanghai Normal University, 100th Guilin Road, Xuhui District, Shanghai, China. E-mail address: yliux@outlook.com (Y. Liu). 1 Shared senior authorship. https://doi.org/10.1016/j.neubiorev.2023.105127 Received 5 November 2022; Received in revised form 8 March 2023; Accepted 9 March 2023 Available online 13 March 2023 0149-7634/© 2023 Elsevier Ltd. All rights reserved. Y. Liu et al. Neuroscience and Biobehavioral Reviews 148 (2023) 105127 as the mesocorticolimbic dopamine reward pathway (Arias-Carrión et al., 2010; Navarri et al., 2022); 2) both addiction and neurological disease show post-mortem alterations in neurotransmitters (meta bolism) involved in error monitoring (e.g., Dopamine, Norepinephrine, Serotonin, GABA; Jocham and Ullsperger, 2009) and in the frontal cortex (Cadet et al., 2014; Goldstein et al., 2011; Gramage and Herra don, 2011; Kärkkäinen et al., 2021; Reinikainen et al., 1988); 3) the existence of bidirectional interactive relationships between addiction and neurological disease (for instance, a substantial proportion of TBI occurs in intoxicated individuals, and TBI can serve as a risk factor for alcohol use disorder; Weil et al., 2018). Notably, we did not aim to verify addiction and neurological disease are the same thing, or they are comparable, but to gauge the degree of brain disorder in terms of error processing in addiction by comparing it with a population with well-recognized brain disorders. disorder present a significant reduction of ERN amplitude. Only two of these reviews examined Pe, and found no apparent difference between the substance use group and the healthy controls (Luijten et al., 2014; Lutz et al., 2021). With a few exceptions (e.g., Lutz et al., 2021), in most of these meta-analyses, effect size dependency was not yet handled in accordance with recent methodological advances (Pustejovsky and Tipton, 2022). There are numerous sources of dependency, either be tween studies (e.g., studies from the same research group) or within the study (e.g., a study reported multiple outcomes). For studies that pro vided more than one effect size, researchers selected the value they predetermined (e.g., use data from certain electrode(s) for ERN scoring), or halved the sample size of the control group when there were two experimental groups and then treated the two effect sizes as if they were from different studies. Such practices may result in hampered modera tion analyses and compromised aggregated effect size estimation accu racy (Scammacca et al., 2014). Furthermore, only two of the four reviews reported results on Pe. ERN and Pe show more dissociation than associations (Di Gregorio et al., 2018; see review: Overbeek et al., 2005), and substance use may be accompanied by aberrant ERN but intact Pe, or vice versa. As for neurological diseases, there are three relevant systematic re views summarizing ERN and Pe findings (Lenzoni et al., 2022; Pezzetta et al., 2021; Pyasik et al., 2022). It was found that for most of the neurological patient groups, ERN amplitude was reduced, whereas Pe was unabated when compared to healthy controls (Pezzetta et al., 2021), suggesting distinct patterns for the processes underlying these two error-related potentials. However, the overall effect size of error-related mechanisms in neurological patients has not been quantitatively eval uated via a meta-analysis. 1.3. Error processing Error processing refers to the ability to detect errors and evaluate performance, which is considered a fundamental aspect of cognitive control that allows for flexible behavior modification to optimize future decisions (Desender et al., 2021; Ridderinkhof et al., 2004). This aspect is of particular relevance in the clinical context, as self-control failure (Krönke et al., 2018) and maladaptive behavior (i.e., post-error adap tation, Wessel, 2018) - which are salient characteristics of error pro cessing - have been previously reported in patients with addiction (Inzlicht et al., 2015; Kotabe and Hofmann, 2015; Luijten et al., 2014) as well as neurological disorders (Lenzoni et al., 2022; Pezzetta et al., 2021). Dysfunctional error monitoring may be observed at the behavioral level and as modulation of error-related brain responses or oscillatory activity. In normal conditions, error commission robustly elicits two event-related potentials (ERPs) components: an early component, known as error-related negativity (Ne or ERN), and a late component, termed error positivity (Pe). ERN is a frontocentral negativity occurring 50–150 ms after an incorrect response, signaling the commission of error (Coles et al., 2001; O’Connell et al., 2007). Source localization analysis indicated that the ERN originated from the posterior medial frontal cortex (pMFC) and anterior cingulate cortex (ACC) (Holroyd et al., 2004; Veen and Carter, 2002). A correct response can trigger a similar component, namely correct-response negativity (CRN). The difference wave (i.e., delta score, ERN − CRN) was also used in some studies as the index of performance monitoring (Gorka et al., 2019). Pe is a positive wave that can be recorded 200–500 ms after the error over centro-parietal regions (Falkenstein et al., 2000), and has been proposed to represent error awareness (Leuthold and Sommer, 1999; Overbeek et al., 2005) and motivational salience of an error (Falkenstein et al., 2000; Ridderinkhof et al., 2009). A recent study indicates that instead of signaling the consciousness of an error directly, Pe reflects the trajectory of evidence accumulation after a decision, and such accumulation ter minates when the agency realizes the mistake (Desender et al., 2021). Contrary to ERN, the source of Pe remains equivocal, with studies showing it arises from activity in the parietal cortex (O’Connell et al., 2007), rostral ACC (rACC; Herrmann et al., 2004; van Boxtel et al., 2005), anterior insula (Harsay et al., 2012), and prefrontal cortex (Masina et al., 2019). 1.5. The present study The neural basis of error processing is among the most adequately studied fields for both addiction and neurological disease. However, so far, no study has investigated and quantified whether and how these populations share a similar impairment in the functional mechanisms of error processing - in line with the brain disease model - or whether addiction and neurological diseases can be considered functionally distinct. This study aimed to shed new light on the debate of ‘addiction as a brain disease’ through intergroup comparisons of neural activities (i.e., ERN and Pe) underlying error processing via meta-analyses. Based on the overview above, we expect both groups to show similar patterns of error-related brain response deficits, but the degree of impairment is larger in neurological disorders. Thus, in the present unregistered metaanalysis, we examined the role of Group (addiction vs. neurological disease) as a focal moderator that may explain between-study heterogeneity. As a secondary focal moderator, we included the effect of Age. The relationship between ERN (or Pe) amplitude and age may be an inverted U-shaped curve, such that it peaks in young adults (Boen et al., 2022), and then declines due to the degradation of the mesocorticolimbic DA-system during normal aging (Beste et al., 2009). In addition, we examined several parity moderators, including medicine (Seer et al., 2017; Stemmer et al., 2007) and task (Pasion and Barbosa, 2019; Vallet et al., 2021) since discrepancies exist concerning their role in ERN and Pe. As to certain other potential parity moderators, robust evidence for their moderating effects is still lacking (e.g., ERN peaking score: delta vs. error-related; electrode analyzed; the percentage of males), and we therefore examined their roles in exploratory analyses. 1.4. Error processing and addiction/neurological disorders To date, four reviews/meta-analyses have summarized ERN and Pe findings in addiction research (systematic review: Luijten et al., 2014; meta-analysis: Lutz et al., 2021; Pasion and Barbosa, 2019; Zhang et al., 2021). Pasion and Barbosa (2019) and Zhang et al. (2021) found that ERN amplitude did not differ significantly between substance users and healthy controls. In contrast, Lutz et al. (2021) showed that the severity of addiction matters, such that only those diagnosed with substance use 2. Methods This review has been reported following the Preferred Reporting Items for Systematic Review and Meta-analysis (Page et al., 2021). 2 Y. Liu et al. Neuroscience and Biobehavioral Reviews 148 (2023) 105127 2.1. Literature search and screening For neurological disease, the updated literature search (from 9/4/2020–5/10/2021) resulted in 22 new articles. Out of them, only one was finally included (Niessen et al., 2020). Another 31 studies from Pezzetta et al. (2021) met our inclusion criteria and were also included. Literature searches were carried out separately for addiction and neurological disease. For addiction, 6 databases (Medline, PsycINFO, Embase, CINAHL, Web of Science, and Scopus) were searched until October 5th, 2021. Search terms and synonyms indicating substance use (e.g., alcohol, amphetamine, cocaine, cannabis, heroin, ketamine, methamphetamine, benzodiazepines) were combined with terms indicative of the ERP component (e.g., ERN, Pe). For neurological dis ease, an updated literature search was carried out based on RP’s recent review (Pezzetta et al., 2021). Search strategies for all databases are available in the Supplementary Materials. For addiction, the inclusion criteria were that the studies: (a) were presented in English and published in peer-reviewed journals; (b) were conducted on human participants; (c) assessed the amplitude and/or latency of ERN and/or Pe during error processing; (d) compared a group with current or past substance use (or a group diagnosed with the neurological disease) to a healthy control group, and each group had no less than 5 participants; (e) statistics reported in the paper or data provided by authors via contacts allowed effect size calculation. We excluded studies: (a) focused on the acute effect of substance use; (b) in which a pure substance use/neurological disease effect cannot be iso lated (e.g., comorbid disorders such as depression, anxiety, etc.). For neurological disease, in line with Pezzetta et al. (2021), the search included the following neurological disorders: “Parkinson’s disease”, “multiple sclerosis”, “spinal cord”, “stroke”, “brain damage”, “lesion/s”, “neurological”, “mild cognitive impairment”, “dementia”, “Alzheimer”, and “Huntington”. For addiction, a total of 174 articles were initially identified after deduplication. The title, abstract, and full text were double-blindly screened by YL and RP through Rayyan (Ouzzani et al., 2016). Con flicts were first addressed between the two raters, then the whole group for complicated ones. After title and abstract screening, there were 27 articles left, of which another 10 were excluded (see details in Fig. 1). 2.2. Quality assessment and data extraction The Appraisal tool for Cross-Sectional Studies (AXIS) was used to assess the quality of included studies (Downes et al., 2016). It evaluates the quality of a study in terms of study aim declaration, sample size justification, study design and measurements appropriateness, results reporting sufficiency, significance justification, inference from the re sults to the conclusions, limitations, and ethics. Items 13, 14, and 19 were deleted as they were inappropriate for the present aim. Each item was rated with ‘Yes’ or ‘No’. The overall quality of a study was repre sented by the percentage of ‘Yes’ answers. YL and FM assessed the quality independently, and the agreement was high (93.37%, 732/784). Information coded mainly includes those used to calculate the effect size and its variance, as well as moderators examined in the metaregression analyses. Some other important study characteristics were also coded and presented in Tables 1a; 1b. Each paper was blindly coded by two authors (YL, RP, or FM). 2.3. Meta-Analytic Procedures The analyses were conducted in the following steps: 1) effect size computation; 2) outlier detection and influence analysis; 3) overall ef fect size calculation; 4) heterogeneity test and moderator analysis; 5) small sample bias assessment. All analyses were performed separately for the four outcomes (ERN amplitude, Pe amplitude, ERN latency, and Pe latency) in R-4.1.0. 2.3.1. Effect size computation The standardized mean difference Hedges’ g was calculated since it Fig. 1. PRISMA flowchart. 3 Study Sample size The sample Education (gender, age: size for years (M±SD) mean±sd) ERN/Pe calculation 4 Measures PolyTreatment Treatment Abstinence substance duration duration use/usage (month) of other substance Medicine Task Reference Baseline/ Time Nr of Electrodes ms window/ electrode analyzed ms SUG: 8.4 ± 3.0; Heroin CON: 13.7 ± 4.6 12.4 ± 3.1 Clinical DSM-IV No Comparable, all Cocaine had lower or intermediate education levels NA Tobacco NA Clinical DSM-IV >5 cigarettes/ day for one year NA Subclinical Comparable, all Alcohol follow higher education Yes 8.2 ± 2.0 NA NA Flanker Average NA ERN: 0–100 4 also used Yes heroin NA 1–13 months NA Flanker Linked mastoids -200–0 FTND NA No NA No abstain NA Flanker Linked mastoids -200–0 Subclinical QFV NA No No No abstain NA Flanker Linked mastoids -100–0 Continuous Nose performance task ERN: 64 25–75; Pe: 200–400 ERN: 32 25–100; Pe: 200–400 ERN: 32 25–75; Pe: 200–400 ERN: 37 - 50–150; Pe: 100–450 ERN: Fz, Fcz, Cz ERN: Fz, Fcz, Cz; Pe: FCZ, Cz, Pz ERN & Pe: Fz, Cz, Pz; ERN: Fz, Fcz, Cz; Pe: Cz, CPz, Pz ERN: FCz; Pe: CPz SUG: 13.07 ± 1.28; CON: 13.97 ± 1.49 Cannabis 4.20 ± 3.71 Clinical DSM-IV No No No 2 days on average No NA Alcohol NA Clinical: DSM-5 AUD; Subclinical: AUD in remission and high risk Yes No No only for the remission group Some Flanker participants had Linked mastoids -500~ 300 ERN: 0–100 SUG: 14.92 ± 1.00; CON: 15.20 ± 1.32 Alcohol NA Subclinical AUDIT & NA AUQ No No No NA Flanker Average -100–0 ERN: 64 50–150; Pe: 150–400 ERN & Pe: Fz, FCZ, Cz & Pz Undergraduate students Alcohol NA Subclinical AUDIT & No AUQ No No No abstain No GNG Average -500–0 ERN: -50–80; Pe: 80–300 128 ERN & Pe: Cz, Fz, FCz Undergraduate students Tobacco No No No No abstain No Flanker Linked mastoids -200–0 Cocaine NA No No the first week No in detoxification treatment Flanker Linked mastoids -100–0 ERN: 34 25–75; Pe: 250–350 ERN: 32 25–100; ERN & Pe: FCz, Cz, CPz Comparable ≥ 10 cig/day Subclinical: FTND for two years medium level of dependence 12.2 ± 6.8 Clinical DSM-IV ERN: Fz, FCz, Cz NA (potential: NA(8 never NA get any treatment) 3.9 days on No average (1 day to 1 week) GNG Nasion NA ERN: 30–70; ERN: FCz; Pe: CPz SUG:12.5 ± 2.3; Cocaine CON: 12.2 ± 1.4 NA Clinical DSM-IV -200–0 32 64 ERN: Cz (sample 1) and 34 (sample 2) 168 (continued on next page) Neuroscience and Biobehavioral Reviews 148 (2023) 105127 Chen et al. (2013) 20SUG(M, 17SUG; 37.1 ± 9.5); 15CON 15CON(M, 32.5 ± 9.9) Franken et al. 14SUG(13 M, Same (2007) 38.1 ± 10.2); 13CON(8 M, 32.0 ± 13.8) Franken et al. 23SUG(NA, 21SUG (2010) 21.7 ± 2.7); (11 M); 28CON(NA, 25CON 21.3 ± 2.8) (11 M) Franken et al. 48SUG(25 M, 44SUG; (2017) 23.4 ± 10); 43CON 49CON(24 M, 22.9 ± 8.5) Fridberg et al. 30SUG(24 M, Same (2013) 20.20 ± 2.86), 32CON(20 M, 20.84 ± 2.95) Gorka et al. 39 current Same (2019) AUD(19 M, 23.3 ± 3.2); 60 remitted AUD(27 M, 23.2 ± 2.8); 43 at-risk for AUD(18 M, 22.4 ± 3.6); 53CON(21 M, 22.2 ± 2.5) Kim and Kim 25SUG(F, Same (2019) 21.44 ± 1.89); 25CON(F, 21.72 ± 2.44) Lannoy et al. 20SUG(8 M, Same (2017) 20.25 ± 1.62); 20CON(7 M, 21.20 ± 2.59) Luijten et al. 13SUG(9 M, Same (2011) 20.7 ± 1.3); 14CON(10 M, 21.4 ± 2.6) Marhe et al. 49SUG Same (2013) (44 M,39.6 ± 8.4); 23CON(17 M, 39.9 ± 9.4) Morie et al. (2014) 23SUG(16 M, Same 44.0 ± 6.6); Substance Duration of Severity substance use/yrs Y. Liu et al. Table 1a Characteristics of included studies (Addiction). Study Sample size The sample Education (gender, age: size for years (M±SD) mean±sd) ERN/Pe calculation 27CON(20 M, 41 ± 8.5) Padilla et al. 14SUG(M, (2011) 37.86 ± 9.3); 14CON(M, 43.5 ± 14.5) Rass et al. (2014) 22 daily smokers (13 M, 27.2 ± 5.3); 31 intermittent smokers (12 M, 23.9 ± 4.4); 30CON(14 M, 25.2 ± 4.3) Smith et al. (2015)-female 5 Smith et al. (2015)-male Smith et al. (2016)-female Smith et al. (2016)-male Smith et al. (2017)unaware Sokhadze et al. (2008) Medicine Task Reference Baseline/ Time Nr of Electrodes ms window/ electrode analyzed ms DSM-IV nicotine, alcohol) NA NA NA Yes (1–4month) No Flanker Linked mastoids Pe: 100–300 -200–100 ERN: 64 20–120 Same SUG: 13.5 ± 2.2; CON: 16.07 ± 3.5; Alcohol NA Flanker: 21 daily smokers; 26 intermittent smokers; 25 controls GNG: 21 daily smokers; 29 intermittent smokers; 26 controls Same Daily smokers:14.8 ± 1.6; Intermittent smokers: 15.8 ± 1.6; CON: 16.6 ± 2.0 Tobacco Daily Subclinical smokers: 9.4 ± 6.4; Intermittent smokers: 5.6 ± 4.4 FTND No No No No abstain No Flanker & GNG Nose -200–0 ERN: 34 -50–100; Pe: 100–250 ERN: FCz; Pe: Cz NA Alcohol Regular Subclinical drinking: 3.9 ± 2.8 AUDIT Cannabis No No No NA Flanker Linked mastoids -500–0 ERN: 0–150 58 ERN: Fz NA Alcohol Regular Subclinical drinking: 6.3 ± 3.1 AUDIT Cannabis No No No NA Flanker Linked mastoids -500–0 ERN: 0–150 58 ERN: Fz NA Alcohol NA Subclinical AUDIT No No No No No Stop task Linked mastoids -500–0 ERN: 0–250 58 ERN: Fz NA Alcohol NA Subclinical AUDIT No No No No No Stop task Linked mastoids -500–0 ERN: 0–250 58 ERN: Fz NA Alcohol NA Subclinical AUDIT No(only tobacco) No No No No Error awareness task Linked mastoids -200–0 NA Alcohol NA Subclinical AUDIT No(only tobacco) No No No No Error awareness task Linked mastoids -200–0 NA Cocaine NA Clinical DSM-IV Cannabis, tobacco No No < 2 months No GNG Average -700~ 500 ERN: 58 0–120; Pe: 300–500 ERN: 58 0–120; Pe: 300–500 ERN: 128 50–200 15SUG(F, 21.0 ± 2.3); 17CON(F, 21.4 ± 2.0) 16SUG(M, Same 23.0 ± 2.2); 18CON(M, 22.1 ± 2.4) 13SUG(F, Same 20.0 ± 1.2); 17CON(F, 20.1 ± 1.2) 21SUG(M, Same 19.8 ± 1.2); 20CON(M, 20.1 ± 1.1) 25SUG(12 M, Same 22.2 ± 2.5); 35CON(18 M, 21.8 ± 2.2) 16SUG(8 M, Same 22.0 ± 2.8); 13CON(6 M, 21.0 ± 1.6) 19 SUG(12 M, 6SUG(4 M, 42.1 ± 5.5); 42.11 15 CON(7 M, ± 5.60); 37 ± 9.4) 6CON(3 M, 39.07 ± 9.48) Clinical Measures PolyTreatment Treatment Abstinence substance duration duration use/usage (month) of other substance ERN: FCz ERN: FCz; Pe: CPz ERN: FCz; Pe: CPz ERN: Fz, Afz, F1, F2 Note. NA: not available; Gender: M: all are males, F: all are females, X: mixed (the specific number for males and females was coded if it was reported, e.g., 17 M); SUG: substance use group. CON: control group; AUD: alcohol use disorder; DSM: Diagnostic and Statistical Manual of Mental Disorders; AUDIT: Alcohol Use Disorder Identification Test; FTND: Fagerstrom Test for Nicotine Dependence; QFV: Quantity-Frequency-Variability index; AUQ: Alcohol Use Questionnaire Neuroscience and Biobehavioral Reviews 148 (2023) 105127 Smith et al. (2017)-aware Substance Duration of Severity substance use/yrs Y. Liu et al. Table 1a (continued ) Y. Liu et al. Neuroscience and Biobehavioral Reviews 148 (2023) 105127 corrects for small study bias (Hedges, 1981). Hedges’ g and its variance were calculated from M(SD) of the experimental group (addiction/ neurological disease) and the control group by the escalc() function (‘metafor’ package, Viechtbauer, 2010). When M and/or SD were un available, the corresponding t-test or F-test result was used, and Hedges’ g was calculated by the ‘esc’ package (Lüdecke and Lüdecke, 2019). For ERN amplitude, a positive Hedge’s g indicates lower amplitude (i.e., less negative values) in the experimental group, since its polarity is negative; for Pe amplitude, ERN latency, and Pe latency, a positive Hedge’s g in dicates enlarged amplitude/delayed latency in the experimental group. 2.3.5. Small sample bias assessment To examine the small sample effect (i.e., small studies can only be published with very high effect sizes), we first visualized the symmetry of the contour-enhanced funnel plot, then used a modified form of Egger’s test (i.e., used RVE with CHE working model to account for ef fect size dependency) to examine the symmetry empirically (Atit et al., 2022). 3. Results 3.1. ERN amplitude 2.3.2. Outlier detection and influence analysis Studentized residuals and Cook’s distance were used to detect out liers and influential cases (Berres and Erdfelder, 2021). Since these methods were not available for models with Robust Variance Estimation (RVE), they were applied to a randomized multilevel model without RVE. Effect size with a z score outside of ± 1.96 was deemed an outlier, and when its Cook’s distance value was larger than 4/N (N being the number of effect sizes), it was considered an influential case. One influential case was detected (Beste et al., 2017), and the following analyses were performed without it. There were 103 (k) effect sizes from 46 (m) studies included. The overall mean weighted effect size was significant (g = 0.471, 95% CI [0.288, 0.655], t(42.92) = 5.19, p < 0.001), implying greater negativity in the experimental group compared to healthy controls across groups (see the forest plot in Fig. 2). Considerable heterogeneity was observed in the results obtained cross studies (Q(102) = 332.01, p < 0.001; τ2 = 0.31; σ2b = 0.20; σ2w = 0.11; I2t = 72.52%; I2b = 47.34%; I2w = 25.18%).2 A sensitivity analysis with a correlation level of within-stfudy outcomes set to the range 0–1 showed no differences in effect size or SE. Beyond the overall significant mean effect size, our main concern was to examine the difference between addiction and neurological dis ease. We found a marginally significant difference between these two groups, F(1, 34.7) = 3.97, p = 0.054. The effect size was larger in the neurological disease group (g = 0.61, t(26.93) = 4.63, p < 0.001) than in the addiction group (g = 0.27, t(15.27) = 2.53, p < 0.001), though both were significant. The results for all moderating effects examined were summarized in Table 2. The effect of ‘Specific category’ was also significant, F(6, 3.84) = 7.7, p = 0.04, and the post-hoc tests revealed that the effect size of cocaine was larger than that of alcohol (β = 0.64, p < 0.05), TBI (β = 0.64, p < 0.05), and cerebellar lesion (β = 0.76, p < 0.05); the effect size of Parkinson disease was larger than that of alcohol (β = 0.74, p < 0.001), tobacco (β = 0.60, p < 0.05), TBI (β = 0.76, p < 0.01), and cerebellar lesion (β = 0.87, p < 0.05); and the effect size of prefrontal lesion was larger than that of alcohol (β = 0.76, p < 0.05) and TBI (β = 0.78, p < 0.05). However, these post-hoc test results were inconclusive since at least one party of the paired com parison was with less than 4 degrees of freedom, indicating that too few studies were available to allow firm conclusions. Another significant moderator was Age (F(2, 20.6) = 4.42, p < 0.05). The effect size was significantly larger for older than for young adults (β = 0.54, p < 0.01), while the middle-aged group did not differ significantly from the other two groups (young adults: β = − 0.20, p = 0.30; older: β = 0.34, p = 0.15). Other moderators examined did not significantly contribute to the heterogeneity of effect sizes, whereas certain levels of a moderator may differ from each other (see details in Table 2). 2.3.3. Overall effect size calculation Many studies contributed more than one effect size, including results from different electrodes, different tasks, multiple experimental groups, medicine on/off, etc. To account for effect size dependency, the corre lated and hierarchical effects (CHE) model considers both correlated (e. g., effect sizes from the same study) and hierarchical (e.g., effect sizes from the same lab across studies) effects using the RVE methods (Pus tejovsky and Tipton, 2022) was built by the ‘metafor’ (Viechtbauer, 2010) and ‘clubSandwich’ packages (Pustejovsky, 2022). The correla tion of within-study effect sizes was set to 0.6, and the robustness of the results was examined through sensitivity tests by varying the coefficient from 0 to 1. 2.3.4. Heterogeneity test and moderator analysis We estimated the heterogeneity using the Q test, τ (SD of the dis tribution of true effects), and I2 (proportion of true heterogeneity to the total variance of observed effects). In general, a significant Q test and an I2 higher than 75% indicate high heterogeneity between study outcomes (Higgins and Thompson, 2002). Meta-regression analyses were conducted to examine the source of heterogeneity in research findings. Categorical moderators analyzed included Group (addiction or neurological disease), Task (flanker, go/ no-go, or others), Medicine (on, off, or washout), ERN/Pe definition (response-locked or delta), Electrode (ERN: FCz, Fz, Cz, or others; Pe: Cz, CPz, Pz, or others). For the effect of Group, to get more detailed infor mation about the difference, we reclassified it into two other modera tors: Broad category (stimulants, depressants, acquired brain injury, basal ganglia disorders, white matter diseases, etc.) and Specific cate gory (alcohol, tobacco, cannabis, cocaine, Parkinson, TBI, cerebellar lesion, prefrontal lesions, etc.). Three addiction-unique moderators (Severity: clinical vs. subclinical; Abstain status: yes vs. no; Treatment: yes vs. no) were examined within addiction-related studies only. Two continuous variables were transformed into categorical variables since they violated the assumption of normal distributions. They were Age (young: ≤39 yrs, middle-aged: 40–59 yrs, older: over 59 yrs) and the Percentage of males (majority are males vs. females). Since this is not individual-level meta-analyses, the mean age for each group (experi mental and control) from all included papers were used. The function of Wald_test() from ‘clubSandwich’ was used to test differences across levels of moderators with small-sample corrections, which is similar to F-test but with denominator degrees of freedom that can take noninteger values (i.e., approximate Hotelling’s T2 [AHT] F tests; Puste jovsky and Tipton, 2018). Because fewer than four effect sizes cause problems for small-sample corrected F tests in terms of power (Tan ner-Smith et al., 2016; Tipton and Pustejovsky, 2015), we excluded all moderator levels with less than four effect sizes. 3.2. Pe amplitude Three outliers were detected (Ito and Kitagawa, 2006, 2005; Solbakk et al., 2014), and the analyses included 47 effect sizes from 23 studies. The overall mean weighted effect size was not significant (g = − 0.14, 95% CI [− 0.297, 0.007], t(18.15) = − 2.00, p = 0.06) (see the forest plot in Fig. 3). Moderate heterogeneity was observed in the results obtained Q = Cochrane’s Q test for effect size heterogeneity; τ2 = total variance; σ2b = variance between samples; σ2w = variance within samples; I2t = total amount of heterogeneity relative to the total amount of variance; I2b = amount of het erogeneity between samples relative to the total amount of variance; I2w = amount of heterogeneity within samples relative to the total amount of variance. 2 6 Study Sample size (gender, age: mean±sd) The sample size for ERN/ Pe calculation Education years (M ±SD) Neurological disorder Diagnostic criteria Brain lesions Task Reference Baseline/ ms Time window/ ms Nr of electrode Electrodes analyzed Beste et al. (2006) 11NDG(NA, 39.81 ± 8.96); 12CON(NA, 38.12 ± 7.56) 15NDG (8 M,37.1 ± 7.4); 15 young CON (7 M,34.5 ± 5.5) 17NDG off med (9 M,66.8 ± 8.5); 17 old CON(9 M,65.2 ± 7.2) 17NDG de novo(7 M,59.6 ± 10.4); 17 old CON(9 M, 65.2 ± 7.2) 21NDG(M, 47.6, SD=NA); 21CON(M, 47.1, SD=NA) 15NDG (8 M,60.1, SD=NA); 15CON (8 M,60.2, SD=NA) Same NA Huntington’s disease Genetic Basal ganglia Flanker Linked mastoids -200–0 ERN: NA; Pe: 200–500 32 ERN: Fz, Fcz, Cz; Pe: Pz Same comparable (mean years NA) Huntington’s disease Genetic Basal ganglia Flanker Linked mastoids -200–0 ERN: NA 32 ERN: Fz, FCz, Cz Same Parkinson’s disease Clinical (Unified Parkinson’s Disease Rating Scale) Same Parkinson’s disease Clinical (Unified Parkinson’s Disease Rating Scale) Beste et al. (2009)-a Beste et al. (2009)-b Beste et al. (2009)-c 7 Beste et al. (2017) Falkenstein et al. (2001a) 15NDG (8 M,60.1, SD=NA); 15CON (8 M,60.2, SD=NA) Holroyd et al. (2002) 9NDG(M,56.1 ± 4.6); 9CON (M, 57.3 ± 5.9) Ito and Kitagawa (2006) 17NDG (8 M,64.1, SD=NA); 15CON (7 M,63.8, SD=NA) NA Parkinson’s disease Genetic (X-linked dystonia parkinsonism) Basal ganglia Flanker Free/CSD -800~ 600 ERN: 50–100 28 ERN: FCz, Cz Task 1: 13NDG, 13CON; Task 2: 13NDG, 13CON; Task 3: 14NDG, 14CON Task 1: 13NDG, 13CON; Task 2: 13NDG, 13CON; Task 3: 14NDG, 14CON Same comparable (mean years NA) Parkinson’s disease Clinical (Unified Parkinson’s Disease Rating Scale) Basal ganglia Task 1: Flanker; Task 2: Simon; Task 3: GNG Average -200–0 ERN: -20–180 24 ERN: FCz comparable (mean years NA) Parkinson’s disease Clinical (Unified Parkinson’s Disease Rating Scale) Basal ganglia Task 1: Flanker; Task 2: Simon; Task 3: GNG Average -200–0 Pe: 250–550 24 Pe: Pz NA Parkinson’s disease Basal ganglia Flanker Linked mastoids NA ERN: 0–200 20 ERN: Cz, Fz comparable (mean years NA) Parkinson’s disease Clinical (Unified Parkinson’s Disease Rating Scale and the Hoehn and Yahr rating scale) Clinical (Hoehn and Yahr rating scale) Basal ganglia Lexical decision task Linked mastoids -200–0 ERN: 10–200; Pe: NA 6 ERN: Fz; Flanker Average -200–0 ERN: NA 30 Same Same Basal ganglia Pe: Pz ERN: FCz (continued on next page) Neuroscience and Biobehavioral Reviews 148 (2023) 105127 Falkenstein et al. (2005) Same Y. Liu et al. Table 1b Characteristics of included studies (Neurological disease). Sample size (gender, age: mean±sd) Seer et al. (2017) 13NDG (10 M,64.31 ± 8.6); 13CON (10 M, 63.15 ± 11.15) 9NDG nonmedicated (4 M,64.2, SD=NA); 14CON (5 M,65.6, SD=NA) 9NDG medicated (6 M, 63.4, SD=NA); 14CON (5 M,65.6, SD=NA) 12NDG (9 M,65, SD=NA); 12CON (4 M,68, SD=NA) 10NDG (7 M,56.3 ± 3.24); 10CON(4 M, 57.9 ± 7.5) 20NDG (12 M,64.5 ± 9.7); 20CON (12 M, 64.3 ± 8.9) 14NDG de novo(7 M,58.9 ± 10.4); 14CON(7 M, 59 ± 11.0) 16NDG (7 M,65.4, SD=NA); 15CON (7 M;63.8, SD=NA) 12NDG (4 M,76.2 ± 5.7); 10CON (4 M, 75.3 ± 5.1) Stemmer et al. (2007)-a Stemmer et al. (2007)-b Verleger et al. (2013) 8 Volpato et al. (2016) Willemssen et al. (2008) Willemssen et al. (2009) Ito and Kitagawa (2005) Mathalon et al. (2003) The sample size for ERN/ Pe calculation Education years (M ±SD) Neurological disorder Diagnostic criteria Brain lesions Task Reference Baseline/ ms Time window/ ms Nr of electrode Electrodes analyzed NDG(13.04 ± 3.82); CON (13.65 ± 3.24) Parkinson’s disease Clinical (experienced attending neurologists) Same NDG nonmedicated (14.0, SD=NA); CON (14.5, SD=NA) Parkinson’s disease Clinical (Unified Parkinson’s Disease Rating Scale and the Hoehn and Yahr rating scale) Basal ganglia Flanker Linked mastoids -400~ 200 ERN: 0–150 64 ERN: Fz, FCz, Cz, CPz, Pz Same NDG medicated (14.3, SD=NA); CON (14.5, SD=NA) Parkinson’s disease Clinical (Unified Parkinson’s Disease Rating Scale and the Hoehn and Yahr rating scale) Basal ganglia Flanker Linked mastoids -400~ 200 ERN: 0–150 64 ERN: Fz, FCz, Cz, CPz, Pz 9NDG; 9CON NA Parkinson’s disease Basal ganglia Flanker Nose-tip NA ERN: 0–150 20 ERN: Fz, Cz, Pz Same NDG(12.0 ± 4.24); CON (13.1 ± 4.14) Parkinson’s disease Basal ganglia Reinforcement learning task Fpz -800~ 700 ERN: 0–150 29 ERN: Cz 18NDG(66.3 ± 8.3); 18CON (66.0 ± 7.3) comparable educational background (mean years NA) comparable (mean years NA) Parkinson’s disease Clinical (Unified Parkinson’s Disease Rating Scale and the Hoehn and Yahr rating scale) Clinical (United Kingdom PD Society brain bank diagnostic criteria) Clinical (Unified Parkinson’s Disease Rating Scale) Basal ganglia Flanker Average NA ERN: 20–120 26 ERN: FCz Parkinson’s disease Clinical (experienced attending neurologists) Basal ganglia Flanker Average NA ERN: 20–120 26 ERN: FCz comparable (mean years NA) Alzheimer’s disease Clinical (NINCDSADRDA Alzheimer’s Criteria) Neurodegenerative disorders Lexical recognition paradigm Linked earlobes -200–0 ERN: 10–150; Pe: NA 5 ERN: Fz; NDG(15.9 ± 1.8); CON (16.7 ± 1.6) Alzheimer’s disease Clinical (NINCDSADRDA Alzheimer’s criteria) Neurodegenerative disorders Picture-name verification task NA -50–0 Pe: 200–500 NA Flanker Average ERN: 0–130 256 Same 12NDG (6 M,61.1, SD=NA); 15CON (7 M,63.8, SD=NA) Same Same Pe: Cz ERN: Fz, Cz; Pe: Pz; (continued on next page) Neuroscience and Biobehavioral Reviews 148 (2023) 105127 Study Y. Liu et al. Table 1b (continued ) Study Sample size (gender, age: mean±sd) Thurm et al. (2013) 14NDG (5 M,70.6 ± 8.2); 16CON (7 M, 68.0 ± 3.4) 25NDG (20 M,21.0 ± 2.3); 22CON (13 M, 20.5 ± 2.0) 20NDG (16 M,30.9 ± 12.0); 20CON(9 M, 26.1 ± 9.9) Olson et al. (2018) Larson et al. (2009) The sample size for ERN/ Pe calculation 9 Education years (M ±SD) Neurological disorder Diagnostic criteria Brain lesions Task Reference Baseline/ ms NDG(13.8 ± 2.3); CON (14.8 ± 2.7) Mild Cognitive Impairment Clinical (MCI criteria) Neurodegenerative disorders 25NDG; 20CON NDG(14.0 ± 1.7); CON (13.8 ± 1.3) TBI Clinical (interview) Diffused Flanker Linked mastoids -400~ 200 ERN:0–100; Pe:200–400 64 ERN: Fz, FCz; Pe: Cz, Pz Same NDG(13.3 ± 1.7); CON (14.2 ± 1.4) TBI Clinical (Glasgow Coma Scale) and structural (computerized tomography) Diffused Stroop task Average -200–0 ERN: 0–100; 64 EGI system positions, ERN: 4(FCz), 65(Cz), 5 and 55; Pe: 65(Cz), 18, 43, 30, 34(Pz) EGI system positions, ERN: 7, 31, 55, 80, 106; Pe: 54, 55, 61, 62 [Pz], 78, 79 ERN: FCz; Same NDG(14.3 ± 1.2); CON (14.1 ± 1.5) TBI Clinical (interview) Diffused Stroop task Average NA Pontifex et al. (2009) 30NDG (23 M,19.9 ± 1.2); 36CON (21 M, 19.4 ± 1.4) 6NDG(4 M,69, SD=NA); 10 older CON (4 M,70, SD=NA) Same NDG(14.2 ± 1.1); CON (13.6 ± 1.4) TBI Diffused Flanker Linked mastoids -100–0 Same NA Acquired brain injury NDG(11.0 ± 4.66); CON (13.7 ± 3.44) Acquired brain injury NDG(10.7 ± 5.37); CON (13.7 ± 3.44) Acquired brain injury NDG(13.1 ± 2.5); CON (13.1 ± 2.5) Tumor or TBI Clinical (American Academy of Neurology injury definition) Structural (computerized tomography or magnetic resonance imaging) Structural (computerized tomography or magnetic resonance imaging) Structural (computerized tomography or magnetic resonance imaging) Structural (magnetic resonance imaging) 7NDG (7 M,54.7 ± 10.05); 7CON(7 M, 48.0 ± 8.81) Same Maier et al. (2015)-b 7NDG(M,57.6 ± 13.33); 7CON(M, 48.0 ± 8.81) Same Solbakk et al. (2014) 12NDG (6 M,48.0 ± 6.7); 14CON (9 M, 41.1 ± 12.4) Same Electrodes analyzed ERN: Fz, FCz, Cz, PCz; Pe: 200–400 ERN: 0–200; Pe: 200–400 128 ERN: 0–200; Pe: 200–500 64 Pe: Pz PFC lesion Letterdiscrimination task Linked mastoids -100–0 ERN: 0–50; 19 ERN: Cz PFC lesion Flanker Linked mastoids -150~ -50 ERN: -10–90; Pe: 200–400 27 ERN: FCz; Linked mastoids -150~ -50 ERN: -10–90; Pe: 200–400 27 Average -300–0 ERN: 60–140; 128 Brain damage lesion (non-frontal lesion) Flanker OFC lesion Stop task Pe: Pz ERN: FCz; Pe: Pz 128 channel Geodesic Sensor Net, ERN: 7, 106, 31, 80; (continued on next page) Neuroscience and Biobehavioral Reviews 148 (2023) 105127 36NDG (18 M,21.6 ± 2.4); 46CON (22 M, 20.7 ± 2.2) Maier et al. (2015)-a Nr of electrode -200~ 100 Larson et al. (2012) Gehring and Knight (2000) Time window/ ms Y. Liu et al. Table 1b (continued ) Study Ullsperger et al. (2002)-a Sample size (gender, age: mean±sd) The sample size for ERN/ Pe calculation Education years (M ±SD) Neurological disorder Diagnostic criteria Brain lesions Task Reference Baseline/ ms Time window/ ms Pe: 200–400 ERN: NA; Pe: 300–450 Nr of electrode 7NDG (6 M,50.7 ± 11.3), 9 older CON (4 M, 51.1 ± 8.5) 6NDG (5 M,38.8 ± 9.5); 9 younger CON (7 M, 38.4 ± 8.9) 7NDG (5 M,54.6, SD=NA); 7CON (5 M,54.7, SD=NA) Same NA Acquired brain injury Structural (magnetic resonance imaging) Unilateral frontal lesion Flanker Linked mastoids -100–0 Same NA Acquired brain injury Structural (magnetic resonance imaging) Bilateral OFC lesion Flanker Linked mastoids -100–0 ERN: NA; Pe: 300–450 29 Same NDG(10.9, SD=NA); CON (11,SD=NA) Acquired brain injury Structural (magnetic resonance imaging) PFC lesion Flanker Linked mastoids NA ERN: 0–120 28 Ullsperger and von Cramon (2006)-b 9NDG (8 M,49.1, SD=NA); 9CON (8 M,49.8, SD=NA) Same NDG(11.3, SD=NA); CON (11.6, SD=NA) Acquired brain injury Structural (magnetic resonance imaging) Basal ganglia Flanker Linked mastoids NA ERN: 0–120 28 Hogan et al. (2006) NDG(NA, 18, SD=NA); 11CON(NA, 17,SD=NA) 27NDG (11 M,34.5 ± 7.5); 31CON (12 M, 37.5 ± 8.9) 24NDG(20 M, 56.4 ± 12.5); 32CON(20 M, 56.4 ± 10.1) Same NA White matter diseases Structural (magnetic resonance imaging) Clinical (modified McDonald’s criteria) White matter lesion Choice-response task Linked mastoids -100–0 21 White matter lesion Stop task Linked mastoids -50–0 ERN: 0–200; Pe: 200–500 ERN: 0–100 Structural (magnetic resonance imaging) Left hemisphere GNG Free/CSD -100–0 Ullsperger et al. (2002)-b Ullsperger and von Cramon (2006)-a 10 Niessen et al. (2020) Peterburs et al. (2011) Peterburs et al. (2012) 6NDG (2 M,54.3 ± 12.5); 28CON(12 M, 47.0 ± 12.0) 8NDG (4 M,42.9 ± 8.9); 22CON Same NDG(14.4 ± 2.8); CON (14.9 ± 3.0) Multiple sclerosis 17NDG (13 M,53.6 ± 12.2); 24CON(15 M, 56.0 ± 9.4) Same NA Stroke NA Stroke Structural (magnetic resonance imaging) Thalamus Antisaccade task Linked mastoids Same NA Stroke Structural (magnetic Cerebellum Antisaccade task Linked mastoids Pe: 7, 106, 31, 80 ERN & Pe amplitude: overall(F3, Fz, F4, FC3, FCz, FC4, C3, Cz, C4, P3, Pz, P4) ERN latency: FCz ERN amplitude: overall(F3, FCz, F4, C3, Cz, C4, P3, Pz, P4); ERN latency: FCz ERN amplitude: overall(F3, FCz, F4, C3, Cz, C4, P3, Pz, P4); ERN latency: FCz ERN: FCz; Pe: NA 19 ERN: Fz, Cz ERN: 0–150; Pe: 150–300 64 ERN: FCz; -100–0 ERN: 0–160; 30 ERN: FCz -100–0 or -200~ 100 ERN: 0–160; 30 ERN: FCz; Pe: Cz Pe: CPz (continued on next page) Neuroscience and Biobehavioral Reviews 148 (2023) 105127 López-Góngora et al. (2015) 29 Electrodes analyzed Y. Liu et al. Table 1b (continued ) Y. Liu et al. Seifert et al. (2011) cross studies (Q(46) = 99.24, p < 0.001; τ2 = 0.07; σ2b < 0.01; σ2w = 0.07; I2t = 42.46%; I2b < 0.01%; I2w = 42.46%). This indicated that the heterogeneity mainly derived from within-study variance. Therefore, we only analyzed the three moderators (i.e., Electrode, Task, and Pe defi nition) with levels varied within studies and the effect of Group as of our interest. It revealed that, most importantly, the effect of Group was not significant (F(1, 18.1) = 0.003, p = 0.96), indicating that effect sizes were negligible in both addiction and neurological disease groups; nor were the effects of Electrode (F(3, 8.62) = 0.604, p = 0.63), Task (F(2, 7.44) = 1.31, p = 0.33), and Pe definition (F(1, 5.76) = 0.19, p = 0.68, Table 3). Note. NA: not available; Gender: M: all are males, F: all are females, X: mixed (the specific number for males and females was coded if it was reported, e.g., 17 M); NDG: neurological disorder group. CON: control group; GNG: go/no-go task; TBI: Traumatic brain injury ERN & Pe: F3, FCz, F4, C3, Cz, C4, P3, Pz, P4 64 ERN: 0–140; Pe: 300–500 Linked mastoids Flanker NDG(10.8, SD=NA); CON (10.9, SD=NA) Same Acquired brain injury Structural (magnetic resonance imaging) Thalamus Linked mastoids Brain lesions (cerebellar lesions) NA Peterburs et al. (2015) (11 M, 43.1 ± 2.1) 16NDG (9 M,56.3 ± 9.7); 16CON (7 M, 56.1 ± 8.2) 15NDG (10 M;44.7 ± 13.07); 15CON(10 M; 44.4, SD=NA) 16NDG; 15CON Acquired brain injury resonance imaging) Clinical (experienced attending neurologist) Antisaccade task NA Pe: Pz ERN: Fz; 28 Pe: 150–450 ERN: 0–120; Pe: 200–400 -200~ 100 Nr of electrode Reference Brain lesions Education years (M ±SD) The sample size for ERN/ Pe calculation Sample size (gender, age: mean±sd) Study Table 1b (continued ) Neurological disorder Diagnostic criteria Task Baseline/ ms Time window/ ms Electrodes analyzed Neuroscience and Biobehavioral Reviews 148 (2023) 105127 3.3. ERN latency Two outliers were detected (Ito and Kitagawa, 2005; Larson et al., 2012), and the analyses were performed with the remaining 37 effect sizes from 19 studies. The overall mean weighted effect size was not significant (g = − 0.13, 95% CI [− 0.297, 0.04], t(15.26) = − 1.63, p = 0.12) (see the forest plot in Fig. 4). Small heterogeneity was observed in the results obtained cross studies (Q(36) = 49.36, p = 0.07; τ2 = 0.03; σ2b < 0.01; σ2w = 0.03; I2t = 19.34%; I2b < 0.01%; I2w = 19.34%). Therefore, we only analyzed the effect of Group as of our interest. It showed that the effect size of the neurological disease group (g = − 0.30, t(8.37) = − 3.95, p < 0.01) was more negative than that of the addiction group (g = − 0.01, t(6.96) = − 0.10, p = 0.93, F(1, 15.2) = 4.96, p = 0.04). 3.4. Pe latency One outlier was detected (Ito and Kitagawa, 2005), and the analyses were performed with the remaining 20 effect sizes from 10 studies. The overall mean weighted effect size was not significant (g = − 0.16, 95% CI [− 0.372, 0.057], t(7.37) = − 1.72, p = 0.13) (see the forest plot in Fig. 5). Small heterogeneity was observed in the results obtained cross studies (Q(19) = 27.73, p = 0.09; τ2 = 0.01; σ2b < 0.01; σ2w = 0.01; I2t = 9.95%; I2b < 0.01%; I2w = 9.95%). Therefore, we only analyzed the effect of Group, and it was not significant (F(1, 6.85) = 0.179, p = 0.69), indicating that effect sizes were negligible in both addiction and neurological disease groups. 3.5. Small sample bias assessment For ERN amplitude, there was a significant relationship between effect size estimate and precision, β = 2.92, SE = 0.77, t(18.9) = 3.78, p < 0.01, indicating possible small sample bias (see funnel plot in Fig. 6). However, the Egger’s test for neurological disease (β = 2.78, SE = 1.47, t(9.8) = 1.90, p = 0.09) and addiction (β = 1.99, SE = 1.41, t (6.14) = 1.41, p = 0.21) separately did not show publication bias. For Pe amplitude (β = − 1.38, SE = 1.12, t(9.94) = − 1.23, p = 0.25), ERN latency (β = − 0.28, SE = 1.19, t(7.93) = − 0.24, p = 0.83), and Pe la tency (β = 1.16, SE = 2.12, t(3.50) = 0.55, p = 0.62), the Egger’s test were not significant and their funnel plots did not indicate severe asymmetry (Fig. S1-S3), which implied no small sample bias. 3.6. Quality assessment The 49 included studies were of moderate to high quality, with a mean score of 83.98% (68.75% ~100%), according to the AXIS (Table 4). The main limitations were the small sample size and the un clear methods used to determine the sample size. For a few studies, in formation about participants’ recruitment channels, descriptive statistics, and limitations was missing. 4. Discussion The present meta-analysis examined whether addiction and 11 Y. Liu et al. Neuroscience and Biobehavioral Reviews 148 (2023) 105127 Fig. 2. Forest plot of ERN amplitude. The pooled effect size in this figure was calculated by making use of aggregated effect sizes per study, and therefore differs from the values reported in the main text when effect size dependency was controlled by RVE-CHE. This also applies to the other three forest plots. neurological disorders were associated with compromised error pro cessing (reflected by ERN amplitude, ERN latency, Pe amplitude, and Pe latency). Given the considerable degrees of heterogeneity, the effect of Age, Task, Medicine, Electrode, ERN/Pe definition, and Percentage of males were examined. We found that: 1) both addiction and neurolog ical disorders were associated with compromised ERN amplitude, and the impairment was marginally larger in the latter sample; 2) Age was a significant moderator of ERN amplitude, with older adults showing more severe impairment than young adults; 3) neurological disorders presented shorter ERN latencies than addiction when compared with controls; 4) for Pe amplitude and Pe latency, all examined moderators including Group were not significant. In the following sections, we start discussing the main findings and the explanation, followed by implica tions, limitations, and suggestions for future studies. 4.1. Summary of results Out of the four examined ERP indexes, ERN amplitude represented the most robust evidence and the richest results. We found that a decrement accompanied both addiction and neurological disorders in ERN amplitude, and such impairment was considerably larger in neurological disorders, although the group difference attained only marginal significance. The results for addiction were in line with some previous meta-analyses (Lutz et al., 2021; Pasion and Barbosa, 2019) but 12 Y. Liu et al. Neuroscience and Biobehavioral Reviews 148 (2023) 105127 Table 2 Moderator analysis of the overall effect size of ERN amplitude. Moderator m k F Group Addiction Neurological disease Specific category Alcohol Cocaine Tobacco Parkinson TBI Cerebellar lesions Prefrontal lesions Broad category Acquired brain injury Basal Ganglia Disorders Depressants Stimulants White matter diseases Age Young Middle-aged Old Task Flanker GNG Others Medicine On Off Overnight washout ERN definition Delta Response locked Electrode Cz FCz Fz Others Percentage male Female major Male major Addiction: severity Clinical Subclinical Addiction: abstain No Yes 46 17 29 36 8 4 3 11 4 2 4 43 14 11 9 7 2 46 21 16 10 46 27 6 16 22 7 14 5 46 13 39 46 15 23 15 17 44 28 20 17 9 10 17 13 5 103 40 63 83 20 8 10 30 6 4 5 99 22 33 21 18 5 103 49 28 26 103 66 9 28 56 15 33 8 103 20 83 103 21 35 21 26 99 49 50 40 14 26 40 31 9 F(1, 34.7) = 3.97 F(6, 3.82) = 7.76 F(4, 5.31) = 2.38 F(2, 20.6) = 4.42 F(2, 7.8) = 0.522 F(2, 2.48) = 0.102 F(1, 11.2) = 1.51 F(3, 12.1) = 0.295 F(1, 17.3) = 0.0128 F(1, 6.5) = 2.57 F(1, 4.72) = 0.159 g SE t df 0.271 0.610 0.107 0.132 2.532 4.629 15.27 26.93 0.110 0.746 0.249 0.853 0.092 -0.018 0.867 0.140 0.083 0.179 0.134 0.273 0.036 0.305 0.785 8.964 1.387 6.364 0.336 -0.503 2.846 6.19 2.31 1.91 9.39 2.77 1.00 2.97 0.344 0.891 0.136 0.516 -0.076 0.165 0.147 0.140 0.146 0.861 2.091 6.052 0.971 3.543 -0.088 12.04 9.70 7.47 5.41 1.00 0.299 0.495 0.837 0.130 0.147 0.117 2.299 3.356 7.148 19.25 14.05 8.39 0.517 0.310 0.441 0.112 0.175 0.162 4.609 1.768 2.719 25.34 4.13 14.37 0.572 0.637 0.532 0.197 0.173 0.170 2.899 3.676 3.135 6.54 15.30 4.95 0.673 0.421 0.180 0.108 3.744 3.885 12.98 39.14 0.458 0.481 0.379 0.524 0.128 0.116 0.127 0.178 3.586 4.145 2.975 2.935 17.02 24.00 17.71 14.91 0.430 0.449 0.146 0.107 2.951 4.191 18.46 26.08 0.414 0.162 0.151 0.111 2.752 1.462 7.24 9.08 0.239 0.335 0.098 0.240 2.450 1.395 10.97 3.61 p 0.054 0.023 0.000 0.037 0.461 0.008 0.305 0.000 0.761 0.703 0.066 0.177 0.058 0.000 0.362 0.014 0.944 0.025 0.033 0.005 0.000 0.612 0.000 0.150 0.016 0.907 0.025 0.002 0.026 0.244 0.002 0.000 0.829 0.002 0.000 0.008 0.010 0.911 0.008 0.000 0.156 0.028 0.178 0.707 0.032 0.243 CI low CI high 0.043 0.339 0.500 0.880 -0.230 0.430 -0.558 0.551 -0.820 -0.478 -0.108 0.449 1.062 1.055 1.154 1.003 0.442 1.842 -0.014 0.561 -0.191 0.150 -11.020 0.703 1.220 0.464 0.883 10.868 0.027 0.179 0.569 0.570 0.811 1.104 0.286 -0.171 0.094 0.748 0.790 0.789 0.099 0.268 0.095 1.045 1.006 0.970 0.285 0.202 1.061 0.640 0.189 0.241 0.111 0.143 0.728 0.720 0.647 0.904 0.124 0.229 0.735 0.670 0.061 -0.088 0.768 0.412 0.024 -0.361 0.454 1.030 Note. F: AHT F test; p: p-value of AHT-F test for moderator effect or t-tests that compare each level against zero not all (Zhang et al., 2021). For instance, Lutz et al. (2021) found an overall significant effect for external behaviors, including addiction, and the diagnosis was not a moderator. Pasion and Barbosa (2019) reported that the effect size was not significant for alcohol-related studies, similar to what we found. Zhang et al. (2021) highlighted how the effect size was significant only when small sample-sized studies were excluded (N < 20 in each group), whereas we found it significant irrespective of sample size. This inconsistency may be due to different inclusion and exclusion criteria applied (e.g., only use data from Cz or Fz electrodes in Zhang et al., 2021 vs. making use of as much data as possible in ours) and different analytic approaches used to address the effect size de pendency. Furthermore, we found that stimulants in general and cocaine in specific were associated with more severe impairments in ERN amplitude, which resembled the results of a meta-analysis on motor inhibition (Smith et al., 2014). These findings stressed the effect of stimulant use, with cocaine in particular, on dopaminergic and seroto nergic prefrontal-subcortical pathways that are critical for cognitive control. As to interpreting the cross-group similarities, we may speculate that addiction as well as dopamine-related neurological disorders such as Parkinson’s disease are characterized by a significant loss of dopa minergic neurons in the substantia nigra and of their projections into the striatum and prefrontal cortex, whose dopamine levels play an important role in the generation of the ERN (Gramage and Herradon, 2011; Jocham and Ullsperger, 2009). Furthermore, Age was found as an important moderator of ERN amplitude, such that the effect size for the older group (over 59 yrs) was larger than that of the young adult group (≤39 yrs), implying more serious error processing deficits. This is in line with Boen et al. (2022), who showed that ERN amplitude peaks in young adults. One potential reason is that normal aging is accompanied by gradual degradation of the mesocorticolimbic dopaminergic system, which plays an important role in error processing (Beste et al., 2009; Falkenstein et al., 2001b; Nieuwenhuis et al., 2002; Ridderinkhof and Krugers, 2022). The absence of an age effect in two other relevant meta-analyses (Lutz et al., 2021; Vallet et al., 2021) likely pertains to the absence of a group of older adults since they did not include neurological disorders-related studies. In addition, we did not find a moderating effect of Medicine, which was in line with previous studies that showed neither an acute nor a chronic medicine effect (Stemmer et al., 2007; Willemssen et al., 2008), implying that the reduction of ERN was unaffected by medicine status. We also did not find the Task used to elicit errors to be a significant moderator. This seemingly conflicts with Lutz et al. (2021); however, ERN amplitude has been shown to be highly correlated across different tasks (Riesel et al., 2013; Segalowitz et al., 2010). We also did not find 13 Y. Liu et al. Neuroscience and Biobehavioral Reviews 148 (2023) 105127 Fig. 3. Forest plot of Pe amplitude. Table 3 Moderator analysis of the overall effect size of Pe amplitude. Moderator m k F Group Addiction Neurological disease Electrode CPz Cz Pz Others Task Flanker GNG Others Pe definition Delta Response locked 23 10 13 23 6 8 8 12 23 12 5 9 23 6 22 47 28 19 47 8 11 11 17 47 26 8 13 47 7 40 F(1, 18.1) = 0.003 F(3, 8.62) = 0.604 F(2, 7.44) = 1.31 F(1, 5.76) = 0.188 g SE t df -0.142 -0.150 0.094 0.119 -1.505 -1.262 8.12 10.04 -0.019 -0.162 -0.291 -0.120 0.154 0.117 0.131 0.096 -0.124 -1.384 -2.223 -1.250 4.89 6.51 6.97 9.63 -0.210 -0.276 0.038 0.072 0.186 0.140 -2.917 -1.487 0.271 8.99 3.65 6.78 -0.073 -0.159 0.185 0.078 -0.391 -2.039 4.94 17.11 Note. F: AHT F test; p: p-value of AHT-F test for moderator effect or t-tests that compare each level against zero 14 p 0.958 0.170 0.236 0.629 0.906 0.212 0.062 0.241 0.326 0.017 0.218 0.795 0.680 0.712 0.057 CI low CI high -0.358 -0.414 0.075 0.115 -0.419 -0.442 -0.600 -0.335 0.381 0.119 0.019 0.095 -0.372 -0.813 -0.296 -0.047 0.260 0.372 -0.551 -0.324 0.406 0.005 Y. Liu et al. Neuroscience and Biobehavioral Reviews 148 (2023) 105127 Fig. 4. Forest plot of ERN latency. the Task used to elicit errors to be a significant moderator, which was in line with Lutz et al. (2021). Furthermore, ERN amplitude has been shown to be highly correlated across different tasks (Riesel et al., 2013; Segalowitz et al., 2010). Though Pasion and Barbosa (2019) found a larger effect size when go/no-go rather than other tasks were used, an evaluation of the moderating effect of Task is lacking. In exploratory analyses, we did not find that Electrode, Percentage of males, or ERN definition (delta/response-locked) contributed significantly to the effect size heterogeneity, nor did Severity (clinical/subclinical) and Absti nence status within addiction-related studies. For ERN latency, the effect size for neurological disorders was more negative than that of addiction, indicating faster error processing speed. This result seems counterintuitive since the impaired error-processing ability is expected to be accompanied by delayed error-processing speed. One might speculate that impairment in error-processing cir cuitry is associated with a ‘quick-and-dirty’ trajectory of error detection, leading to premature false negatives that could have been prevented by more deliberate error processing. Of note, though ERN latency alteration was significant in neurological disorders, only one included study reported significantly shorter latency in Parkinson’s disease patients than in healthy controls, which was explained as a reflection of reduced ERN amplitude instead (Falkenstein et al., 2001a). Several others indicated numerical differences between patients and controls (Hogan et al., 2006; Olson et al., 2018; Peterburs et al., 2012; Thurm et al., 2013). For addiction-related studies, most effect sizes approached zero, suggesting no apparent difference from the control group. In sum, the effect of Group should be explained with caution and more studies were needed to validate the present finding. For Pe amplitude, none of the examined within-study moderators showed significant effects. The negative result of Task conflicts with that reported by Lutz et al. (2021), where the go/no-go task was found to induce larger group differences than the flanker task. Although a sig nificant effect size was found here for the flanker task, it did not differ significantly from the ones when other tasks were used, resulting in a non-significant moderating effect of Task. Reasons similar to those used to explain the discrepant findings about ERN amplitude may also apply here. As exploratory moderators, Electrode and Pe definition (delta/r esponse-locked) were not found to contribute significantly to the effect 15 Neuroscience and Biobehavioral Reviews 148 (2023) 105127 Y. Liu et al. Fig. 5. Forest plot of Pe latency. size heterogeneity. Furthermore, no Group effect was found for both Pe amplitude and Pe latency, implying impairments in later stages of error processing that were comparable between addiction and neurological disorders. this research topic. Furthermore, several prospective studies suggested that ERN and Pe indices may serve as biomarkers of behavioral changes in treatment for substance use disorder (e.g., treatment discontinuance: Steele et al., 2014; relapse: Luijten et al., 2016; Marhe et al., 2013), indicating their clinical relevance. We do not mean to neglect the role of other core constructs in the development of addiction, as defined by a recent Delphi study (Yücel et al., 2018). However, compared to error-processing impairments, these other components were relatively less studied in neurological disorders, thus hampering intergroup comparison. In line with the brain disease model of addiction, a previous review reported that people with long-term abuse showed neuropsychological impairments of executive (inhibitory) control, working memory, and decision-making, together with neurobiological alterations in the fron totemporal and basal ganglia circuits (Yücel et al., 2007). Also a recent meta-analysis found alterations in gray matter and white matter in substance use disorders depending on the severity of the consumption pattern and type of substance used (Pando-Naude et al., 2021). The present study is the first attempt to examine the degree of brain disorder 4.2. Implications A central feature of addictive behavior is the person’s apparent lack of voluntary control over drug self-administration, despite their awareness of the severe adverse consequences of continued use. Ac cording to the process model of self-control failure, a deficient moni toring network leads to insufficient mobilization of the cognitive control network, which results in self-control failure by increasing the value of short-term over long-term outcomes when computing the integrated value of an action (Goschke, 2014; Inzlicht et al., 2015; Kotabe and Hofmann, 2015). Similarly, Luijten et al. (2014) postulated that error processing deficits might indirectly influence other cognitive control domains, including inhibitory control. This fundamental role of error processing in self-control was the main impetus for our present focus on 16 Y. Liu et al. Neuroscience and Biobehavioral Reviews 148 (2023) 105127 Fig. 6. Funnel plot of ERN amplitude. 4.3. Limitations and future study suggestions of addiction by contrasting it with a sample of well-defined brain dis orders. We approached this aim by using available data and applying recently developed methods to address effect size dependency (Puste jovsky and Tipton, 2022). Our analyses confirm that addiction is asso ciated with brain functional changes, expressed in significantly diminished ERN amplitude. While it remains difficult to establish beyond doubt whether the changes in ERN are caused by drug abuse or are predisposing factors, it may be noted that some endophenotype research indicated that diminished ERN (as observed in unaffected rel atives) holds potential as a risk factor for substance use disorders (Euser et al., 2013; Riesel et al., 2019). For instance, Euser et al., (2013) showed that ERN was already reduced before substance use onset in the offspring of substance use disorder parents. Alterations in ERN ampli tude in unaffected first-degree relatives with a family history of sub stance abuse disorder support the utility of the ERN as a transdiagnostic endophenotype. Reduced ERNs may indicate vulnerability for under-controlled behavior and risk for substance (ab)use. The finding that this impairment was not (or only marginally) moderated by Group offers novel support for the suggestion that addiction can be considered a brain disorder, similar to other neurological diseases. Nonetheless, considering the pattern of findings for the ERN (less compromised amplitude and delayed latencies for addiction than neurological disorders, although the former effect just failed to attain statistical significance), the observed error-processing deficiency in addiction does not allow us to simply equate addiction with conven tional brain pathologies such as various neurological disorders in terms of the brain dysfunction level. Therefore, rather than struggling over a label, the crux of the matter is to quantify the degree of dysfunctions in addiction and examine whether these dysfunctions preclude individuals’ capacity to alter their behavior based on foreseeable consequences like typical brain disorders do (Satel and Lilienfeld, 2014). A bold suggestion is that instead of perceiving addiction and neurological disorders as identical or distinct patterns, they may represent nuances within a spectrum of brain pathology, which comes to expression in error processing. Of note, the neurocentric view of addiction does not downplay the influence of social, environmental, developmental, or socioeconomic processes as both causes and consequences of substance use, but instead assumes the brain as the underlying material substrate upon which those factors come to play and from which the responses originate (Heilig et al., 2021). A couple of limitations should be acknowledged. First, all studies included were required to have an experimental group and a control group, which made assessing the continua of the full range of symptoms impossible. For addiction, we coded clinical and subclinical samples to tap this continuum, and it turned out that this factor did not moderate error processing. By contrast, it is hard to quantify the severity of neurological disorders across complaints consistently. For instance, cognitive deterioration develops over time in progressive neurodegen erative disorders (e.g., Alzheimer’s disease and Parkinson’s disease; Gao and Hong, 2008), which is not necessarily the case for traumatic brain injury. Second, we excluded studies having participants with comorbid disorders (e.g., anxiety, ADHD) to eliminate the confounding effect. For instance, alcohol use disorder is often combined with anxiety disorder which is associated with enhanced ERN amplitude (Schellekens et al., 2010). Albeit relatively pure substance and neurological disease effect were obtained, this practice hampered the generalization of conclusions since these two samples are likely to have comorbidities that bias error processing (Pasion and Barbosa, 2019; Seow et al., 2020). Future studies are recommended to record and report comorbid disorders consistently, which allows isolating relevant effects neatly through statistical analysis rather than paper removal. Third, unlike addiction, most neurological diseases need medical treatment. However, detailed medical use history was often unavailable, which if accessible would likely outperform the coarse measure of on/off/overnight washout in evaluating the medical effect. Given the concerns of the last two points, studies in the future are encouraged to increase transparent reports of moderators that may help explain heterogeneity in the results. Fourth, we did not include behav ioral addiction since too little published evidence is available to allow running a separate meta-analysis. However, we can imagine that along with the forthcoming accumulation of evidence, the pharmacological effects of substances on error processing can be isolated from non-pharmacological factors through meta-regression analysis. Fifth, we did not explicitly require all included studies to have at least six error trials for a reliable ERN estimation (Olvet and Hajcak, 2009), but assumed that authors complied with this rule of thumb by default. More strictly speaking, a recent meta-analysis indicated that 16 error trials were needed for internal consistency of 0.80 (Clayson, 2020). Studies in the future may examine the effect number of error trials by treating it as a moderator. Still, seven influential effect sizes were excluded from the 17 18 Q1/ aim clear? Q2/ design proper for aim? Q3/ sample size justified? Q4/target population clear? Q5/ sample frame proper? Q6/ participants selected properly? Q7/IV and DV Q8/IV and DV Q9/ Q10/can measurement measurement significance methods be validity reliability and precision replicated? (p and /or CI) Q11/basic data described adequately? Q12/results Q13/ internally present the consistent? results for all analysis in methods? Q14/ discussion and conclusion justified by results? Q15/ limitation discussed? Q16/ ethics and consent form % score Beste et al. (2006) Beste et al. (2009) Beste et al. (2017) Falkenstein et al. (2001a) Falkenstein et al. (2005) Gehring and Knight (2000) Hogan et al. (2006) Holroyd et al. (2002) Ito and Kitagawa (2005) Ito and Kitagawa (2006) Larson et al. (2009) Larson et al. (2012) Lópezet al. (2015) Maier et al. (2015) Mathalon et al. (2003) Niessen et al. (2020) Olson et al. (2018) Peterburs et al. (2011) Peterburs et al. (2012) Y Y N Y Y Y Y Y y Y N Y Y Y N Y 81.25 Y Y N Y Y Y Y Y Y Y N Y Y Y N Y 81.25 Y Y N Y Y Y Y Y Y N N Y Y Y N Y 75 Y Y N Y Y Y Y Y Y Y N Y Y Y N N 75 Y Y N Y Y Y Y Y Y Y N Y Y Y N N 75 Y Y N Y Y Y Y Y Y Y Y N N Y N Y 75 Y Y N Y Y Y Y Y Y N N Y Y Y Y Y 81.25 Y Y N Y Y Y Y Y Y Y N Y Y Y N Y 81.25 Y Y N Y Y Y Y Y Y Y Y Y N Y N N 75 Y Y N Y Y Y Y Y Y Y Y Y N Y N N 75 Y Y N Y Y Y Y Y Y Y Y Y Y Y Y Y 93.75 Y Y N Y Y Y Y Y Y Y N Y Y Y Y Y 87.5 Y Y N Y Y Y Y Y Y Y Y Y Y Y Y Y 93.75 Y Y N Y Y Y Y Y Y Y Y Y Y Y Y Y 93.75 Y Y N Y Y Y Y Y Y N N Y Y Y N N 68.75 Y Y N Y Y Y Y Y Y Y N Y Y Y Y Y 87.5 Y Y N Y Y Y Y Y Y Y y Y Y Y Y Y 93.75 Y Y N Y Y Y Y Y Y Y N Y Y Y N Y 81.25 Y Y N Y Y Y Y Y Y Y Y Y Y Y N Y 87.5 Y Y N Y Y Y Y Y Y Y N Y N Y Y Y 81.25 (continued on next page) Neuroscience and Biobehavioral Reviews 148 (2023) 105127 Ref Y. Liu et al. Table 4 Quality assessment of studies included according to the AXIS. Ref 19 Q2/ design proper for aim? Q3/ sample size justified? Q4/target population clear? Q5/ sample frame proper? Q6/ participants selected properly? Q7/IV and DV Q8/IV and DV Q9/ Q10/can measurement measurement significance methods be validity reliability and precision replicated? (p and /or CI) Q11/basic data described adequately? Q12/results Q13/ internally present the consistent? results for all analysis in methods? Q14/ discussion and conclusion justified by results? Q15/ limitation discussed? Q16/ ethics and consent form % score Y Y N Y Y Y Y Y Y Y N Y Y Y Y Y 87.5 Y Y N Y Y Y Y Y Y Y N Y Y Y Y Y 87.5 Y Y N Y Y Y Y Y Y Y Y Y Y Y N Y 87.5 Y Y N Y Y Y Y Y Y N N Y Y Y Y Y 81.25 Y Y N Y Y N Y Y Y Y N Y Y Y N Y 75 Y Y N Y Y Y Y Y Y Y N Y Y Y N Y 81.25 Y Y N Y Y Y Y Y Y Y N Y Y Y N Y 81.25 Y Y N Y Y Y Y Y Y Y N Y Y Y N Y 81.25 Y Y N Y Y Y N Y Y N N Y Y Y Y Y 75 Y Y N Y Y Y N Y Y Y Y Y Y Y Y Y 87.5 Y Y N Y Y Y Y Y Y Y N Y Y Y N Y 81.25 Y Y N Y Y Y Y Y Y Y N N Y Y N Y 75 Y Y N Y Y N Y Y Y Y Y Y Y Y Y Y 81.25 Y Y N Y Y N Y Y Y Y N Y Y Y Y Y 81.25 Y Y N Y Y Y Y Y Y Y N Y N Y Y Y 81.25 Y Y Y Y Y Y Y Y Y Y N Y N Y Y Y 87.5 Y Y Y Y Y Y Y Y Y Y N Y Y Y Y Y 93.75 (continued on next page) Neuroscience and Biobehavioral Reviews 148 (2023) 105127 Peterburs et al. (2015) Pontifex et al. (2009) Seer et al. (2017) Seifert et al. (2011) Solbakk et al. (2014) Stemmer et al. (2007) Thurm et al. (2013) Ullsperger et al. (2002) Ullsperger and von Cramon (2006) Verleger et al. (2013) Volpato et al. (2016) Willemssen et al. (2008) Willemssen et al. (2009) Chen et al. (2013) Franken et al. (2007) Franken et al. (2010) Franken et al. (2017) Q1/ aim clear? Y. Liu et al. Table 4 (continued ) Y. Liu et al. Table 4 (continued ) Ref 20 Q2/ design proper for aim? Q3/ sample size justified? Q4/target population clear? Q5/ sample frame proper? Q6/ participants selected properly? Q7/IV and DV Q8/IV and DV Q9/ Q10/can measurement measurement significance methods be validity reliability and precision replicated? (p and /or CI) Q11/basic data described adequately? Q12/results Q13/ internally present the consistent? results for all analysis in methods? Q14/ discussion and conclusion justified by results? Q15/ limitation discussed? Q16/ ethics and consent form % score Y Y Y Y Y N Y Y Y Y Y Y Y Y Y Y 93.75 Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y 100 Y Y N Y Y Y Y Y Y Y Y Y Y Y N Y 87.5 Y Y N Y Y Y Y Y Y Y N Y Y Y Y Y 87.5 Y Y Y Y Y Y Y Y Y Y N Y Y Y Y Y 93.75 Y Y Y Y Y Y Y Y Y Y N N Y Y Y Y 87.5 Y Y N Y Y N Y Y Y Y N N Y Y Y Y 75 Y Y N Y Y Y Y Y Y Y Y Y Y Y Y Y 93.75 Y Y N Y Y Y Y Y Y Y Y Y Y Y N Y 87.5 Y Y N Y Y Y Y Y Y Y Y Y Y Y Y Y 93.75 Y Y N Y Y Y N Y Y Y Y Y Y Y Y Y 87.5 Y Y N Y Y N Y Y N Y N Y Y Y N Y 68.75 Note. N = No; Y = Yes; IV: independent variable; DV: dependent variable; CI: confidence interval Neuroscience and Biobehavioral Reviews 148 (2023) 105127 Fridberg et al. (2013) Gorka et al. (2019) Kim and Kim (2019) Lannoy et al. (2017) Luijten et al. (2011) Marhe et al. (2013) Morie et al. (2014) Padilla et al. (2011) Rass et al. (2014) Smith et al. (2015) Smith et al. (2016) Smith et al. (2017) Sokhadze et al. (2008) Q1/ aim clear? Y. Liu et al. Neuroscience and Biobehavioral Reviews 148 (2023) 105127 analyses, potentially due to uncommon tasks used (e.g., lexical recog nition paradigm: Ito and Kitagawa, 2005) and broad sources of ERN and Pe activities when 128 channel caps were used (e.g., Larson et al., 2012; Solbakk et al., 2014). Studies in the future may want to take these methodological considerations into account. Furthermore, while our understanding of differences in ERN and Pe latencies remains limited at present, and the focus in the empirical literature has been on amplitude, nonetheless some studies have examined (and reported effects on) la tency, and we cannot exclude that such effects are meaningful and might be corroborated by meta-analysis. For instance, several studies have drawn analogies between the Pe and P3, both in terms of functional significance and underlying sources (e.g., Ridderinkhof et al., 2009). P3 latency has often been found to be meaningfully affected by experi mental effects and group or individual differences. For example, No-go P3 latency has been argued to reflect the slowing of inhibition pro cessing (as explored further in a meta-analysis by Cheng et al., 2019). Interpreting delayed latency as impaired cognitive processing speed might not be limited to inhibition, and can potentially be expanded to error processing. Concerning ERN latency, some studies have suggested that delayed ERN latency may be interpreted as a delay in upregulating cognitive control processes following an error (e.g., Larson et al., 2012). Since meta-analytic outcomes might help in generating hypotheses (or, instead, confirm that latency is not fruitful as a measure to include in further studies), we included explorative analyses of latencies in the present meta-analysis. Rita Pasion, Junfeng Sun, and William Vallet. Appendix A. Supporting information Supplementary data associated with this article can be found in the online version at doi:10.1016/j.neubiorev.2023.105127. References Arias-Carrión, O., Stamelou, M., Murillo-Rodríguez, E., Menéndez-González, M., Pöppel, E., 2010. Dopaminergic reward system: a short integrative review. Int. Arch. Med. 3, 1–6. https://doi.org/10.1186/1755-7682-3-24. Atit, K., Power, J.R., Pigott, T., Lee, J., Geer, E.A., Uttal, D.H., Ganley, C.M., Sorby, S.A., 2022. Examining the relations between spatial skills and mathematical performance: a meta-analysis. Psychon. Bull. Rev. 29, 699–720. https://doi.org/10.3758/s13423021-02012-w. Bechara, A., 2005. Decision making, impulse control and loss of willpower to resist drugs: a neurocognitive perspective. Nat. Neurosci. 8, 1458. https://doi.org/ 10.1038/nn1584. Bechara, A., Tranel, D., Damasio, H., 2000. Characterization of the decision-making deficit of patients with ventromedial prefrontal cortex lesions. Brain 123, 2189–2202. https://doi.org/10.1093/brain/123.11.2189. Berres, S., Erdfelder, E., 2021. The sleep benefit in episodic memory: an integrative review and a meta-analysis. Psychol. Bull. 147, 1309. https://doi.org/10.1037/ bul0000350. Beste, C., Saft, C., Andrich, J., Gold, R., Falkenstein, M., 2006. Error processing in Huntington’s disease. PLoS One 1, e86. https://doi.org/10.1371/journal. pone.0000086. Beste, C., Willemssen, R., Saft, C., Falkenstein, M., 2009. Error processing in normal aging and in basal ganglia disorders. Neuroscience 159, 143–149. https://doi.org/ 10.1016/j.neuroscience.2008.12.030. Beste, C., Mückschel, M., Rosales, R., Domingo, A., Lee, L., Ng, A., Klein, C., Münchau, A., 2017. Striosomal dysfunction affects behavioral adaptation but not impulsivity—evidence from X-linked dystonia-parkinsonism. Mov. Disord. 32, 576–584. https://doi.org/10.1002/mds.26895. Boen, R., Quintana, D.S., Ladouceur, C.D., Tamnes, C.K., 2022. Age-related differences in the error-related negativity and error positivity in children and adolescents are moderated by sample and methodological characteristics: a meta-analysis. Psychophysiology 59, e14003. https://doi.org/10.1111/psyp.14003. Cadet, J.L., Bisagno, V., Milroy, C.M., 2014. Neuropathology of substance use disorders. Acta Neuropathol. 127, 91–107. https://doi.org/10.1007/s00401-013-1221-7. Chen, H., Jiang, H., Guo, Q., Du, J., Wang, J., Zhao, M., 2013. Case-control study of error-related negativity among males with heroin dependence undergoing rehabilitation. Shanghai Arch. Psychiatry 25, 141–148. https://doi.org/10.3969/j. issn.1002-0829.2013.03.003. Cheng, C.-H., Tsai, H.-Y., Cheng, H.-N., 2019. The effect of age on N2 and P3 components: a meta-analysis of Go/Nogo tasks. Brain Cogn. 135, 103574 https:// doi.org/10.1016/j.bandc.2019.05.012. Clayson, P.E., 2020. Moderators of the internal consistency of error-related negativity scores: a meta-analysis of internal consistency estimates. Psychophysiology 57, e13583. https://doi.org/10.1111/psyp.13583. Coles, M.G.H., Scheffers, M.K., Holroyd, C.B., 2001. Why is there an ERN/Ne on correct trials? Response representations, stimulus-related components, and the theory of error-processing. Biol. Psychol. 56, 173–189. https://doi.org/10.1016/s0301-0511 (01)00076-x. Desender, K., Ridderinkhof, K.R., Murphy, P.R., 2021. Understanding neural signals of post-decisional performance monitoring: an integrative review. Elife 10, e67556. https://doi.org/10.7554/eLife.67556. Di Gregorio, F., Maier, M.E., Steinhauser, M., 2018. Errors can elicit an error positivity in the absence of an error negativity: evidence for independent systems of human error monitoring. Neuroimage 172, 427–436. https://doi.org/10.1016/j. neuroimage.2018.01.081. Downes, M.J., Brennan, M.L., Williams, H.C., Dean, R.S., 2016. Development of a critical appraisal tool to assess the quality of cross-sectional studies (AXIS. BMJ Open 6, e011458. https://doi.org/10.1136/bmjopen-2016-011458. Falkenstein, M., Hoormann, J., Christ, S., Hohnsbein, J., 2000. ERP components on reaction errors and their functional significance: a tutorial. Biol. Psychol. 51, 87–107. https://doi.org/10.1016/s0301-0511(99)00031-9. Falkenstein, M., Hielscher, H., Dziobek, I., Schwarzenau, P., Hoormann, J., Sundermann, B., Hohnsbein, J., 2001a. Action monitoring, error detection, and the basal ganglia: an ERP study. Neuroreport 12, 157–161. https://doi.org/10.1097/ 00001756-200101220-00039. Falkenstein, M., Hoormann, J., Hohnsbein, J., 2001b. Changes of error-related ERPs with age. Exp. Brain Res 138, 258–262. https://doi.org/10.1007/s002210100712. Falkenstein, M., Willemssen, R., Hohnsbein, J., Hielscher, H., 2005. Error processing in Parkinson’s disease: the error positivity (Pe). J. Psychophysiol. 19, 305–310. https:// doi.org/10.1027/0269-8803.19.4.305. Franken, I.H.A., van Strien, J.W., Franzek, E.J., van de Wetering, B.J., 2007. Errorprocessing deficits in patients with cocaine dependence. Biol. Psychol. 75, 45–51. https://doi.org/10.1016/j.biopsycho.2006.11.003. Franken, I.H.A., van Strien, J.W., Kuijpers, I., 2010. Evidence for a deficit in the salience attribution to errors in smokers. Drug Alcohol Depend. 106, 181–185. https://doi. org/10.1016/j.drugalcdep.2009.08.014. 5. Conclusions The current meta-analyses compared 17 addiction-related and 32 neurological disorder-related studies regarding electrophysiological in dexes of error processing. The main finding is that, in addition to a significant pooled effect size in both groups, neurological disorders were associated with a marginally significantly greater impairment in ERN amplitude than addiction. Neurological disorders also presented shorter ERN latencies than addiction when compared with controls. The effect of Age as a moderator was in the expected direction, while no evidence was found for the influence of other moderators. The present findings can shed light on the ‘addiction as a brain disease’ debate. We suggest that, rather than a black-and-white position in this debate, a fruitful way forward is to quantify the degree of brain dysfunctions in addiction. And then examine to which extent the corresponding consequences preclude individuals’ capacity to alter their drug-related behavior. Funding This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors. Declarations of interest None Data Availability Data will be made available on request. Acknowledgment We would like to acknowledge researchers who provided informa tion to support effect size calculation (in alphabetic order): Michael Falkenstein, Ingmar H.A. Franken, William J. Gehring, Michael J. Lar son, Maartje Luijten, Miranda Christine Lutz, Reshmi Marhe, Kristen P. Morie, Eva Niessen, Jutta Peterburs, Matthew Pontifex, Caroline Seer, Sidney J. Segalowitz, Janette L. Smith, Estate Sokhadze, Anne-Kristin Solbakk, Brigitte Stemmer, Markus Ullsperger, and Jan R. Wessel. And those who gave us data analyses and other suggestions: Enrica Donolato, 21 Y. Liu et al. Neuroscience and Biobehavioral Reviews 148 (2023) 105127 Franken, I.H.A., Luijten, M., van der Veen, F.M., van Strien, J.W., 2017. Cognitive control in young heavy drinkers: an ERP study. Drug Alcohol Depend. 175, 77–83. https:// doi.org/10.1016/j.drugalcdep.2017.01.036. Fridberg, D.J., Skosnik, P.D., Hetrick, W.P., O’Donnell, B.F., 2013. Neural correlates of performance monitoring in chronic cannabis users and cannabis-naive controls. J. Psychopharmacol. 27, 515–525. https://doi.org/10.1177/0269881113477745. Gao, H.-M., Hong, J.-S., 2008. Why neurodegenerative diseases are progressive: uncontrolled inflammation drives disease progression. Trends Immunol. 29, 357–365. https://doi.org/10.1016/j.it.2008.05.002. Gehring, W.J., Knight, R.T., 2000. Prefrontal-cingulate interactions in action monitoring. Nat. Neurosci. 3, 516–520. https://doi.org/10.1038/74899. Goldstein, D.S., Sullivan, P., Holmes, C., Kopin, I.J., Basile, M.J., Mash, D.C., 2011. Catechols in post-mortem brain of patients with Parkinson disease. Eur. J. Neurol. 18, 703–710. https://doi.org/10.1111/j.1468-1331.2010.03246.x. Goldstein, R.Z., Bechara, A., Garavan, H., Childress, A.R., Paulus, M.P., Volkow, N.D., 2009. The neurocircuitry of impaired insight in drug addiction. Trends Cogn. Sci. 13, 372–380. https://doi.org/10.1016/j.tics.2009.06.004. Gorka, S.M., Lieberman, L., Kreutzer, K.A., Carrillo, V., Weinberg, A., Shankman, S.A., 2019. Error-related neural activity and alcohol use disorder: differences from risk to remission. Prog. Neuropsychopharmacol. Biol. Psychiatry 92, 271–278. https://doi. org/10.1016/j.pnpbp.2019.01.011. Goschke, T., 2014. Dysfunctions of decision-making and cognitive control as transdiagnostic mechanisms of mental disorders: advances, gaps, and needs in current research. Int. J. Methods Psychiatr. Res. 23, 41–57. https://doi.org/ 10.1002/mpr.1410. Gramage, E., Herradon, G., 2011. Connecting Parkinson’s disease and drug addiction: common players reveal unexpected disease connections and novel therapeutic approaches. Curr. Pharm. Des. 17, 449–461. https://doi.org/10.2174/ 138161211795164103. Harsay, H.A., Spaan, M., Wijnen, J.G., Ridderinkhof, K.R., 2012. Error awareness and salience processing in the oddball task: shared neural mechanisms. Front. Hum. Neurosci. 6, 246. https://doi.org/10.3389/fnhum.2012.00246. Hedges, L.V., 1981. Distribution theory for Glass’s estimator of effect size and related estimators. J. Educ. Stat. 6, 107–128. https://doi.org/10.2307/1164588. Heilig, M., MacKillop, J., Martinez, D., Rehm, J., Leggio, L., Vanderschuren, L.J.M.J., 2021. Addiction as a brain disease revised: why it still matters, and the need for consilience. Neuropsychopharmacology 46, 1715–1723. https://doi.org/10.1038/ s41386-020-00950-y. Herrmann, M.J., Römmler, J., Ehlis, A.-C., Heidrich, A., Fallgatter, A.J., 2004. Source localization (LORETA) of the error-related-negativity (ERN/Ne) and positivity (Pe). Cogn. Brain Res. 20, 294–299. https://doi.org/10.1016/j.cogbrainres.2004.02.013. Higgins, J.P.T., Thompson, S.G., 2002. Quantifying heterogeneity in a meta-analysis. Stat. Med. 21, 1539–1558. https://doi.org/10.1002/sim.1186. Hogan, A.M., Vargha-Khadem, F., Saunders, D.E., Kirkham, F.J., Baldeweg, T., 2006. Impact of frontal white matter lesions on performance monitoring: ERP evidence for cortical disconnection. Brain 129, 2177–2188. https://doi.org/10.1093/brain/ awl160. Holroyd, C.B., Praamstra, P., Plat, E., Coles, M.G.H., 2002. Spared error-related potentials in mild to moderate Parkinson’s disease. Neuropsychologia 40, 2116–2124. https://doi.org/10.1016/s0028-3932(02)00052-0. Holroyd, C.B., Nieuwenhuis, S., Mars, R.B., Coles, M.G.H., 2004. Anterior cingulate cortex, selection for action, and error processing. In: Posner, M.I. (Ed.), Cognitive neuroscience of attention. The Guilford Press, pp. 219–231. Inzlicht, M., Bartholow, B.D., Hirsh, J.B., 2015. Emotional foundations of cognitive control. Trends Cogn. Sci. 19, 126–132. https://doi.org/10.1016/j.tics.2015.01.004. Ito, J., Kitagawa, J., 2005. Error processing in patients with Alzheimer’s disease. Pathophysiology 12, 97–101. https://doi.org/10.1016/j.pathophys.2005.02.003. Ito, J., Kitagawa, J., 2006. Performance monitoring and error processing during a lexical decision task in patients with Parkinson’s disease. J. Geriatr. Psychiatry Neurol. 19, 46–54. https://doi.org/10.1177/0891988705284716. Jocham, G., Ullsperger, M., 2009. Neuropharmacology of performance monitoring. Neurosci. Biobehav. Rev. 33, 48–60. https://doi.org/10.1016/j. neubiorev.2008.08.011. Kärkkäinen, O., Kokla, M., Lehtonen, M., Auriola, S., Martiskainen, M., Tiihonen, J., Karhunen, P.J., Hanhineva, K., Kok, E., 2021. Changes in the metabolic profile of human male postmortem frontal cortex and cerebrospinal fluid samples associated with heavy alcohol use. Addict. Biol. 26, e13035 https://doi.org/10.1111/ adb.13035. Kim, E.-H., Kim, M.-S., 2019. An event-related potential study of error-monitoring deficits in female college students who participate in binge drinking. Clin. Psychopharmacol. Neurosci. Sci. J. Korean Coll. Neuropsychopharmacol. 17, 80–92. https://doi.org/10.9758/cpn.2019.17.1.80. Kotabe, H.P., Hofmann, W., 2015. On integrating the components of self-control. Perspect. Psychol. Sci. 10, 618–638. https://doi.org/10.1177/1745691615593382. Krönke, K.-M., Wolff, M., Mohr, H., Kräplin, A., Smolka, M.N., Bühringer, G., Goschke, T., 2018. Monitor yourself! Deficient error-related brain activity predicts real-life self-control failures. Cogn. Affect. Behav. Neurosci. 18, 622–637. https:// doi.org/10.3758/s13415-018-0593-5. Lannoy, S., D’Hondt, F., Dormal, V., Billieux, J., Maurage, P., 2017. Electrophysiological correlates of performance monitoring in binge drinking: impaired error-related but preserved feedback processing. Clin. Neurophysiol. J. Int. Fed. Clin. Neurophysiol. 128, 2110–2121. https://doi.org/10.1016/j.clinph.2017.08.005. Larson, M.J., Kaufman, D.A.S., Kellison, I.L., Schmalfuss, I.M., Perlstein, W.M., 2009. Double jeopardy! The additive consequences of negative affect on performancemonitoring decrements following traumatic brain injury. Neuropsychology 23, 433–444. https://doi.org/10.1037/a0015723. Larson, M.J., Clayson, P.E., Farrer, T.J., 2012. Performance monitoring and cognitive control in individuals with mild traumatic brain injury. J. Int. Neuropsychol. Soc. 18, 323–333. https://doi.org/10.1017/S1355617711001779. Lenzoni, S., Baker, J., Sumich, A.L., Mograbi, D.C., 2022. New insights into neural networks of error monitoring and clinical implications: a systematic review of ERP studies in neurological diseases. Rev. Neurosci. 33, 161–179. https://doi.org/ 10.1515/revneuro-2021-0054. Leshner, A.I., 1997. Addition is a brain disease, and it matters. Sci. N. Ser. 278, 45–47. https://doi.org/10.1126/science.278.5335.45. Leuthold, H., Sommer, W., 1999. ERP correlates of error processing in spatial SR compatibility tasks. Clin. Neurophysiol. 110, 342–357. https://doi.org/10.1016/ s1388-2457(98)00058-3. Levy, N., 2013. Addiction is not a brain disease (and it matters. Front. Psychiatry 4, 24. https://doi.org/10.3389/fpsyt.2013.00024. López-Góngora, M., Escartín, A., Martínez-Horta, S., Fernández-Bobadilla, R., Querol, L., Romero, S., Mañanas, M.À., Riba, J., 2015. Neurophysiological evidence of compensatory brain mechanisms in early-stage multiple sclerosis. PLoS One 10, e0136786. https://doi.org/10.1371/journal.pone.0136786. Lüdecke, D., Lüdecke, M.D., Calculator’from David, B.W., 2019. Package ‘esc.’ R Packag. Version 0.5 1, 2019. Luijten, M., van Meel, C.S., Franken, I.H.A., 2011. Diminished error processing in smokers during smoking cue exposure. Pharmacol. Biochem. Behav. 97, 514–520. https://doi.org/10.1016/j.pbb.2010.10.012. Luijten, M., Machielsen, M.W.J., Veltman, D.J., Hester, R., de Haan, L., Franken, I.H.A., 2014. Systematic review of ERP and fMRI studies investigating inhibitory control and error processing in people with substance dependence and behavioural addictions. J. Psychiatry Neurosci. 39, 149–169. https://doi.org/10.1503/ jpn.130052. Luijten, M., Kleinjan, M., Franken, I.H.A., 2016. Event-related potentials reflecting smoking cue reactivity and cognitive control as predictors of smoking relapse and resumption. Psychopharmacol. (Berl. ) 233, 2857–2868. https://doi.org/10.1007/ s00213-016-4332-8. Lutz, M.C., Kok, R., Verveer, I., Malbec, M., Koot, S., van Lier, P.A.C., Franken, I.H.A., 2021. Diminished error-related negativity and error positivity in children and adults with externalizing problems and disorders: a meta-analysis on error processing. J. Psychiatry Neurosci. 46, E615–E627. https://doi.org/10.1503/jpn.200031. Maier, M.E., Di Gregorio, F., Muricchio, T., Di Pellegrino, G., 2015. Impaired rapid error monitoring but intact error signaling following rostral anterior cingulate cortex lesions in humans. Front. Hum. Neurosci. 9, 339. https://doi.org/10.3389/ fnhum.2015.00339. Marhe, R., van de Wetering, B.J.M., Franken, I.H.A., 2013. Error-related brain activity predicts cocaine use after treatment at 3-month follow-up. Biol. Psychiatry 73, 782–788. https://doi.org/10.1016/j.biopsych.2012.12.016. Masina, F., Tarantino, V., Vallesi, A., Mapelli, D., 2019. Repetitive TMS over the left dorsolateral prefrontal cortex modulates the error positivity: An ERP study. Neuropsychologia 133, 107153. https://doi.org/10.1016/j. neuropsychologia.2019.107153. Mathalon, D.H., Bennett, A., Askari, N., Gray, E.M., Rosenbloom, M.J., Ford, J.M., 2003. Response-monitoring dysfunction in aging and Alzheimer’s disease: an event-related potential study. Neurobiol. Aging 24, 675–685. https://doi.org/10.1016/s01974580(02)00154-9. Morie, K.P., De Sanctis, P., Garavan, H., Foxe, J.J., 2014. Executive dysfunction and reward dysregulation: a high-density electrical mapping study in cocaine abusers. Neuropharmacology 85, 397–407. https://doi.org/10.1016/j. neuropharm.2014.05.016. Navarri, X., Afzali, M.H., Lavoie, J., Sinha, R., Stein, D.J., Momenan, R., Veltman, D.J., Korucuoglu, O., Sjoerds, Z., van Holst, R.J., 2022. How do substance use disorders compare to other psychiatric conditions on structural brain abnormalities? A crossdisorder meta-analytic comparison using the ENIGMA consortium findings. Hum. Brain Mapp. 43, 399–413. https://doi.org/10.1002/hbm.25114. Niessen, E., Ant, J.M., Bode, S., Saliger, J., Karbe, H., Fink, G.R., Stahl, J., Weiss, P.H., 2020. Preserved performance monitoring and error detection in left hemisphere stroke. NeuroImage Clin. 27, 102307 https://doi.org/10.1016/j.nicl.2020.102307. Nieuwenhuis, S., Ridderinkhof, K.R., Talsma, D., Coles, M.G.H., Holroyd, C.B., Kok, A., Van der Molen, M.W., 2002. A computational account of altered error processing in older age: Dopamine and the error-related negativity. Cogn. Affect. Behav. Neurosci. 2, 19–36. https://doi.org/10.3758/cabn.2.1.19. O’Connell, R.G., Dockree, P.M., Bellgrove, M.A., Kelly, S.P., Hester, R., Garavan, H., Robertson, I.H., Foxe, J.J., 2007. The role of cingulate cortex in the detection of errors with and without awareness: a high-density electrical mapping study. Eur. J. Neurosci. 25, 2571–2579. https://doi.org/10.1111/j.1460-9568.2007.05477.x. Olson, R.L., Brush, C.J., Ehmann, P.J., Buckman, J.F., Alderman, B.L., 2018. A history of sport-related concussion is associated with sustained deficits in conflict and error monitoring. Int. J. Psychophysiol. 132, 145–154. https://doi.org/10.1016/j. ijpsycho.2018.01.006. Olvet, D.M., Hajcak, G., 2009. The stability of error-related brain activity with increasing trials. Psychophysiology 46, 957–961. https://doi.org/10.1111/j.14698986.2009.00848.x. Ouzzani, M., Hammady, H., Fedorowicz, Z., Elmagarmid, A., 2016. Rayyan—a web and mobile app for systematic reviews. Syst. Rev. 5, 1–10. https://doi.org/10.1186/ s13643-016-0384-4. Overbeek, T.J.M., Nieuwenhuis, S., Ridderinkhof, K.R., 2005. Dissociable components of error processing: On the functional significance of the Pe vis-à-vis the ERN/Ne. J. Psychophysiol. 19, 319–329. https://doi.org/10.1027/0269-8803.19.4.319. Padilla, M.L., Colrain, I.M., Sullivan, E.V., Mayer, B.Z., Turlington, S.R., Hoffman, L.R., Wagstaff, A.E., Pfefferbaum, A., 2011. Electrophysiological evidence of enhanced 22 Y. Liu et al. Neuroscience and Biobehavioral Reviews 148 (2023) 105127 performance monitoring in recently abstinent alcoholic men. Psychopharmacology 213, 81–91. https://doi.org/10.1007/s00213-010-2018-1. Page, M.J., McKenzie, J.E., Bossuyt, P.M., Boutron, I., Hoffmann, T.C., Mulrow, C.D., Shamseer, L., Tetzlaff, J.M., Akl, E.A., Brennan, S.E., Chou, R., Glanville, J., Grimshaw, J.M., Hróbjartsson, A., Lalu, M.M., Li, T., Loder, E.W., Mayo-Wilson, E., McDonald, S., McGuinness, L.A., Stewart, L.A., Thomas, J., Tricco, A.C., Welch, V.A., Whiting, P., Moher, D., 2021. The PRISMA 2020 statement: an updated guideline for reporting systematic reviews. BMJ 372, n71. https://doi.org/10.1136/bmj.n71. Pando-Naude, V., Toxto, S., Fernandez-Lozano, S., Parsons, C.E., Alcauter, S., GarzaVillarreal, E.A., 2021. Gray and white matter morphology in substance use disorders: a neuroimaging systematic review and meta-analysis. Transl. Psychiatry 11, 29 https://doi.org/0.1038/s41398-020-01128-2. Pasion, R., Barbosa, F., 2019. ERN as a transdiagnostic marker of the internalizingexternalizing spectrum: a dissociable meta-analytic effect. Neurosci. Biobehav. Rev. 103, 133–149. https://doi.org/10.1016/j.neubiorev.2019.06.013. Peterburs, J., Pergola, G., Koch, B., Schwarz, M., Hoffmann, K.-P., Daum, I., Bellebaum, C., 2011. Altered error processing following vascular thalamic damage: evidence from an antisaccade task. PLoS One 6, e21517. https://doi.org/10.1371/ journal.pone.0021517. Peterburs, J., Gajda, K., Koch, B., Schwarz, M., Hoffmann, K.-P., Daum, I., Bellebaum, C., 2012. Cerebellar lesions alter performance monitoring on the antisaccade task—an event-related potentials study. Neuropsychologia 50, 379–389. https://doi.org/ 10.1016/j.neuropsychologia.2011.12.009. Peterburs, J., Thürling, M., Rustemeier, M., Göricke, S., Suchan, B., Timmann, D., Bellebaum, C., 2015. A cerebellar role in performance monitoring - evidence from EEG and voxel-based morphometry in patients with cerebellar degenerative disease. Neuropsychologia 68, 139–147. https://doi.org/10.1016/j. neuropsychologia.2015.01.017. Pezzetta, R., Wokke, M.E., Aglioti, S.M., Ridderinkhof, K.R., 2021. Doing it wrong: a systematic review on electrocortical and behavioral correlates of error monitoring in patients with neurological disorders. Neuroscience. https://doi.org/10.1016/j. neuroscience.2021.01.027. Pontifex, M.B., O’Connor, P.M., Broglio, S.P., Hillman, C.H., 2009. The association between mild traumatic brain injury history and cognitive control. Neuropsychologia 47, 3210–3216. https://doi.org/10.1016/j. neuropsychologia.2009.07.021. Pustejovsky, J.E., clubSandwich: cluster-robust (sandwich) variance estimators with small-sample corrections. Version 0.5.8. 〈https://cran.r-project.org/web/packag es/clubSandwich/index.html〉. Accessed August 15, 2022. Pustejovsky, J.E., Tipton, E., 2018. Small-sample methods for cluster-robust variance estimation and hypothesis testing in fixed effects models. J. Bus. Econ. Stat. 36, 672–683. https://doi.org/10.1080/07350015.2016.1247004. Pustejovsky, J.E., Tipton, E., 2022. Meta-analysis with robust variance estimation: expanding the range of working models. Prev. Sci. 23, 425–438. https://doi.org/ 10.1007/s11121-021-01246-3. Pyasik, M., Scandola, M., Moro, V., 2022. Electrophysiological correlates of action monitoring in brain-damaged patients: a systematic review. Neuropsychologia, 108333 https://doi.org/j.neuropsychologia.2022.108333. Rass, O., Fridberg, D.J., O’Donnell, B.F., 2014. Neural correlates of performance monitoring in daily and intermittent smokers. Clin. Neurophysiol. 125, 1417–1426. https://doi.org/10.1016/j.clinph.2013.12.001. Reinikainen, K.J., Paljärvi, L., Huuskonen, M., Soininen, H., Laakso, M., Riekkinen, P.J., 1988. A post-mortem study of noradrenergic, serotonergic and GABAergic neurons in Alzheimer’s disease. J. Neurol. Sci. 84, 101–116. https://doi.org/10.1016/0022510x(88)90179-7. Ridderinkhof, K.R., Krugers, H.J., 2022. Horizons in human aging neuroscience: from normal neural aging to mental (Fr) agility. Front. Hum. Neurosci. 363. https://doi. org/10.3389/fnhum.2022.815759. Ridderinkhof, K.R., van den Wildenberg, W.P.M., Segalowitz, S.J., Carter, C.S., 2004. Neurocognitive mechanisms of cognitive control: The role of prefrontal cortex in action selection, response inhibition, performance monitoring, and reward-based learning. Brain Cogn. 56, 129–140. https://doi.org/10.1016/j.bandc.2004.09.016. Ridderinkhof, K.R., Ramautar, J.R., Wijnen, J.G., 2009. To PE or not to PE: A P3–like ERP component reflecting the processing of response errors. Psychophysiology 46, 531–538. https://doi.org/10.1111/j.1469-8986.2009.00790.x. Riesel, A., Weinberg, A., Endrass, T., Meyer, A., Hajcak, G., 2013. The ERN is the ERN is the ERN? Convergent validity of error-related brain activity across different tasks. Biol. Psychol. 93, 377–385. https://doi.org/10.1016/j.biopsycho.2013.04.007. Riesel, A., Klawohn, J., Grützmann, R., Kaufmann, C., Heinzel, S., Bey, K., Lennertz, L., Wagner, M., Kathmann, N., 2019. Error-related brain activity as a transdiagnostic endophenotype for obsessive-compulsive disorder, anxiety and substance use disorder. Psychol. Med. 49, 1207–1217. https://doi.org/10.1017/ S0033291719000199. Satel, S., Lilienfeld, S.O., 2014. Addiction and the brain-disease fallacy. Front. Psychiatry 4, 141. https://doi.org/10.3389/fpsyt.2013.00141. Scammacca, N., Roberts, G., Stuebing, K.K., 2014. Meta-analysis with complex research designs: dealing with dependence from multiple measures and multiple group comparisons. Rev. Educ. Res. 84, 328–364. https://doi.org/10.3102/ 0034654313500826. Schellekens, A.F.A., de Bruijn, E.R.A., van Lankveld, C.A.A., Hulstijn, W., Buitelaar, J.K., de Jong, C.A.J., Verkes, R.J., 2010. Alcohol dependence and anxiety increase errorrelated brain activity. Addiction 105, 1928–1934. https://doi.org/10.1111/j.13600443.2010.03065.x. Seer, C., Lange, F., Loens, S., Wegner, F., Schrader, C., Dressler, D., Dengler, R., Kopp, B., 2017. Dopaminergic modulation of performance monitoring in Parkinson’s disease: an event-related potential study. Sci. Rep. 7, 1–13. https://doi.org/10.1038/ srep41222. Segalowitz, S.J., Santesso, D.L., Murphy, T.I., Homan, D., Chantziantoniou, D.K., Khan, S., 2010. Retest reliability of medial frontal negativities during performance monitoring. Psychophysiology 47, 260–270. https://doi.org/10.1111/j.14698986.2009.00942.x. Seifert, S., von Cramon, D.Y., Imperati, D., Tittgemeyer, M., Ullsperger, M., 2011. Thalamocingulate interactions in performance monitoring. J. Neurosci. J. Soc. Neurosci. 31, 3375–3383. https://doi.org/10.1523/jneurosci.6242-10.2011. Seow, T.X.F., Benoit, E., Dempsey, C., Jennings, M., Maxwell, A., McDonough, M., Gillan, C.M., 2020. A dimensional investigation of error-related negativity (ERN) and self-reported psychiatric symptoms. Int. J. Psychophysiol. 158, 340–348. https://doi.org/10.1016/j.ijpsycho.2020.09.019. Smith, J.L., Mattick, R.P., Jamadar, S.D., Iredale, J.M., 2014. Deficits in behavioural inhibition in substance abuse and addiction: a meta-analysis. Drug Alcohol Depend. 145, 1–33. https://doi.org/10.1016/j.drugalcdep.2014.08.009. Smith, J.L., Mattick, R.P., Sufani, C., 2015. Female but not male young heavy drinkers display altered performance monitoring. Psychiatry Res 233, 424–435. https://doi. org/10.1016/j.pscychresns.2015.07.014. Smith, J.L., Iredale, J.M., Mattick, R.P., 2016. Sex differences in the relationship between heavy alcohol use, inhibition and performance monitoring: disconnect between behavioural and brain functional measures. Psychiatry Res. Neuroimaging 254, 103–111. https://doi.org/10.1016/j.pscychresns.2016.06.012. Smith, J.L., Mattick, R.P., Sufani, C., 2017. Error detection and behavioural inhibition in young heavy drinkers. Drug Alcohol Depend. 171, 20–30. https://doi.org/10.1016/ j.drugalcdep.2016.11.016. Sokhadze, E., Stewart, C., Hollifield, M., Tasman, A., 2008. Event-related potential study of executive dysfunctions in a speeded reaction task in cocaine addiction. J. Neurother. 12, 185–204. https://doi.org/10.1080/10874200802502144. Solbakk, A.-K., Funderud, I., Løvstad, M., Endestad, T., Meling, T., Lindgren, M., Knight, R.T., Krämer, U.M., 2014. Impact of orbitofrontal lesions on electrophysiological signals in a stop signal task. J. Cogn. Neurosci. 26, 1528–1545. https://doi.org/10.1162/jocn_a_00561. Steele, V.R., Fink, B.C., Maurer, J.M., Arbabshirani, M.R., Wilber, C.H., Jaffe, A.J., Sidz, A., Pearlson, G.D., Calhoun, V.D., Clark, V.P., Kiehl, K.A., 2014. Brain potentials measured during a Go/NoGo task predict completion of substance abuse treatment. Biol. Psychiatry 76, 75–83. https://doi.org/10.1016/j. biopsych.2013.09.030. Stemmer, B., Segalowitz, S.J., Dywan, J., Panisset, M., Melmed, C., 2007. The error negativity in nonmedicated and medicated patients with Parkinson’s disease. Clin. Neurophysiol. 118, 1223–1229. https://doi.org/10.1016/j.clinph.2007.02.019. Tanner-Smith, E.E., Tipton, E., Polanin, J.R., 2016. Handling complex meta-analytic data structures using robust variance estimates: a tutorial in R. J. Dev. Life-Course Criminol. 2, 85–112. https://doi.org/10.1007/s40865-016-0026-5. Thurm, F., Antonenko, D., Schlee, W., Kolassa, S., Elbert, T., Kolassa, I.-T., 2013. Effects of aging and mild cognitive impairment on electrophysiological correlates of performance monitoring. J. Alzheimer’s Dis. 35, 575–587. https://doi.org/10.3233/ JAD-121348. Tipton, E., Pustejovsky, J.E., 2015. Small-sample adjustments for tests of moderators and model fit using robust variance estimation in meta-regression. J. Educ. Behav. Stat. 40, 604–634. https://doi.org/10.3102/1076998615606099. Ullsperger, M., von Cramon, D.Y., 2006. The role of intact frontostriatal circuits in error processing. J. Cogn. Neurosci. 18, 651–664. https://doi.org/10.1162/ jocn.2006.18.4.651. Ullsperger, M., von Cramon, D.Y., Müller, N.G., 2002. Interactions of focal cortical lesions with error processing: evidence from event-related brain potentials. Neuropsychology 16, 548–561. https://doi.org/10.1037//0894-4105.16.4.548. Vallet, W., Neige, C., Mouchet-Mages, S., Brunelin, J., Grondin, S., 2021. Responselocked component of error monitoring in psychopathy: a systematic review and meta-analysis of error-related negativity/positivity. Neurosci. Biobehav. Rev. 123, 104–119. https://doi.org/10.1016/j.neubiorev.2021.01.004. van Boxtel, G.J.M., Van Der Molen, M.W., Jennings, J.R., 2005. Differential involvement of the anterior cingulate cortex in performance monitoring during a stop-signal task. J. Psychophysiol. 19, 1. https://doi.org/10.1027/0269-8803.19.1.1. Veen, V. van, Carter, C.S., 2002. The timing of action-monitoring processes in the anterior cingulate cortex. J. Cogn. Neurosci. 14, 593–602. https://doi.org/10.1162/ 08989290260045837. Verleger, R., Schroll, H., Hamker, F.H., 2013. The unstable bridge from stimulus processing to correct responding in Parkinson’s disease. Neuropsychologia 51, 2512–2525. https://doi.org/10.1016/j.neuropsychologia.2013.09.017. Viechtbauer, W., 2010. Conducting meta-analyses in R with the metafor package. J. Stat. Softw. 36, 1–48. https://doi.org/10.18637/jss.v036.i03. Volkow, N.D., Fowler, J.S., Wang, G.-J., 2004. The addicted human brain viewed in the light of imaging studies: brain circuits and treatment strategies. Neuropharmacology 47, 3–13. https://doi.org/10.1016/j.neuropharm.2004.07.019. Volpato, C., Schiff, S., Facchini, S., Silvoni, S., Cavinato, M., Piccione, F., Antonini, A., Birbaumer, N., 2016. Dopaminergic medication modulates learning from feedback and error-related negativity in parkinson’s disease: a pilot study. Front. Behav. Neurosci. 10, 205. https://doi.org/10.3389/fnbeh.2016.00205. Wessel, J.R., 2018. An adaptive orienting theory of error processing. Psychophysiology 55, e13041. https://doi.org/10.1111/psyp.13041. Willemssen, R., Müller, T., Schwarz, M., Hohnsbein, J., Falkenstein, M., 2008. Error processing in patients with Parkinson’s disease: the influence of medication state. J. Neural Transm. 115, 461–468. https://doi.org/10.1007/s00702-007-0842-1. 23 Y. Liu et al. Neuroscience and Biobehavioral Reviews 148 (2023) 105127 Willemssen, R., Müller, T., Schwarz, M., Falkenstein, M., Beste, C., 2009. Response monitoring in de novo patients with Parkinson’s disease. PLoS One 4, e4898. https://doi.org/10.1371/journal.pone.0004898. Xiao, P., Dai, Z., Zhong, J., Zhu, Y., Shi, H., Pan, P., 2015. Regional gray matter deficits in alcohol dependence: a meta-analysis of voxel-based morphometry studies. Drug Alcohol Depend. 153, 22–28. https://doi.org/10.1016/j.drugalcdep.2015.05.030. Yücel, M., Oldenhof, E., Ahmed, S.H., Belin, D., Billieux, J., Bowden-Jones, H., Carter, A., Chamberlain, S.R., Clark, L., Connor, J., 2018. A transdiagnostic dimensional approach towards a neuropsychological assessment for addiction: an international Delphi consensus study. Addiction. https://doi.org/10.1111/add.14424. Zhang, Y., Ou, H., Yuan, T.-F., Sun, J., 2021. Electrophysiological indexes for impaired response inhibition and salience attribution in substance (stimulants and depressants) use disorders: a meta-analysis. Int. J. Psychophysiol. . J. Int. Organ. Psychophysiol. 170, 133–155. https://doi.org/10.1016/j.ijpsycho.2021.10.008. 24