Background: Internalizing disorders (IDs; e.g. posttraumatic stress disorder (PTSD), obsessive-compulsive disorder (OCD)) are the most prevalent form of psychopathology experienced worldwide. Current first-line therapies (e.g. pharmacotherapy) offer high failure rates and substantial side-effects. Closed-loop brain training of electrophysiological signals, also known as electroencephalographic neurofeedback (EEG-NFB), has been shown to be an effective and safe treatment for these conditions, however, there remains much doubt regarding the existence of specificity (i.e. clinical effects specific to the modulation of the EEG variable(s) of interest). This systematic review and meta-analysis was undertaken to determine if there is evidence for EEG-NFB specificity in the treatment of IDs. Methods: We considered all randomised, double-blind, sham/placebo-controlled trials involving humans with at least one ID diagnosis without exclusion by language, locality, ethnicity, age, or sex obtained. We searched various electronic databases and registries including Scopus, PubMed, MEDLINE (Ovid), Cochrane Central Register of Controlled Trials (Ovid), Embase, Allied and Complementary Medicine (Ovid), PsycInfo (Ovid), PsycExtra (Ovid), the World Health Organization's International Clinical Trials Registry Platform (ICTRP), ClinicalTrials.gov, and the Australia New Zealand Clinical Trials Registry (ANZCTR). Searches for all databases & registries were last run on 15 May, 2021. Outcomes of interest included self/parent/teacher reports and clinician ratings. Because the included trials utilized different measurement scales to assess clinical outcomes, we utilised standardised mean differences (SMD) and 95% confidence intervals (CIs) calculated from change-from-baseline (CFB) scores. In addition, a sensitivity analysis was performed utilising post-treatment (PI) scores. When appropriate, we contacted trial authors to obtain additional information. A meta-analysis was performed utilizing inverse variance and random effects modelling with results displayed in a forest plot. Additionally, the risk of bias for individual studies was assessed via the RoB 2 (revised Cochrane risk of bias tool for randomised controlled trials) and the certainty of the pooled results was rated using the GRADE (Grading of Recommendations, Assessment, Development, and Evaluations) framework. Results: Of 705 unique records identified, 17 full-text reports were evaluated and 4 completed trials (n=152 adult participants) met our criteria, however, only 3 (n=102 adult participants) had publicly available data and were included in our meta-analyses. The overall risk of bias for the individual trials included ranged from ‘some concerns’ to ‘high’. The primary analysis favoured genuine EEG-NFB over sham with a pooled SMD = 0.41 [95% CI 0.01, 0.80]; p = 0.04) with no evidence of heterogeneity (Tau2=0.00, Chi² p=0.96; I² = 0%). Similarly, the sensitivity analysis favoured genuine over sham (SMD = 0.36 [95% CI -0.04, 0.75]; p = 0.07). Per the GRADE, the overall certainty in the evidence was deemed ‘very low’. Discussion: Our primary analysis suggests that EEG-NFB does have specific effects in the treatment of IDs as evidenced by our finding of an overall small mean effect size (0.41). Nonetheless, a mean effect size ranging from nil (0.01) to large (0.80) is also reasonably compatible with the data, given our assumptions. Similar results were evident in our sensitivity analysis. Moreover, the certainty of the pooled results was rated as ‘very low’ due to concerns surrounding reporting biases and imprecision. More randomised, double-blind, sham-controlled trials are needed to verify the existence and, if present, degree of EEG-NFB specificity in the treatment of IDs. Registration: This review was registered on the International Prospective Register of Systematic Reviews (PROSPERO; registration number: CRD42020159702).