The Impact of Citicoline/Phosphatidylserine Supplementation on Cognitive Performance and Executive Functions in Mental Disorders: A Systematic Review of Controlled Trials

ABSTRACT

Introduction

Citicoline and phosphatidylserine (PS) have been proposed as potential neuroprotective and cognitive‐enhancing agents; however, evidence of their therapeutic impact across populations with mental disorders remains limited. This systematic review aimed to evaluate the effects of citicoline and PS, administered as monotherapy or adjunctive treatment, on cognitive performance in children, adolescents, and adults with mental disorders.

Methods

A systematic review of 11 included studies was conducted following PRISMA guidelines. Databases searched were PubMed/MEDLINE, Scopus, Web of Science, and Cochrane CENTRAL through October 2025. Ten studies were randomized controlled trials (RCTs), and one was a non‐randomized controlled trial (non‐RCT). Methodological quality and risk of bias were evaluated using the RoB 2 and ROBINS‐I tools.

Results

Five studies focused on children with ADHD, and the remaining studies included adults with mood disorders, alcohol and substance use disorders, and schizophrenia. Regarding ADHD, PS‐containing supplementation produced significant improvements in attention and impulsivity in three studies, whereas two studies found no significant differences. A single study in schizophrenia reported significant improvements in verbal free recall but no global cognitive gains. Results for alcohol and substance use disorders, as well as mood disorders were mixed. Two studies revealed selective benefits in cognitive performance, while the others showed none. Overall, findings were heterogeneous and inconsistent across diagnostic groups.

Conclusion

Given methodological limitations and high risk of bias, current evidence does not support citicoline or PS‐containing compounds as effective cognitive enhancers in mental disorders. Consequently, these agents should be positioned as adjunctive or investigational rather than as evidence‐based treatments.

1. Introduction

Cognitive impairment and executive dysfunction are increasingly recognized as core transdiagnostic features of mental disorders, including schizophrenia, mood disorders, bipolar disorder, and attention‐deficit/hyperactivity disorder (ADHD) (Millan et al. 2012). Problems with attention, working memory, cognitive control, and processing speed often persist even after symptoms improve, and cognitive deficits often persist, limiting functioning and recovery. Current medications primarily target affective, psychotic, or behavioral symptoms but offer only modest and inconsistent cognitive benefits (Bortolato et al. 2016; McCutcheon et al. 2023). This gap has sparked growing interest in adjunctive strategies that might restore the neuronal integrity and biological processes underlying higher‐order cognitive function.

Citicoline (cytidine‐5′‐diphosphocholine; CDP‐choline) is an endogenous compound involved in phospholipid synthesis and neurotransmitter regulation. When metabolized into cytidine and choline, it supports phosphatidylcholine production and helps maintain membrane stability, especially under metabolic stress or neuroinflammation (Secades and Gareri 2022). Studies suggest citicoline influences several neurotransmitter systems relevant to cognition, particularly cholinergic and catecholaminergic pathways, while also affecting glutamatergic excitotoxicity through regulation of excitatory amino acid transporters (Secades and Gareri 2022; Wignall and Brown 2014). In neurological conditions like age‐related cognitive decline, post‐stroke recovery, and mild cognitive impairment, citicoline has been associated with improvements in memory, attention, and executive functioning (Jasielski et al. 2020; Nakazaki et al. 2021; Bonvicini et al. 2023). However, evidence in mental disorders remains limited and inconsistent, with small trials reporting variable effects across different diagnoses (Wignall and Brown 2014; Brown and Gabrielson 2012; Brown et al. 2019; Fornaro et al. 2025; Hübner et al. 2024).

Phosphatidylserine (PS) is a major phospholipid in neuronal membranes and plays a central role in intracellular signaling and stress regulation. Located predominantly on the inner leaflet of the plasma membrane, PS facilitates activation of key signaling molecules like protein kinase C and Akt, which are essential for synaptic plasticity and cell survival (Kim et al. 2014; Glade and Smith 2015). Beyond these structural and signaling functions, PS appears to modulate the hypothalamic–pituitary–adrenal (HPA) axis, dampening stress‐related cortisol elevations implicated in hippocampal dysfunction and cognitive decline (Hellhammer et al. 2014; Baumeister et al. 2008). These properties make PS worth exploring as a supportive intervention in conditions marked by chronic stress and cognitive vulnerability.

Across mental disorders, several overlapping mechanisms have been implicated in persistent executive dysfunction: disrupted membrane phospholipid composition, altered monoaminergic and cholinergic transmission, dysregulated stress responses, and impaired neuroplasticity. In schizophrenia, deficits in prefrontal dopaminergic signaling and white‐matter connectivity closely relate to impairments in cognitive control and working memory (McCutcheon et al. 2023). Mood disorders similarly show enduring cognitive problems linked to stress‐related hippocampal changes and HPA‐axis dysregulation (Keller et al. 2017; Belvederi Murri et al. 2016). In ADHD, maturational delays in fronto‐striatal circuits and catecholaminergic imbalance contribute to difficulties with sustained attention and inhibitory control (Koirala et al. 2024). These shared substrates suggest that interventions targeting membrane integrity and neuromodulatory balance might have transdiagnostic relevance.

Within this context, citicoline and PS have been proposed as adjunctive agents that could support cognitive and executive functioning across populations with mental disorders. Small randomized trials indicate citicoline may yield modest benefits on negative symptoms and certain cognitive domains in schizophrenia, potentially through selective enhancement of mesocortical dopaminergic transmission and cholinergic modulation (Ghajar et al. 2018). In mood disorders, adjunctive citicoline has been associated with faster antidepressant response in selected samples, particularly those with comorbid substance use (Brown and Gabrielson 2012). PS supplementation, alone or combined with omega‐3 fatty acids, has shown potential benefits for impulsivity, emotional regulation, and working memory in pediatric ADHD populations (Manor et al. 2012; Hirayama et al. 2014). Yet these findings remain scattered across diagnostic categories and vary considerably in methodological quality, making it difficult to draw firm conclusions about the magnitude, consistency, and clinical relevance of cognitive effects.

Despite a growing mechanistic rationale and accumulating clinical studies, no comprehensive review has systematically evaluated the effects of citicoline and PS on cognitive performance and executive functions specifically in populations with mental disorders. Existing reviews have largely focused on neurological conditions or healthy aging, while mental disorders—where cognitive dysfunction is both prevalent and functionally disabling—have received less integrative attention. Given the heterogeneity in study designs, outcome measures, and diagnostic groups, a systematic review of controlled trials is warranted.

This systematic review therefore, aims to evaluate controlled clinical trials examining the effects of citicoline and/or PS‐containing supplementation on cognitive performance and executive functions in individuals with mental disorders. Specifically, we seek to: (i) assess the extent to which these interventions influence core cognitive domains, (ii) explore potential disorder‐specific and transdiagnostic effects, and (iii) identify key methodological limitations to inform future research. By synthesizing the available evidence, we aim to provide a balanced assessment of whether membrane‐targeting supplements represent clinically meaningful adjunctive strategies or supportive interventions with limited real‐world impact in psychiatric care.

2. Methods

2.1. Search Strategy and Study Selection Procedure

The current systematic literature review was performed in accordance with the Preferred Reporting Items for Systematic Reviews and Meta‐Analyses Protocols (PRISMA‐P) 2020 guidelines during the development of the review protocol (Moher et al. 2015), and the final review was reported in accordance with the PRISMA 2020 guidelines This study was registered with the PROSPERO database under registration number CRD420251167224 prior to data extraction.

Searches were conducted in several databases, including PubMed/MEDLINE, Scopus, Web of Science, and Cochrane CENTRAL in October 2025 to identify all relevant articles on the topic. The keywords employed for the search were (citicoline OR CDP‐choline OR “cytidine diphosphocholine” OR phosphatidylserine OR “phosphatidyl serine”) AND (cogn* OR memory OR attention OR working memory OR “executive function”). The reference lists of all included studies and relevant reviews were also manually screened for additional eligible articles. The eligibility of studies was independently assessed by two reviewers (NGK, TTK), with any disagreements resolved by a consulting third reviewer (ROC). The detailed search strategy is provided in the Table S1.

2.2. Eligibility Criteria

Studies were selected according to the PICOS framework. The population comprised children, adolescents, and adults diagnosed with any mental disorder. The inclusion of a broad range of mental disorders was not determined a priori. Instead, the diagnoses were included based on the studies that met the predefined eligibility criteria identified through the literature search. The identified diagnoses were subsequently categorized, taking into account the psychiatric comorbidities reported in the included studies. The interventions involved the administration of citicoline or PS, either as monotherapy or as an adjunctive treatment. The comparators included placebo, standard pharmacological treatments, or other interventions. Changes in cognitive function and psychiatric symptom severity were recorded to evaluate the impact of citicoline or PS. Eligible study designs included randomized controlled trials (RCTs), non‐randomized controlled studies, and crossover trials. Only studies published in English were included, with no restrictions on publication date, to ensure a comprehensive synthesis of the available evidence.

Studies without a control group, including single‐arm and pre–post study designs, were excluded. In vitro studies, case reports, animal studies, conference abstracts, poster presentations, systematic reviews, meta‐analyses, traditional reviews, and studies whose full text was inaccessible were also excluded.

2.3. Data Extraction

The screening and management of articles were performed using Rayyan intelligent systematic review software (Ouzzani et al. 2016). Data from the included studies were extracted into a predefined and standardized, pre‐piloted Microsoft Excel spreadsheet. Two reviewers (NGK, TTK) independently extracted the data. The completed Excel forms were compared, and any discrepancies were resolved through consultation with a third reviewer (ES). To ensure consistency and clarity, the data extraction form was pilot‐tested on three randomly selected studies from the included articles, and minor modifications were made based on the findings (Page et al. 2021).

The following data items were extracted from each included study: study characteristics (authors, publication year, country, study design, clinical trial registration number); sample characteristics (sample size, mean age, gender distribution, diagnostic criteria); methodological details (assessments, follow‐up duration); intervention details (dose, duration, frequency, type); and outcomes (cognitive and clinical outcomes, adverse events). Additionally, reported statistical findings (e.g., mean, standard deviation, effect size, and significance levels) were recorded.

2.4. Risk of Bias Assessment

The quality of randomized studies was assessed using the Cochrane Risk of Bias 2 (RoB 2) tool for RCTs (Sterne et al. 2019). The RoB 2 checklist consists of five items evaluating different sources of bias, including the randomization process, deviations from intended interventions, missing outcome data, measurement of outcomes, and selective reporting. Each item was rated as low risk, high risk, or some concerns (Sterne et al. 2019).

The Risk of Bias In Non‐randomized Studies—of Interventions (ROBINS‐I) tool was used for non‐RCTs (Sterne et al. 2016). The ROBINS‐I tool assesses bias across the following domains: confounding, selection of participants into the study, classification of interventions, deviations from intended interventions, missing outcome data, measurement of outcomes, and selection of the reported results. Responses are classified as low, moderate, serious, or critical risk of bias (Sterne et al. 2016).

Risk of bias using both the RoB 2 and ROBINS‐I tools was evaluated independently by two reviewers (N.G.K. and T.T.K.), and any discrepancies were resolved by consensus. The results of the risk of bias assessments were visualized using the robvis web application (McGuinness and Higgins 2021).

2.5. Data Analysis

The selected articles were summarized using a standardized data extraction template, reporting each study's main characteristics and corresponding outcomes (Table 1). Studies were first grouped by diagnostic category, and cognitive outcomes were synthesized within each diagnostic group. Due to substantial heterogeneity across studies in diagnostic groups, interventions, and outcome measures, a quantitative synthesis (meta‐analysis) was not performed; therefore, findings were synthesized narratively. Findings were presented in chronological order of publication.

TABLE 1.

Characteristics of the included studies examining the effects of citicoline and phosphatidylserine in mental disorders.

Author/year	Country	Study design	Participant characteristics	Psychiatric diagnosis	Intervention type and dosage	Duration (week)	Cognitive measurements	Clinical measurements	Main outcomes
Ayuso‐Gutierrez and Saiz‐Ruiz (1982)	Spain	Placebo‐controlled, d.b.	IG: n = 11 31.00 ± 15.14 years 54.5% M PG: n = 11 34.73 ± 11.25 years 27.3% M	Endogen ous depression	Citicoline 800 mg/day, iv Bilateral ECT	8 days	TEA memory test	Recovery time of spontaneous respiration and orientation	No significant group differences in numerical or associative memory reduction after 4 ECTs. No significant differences in post‐ECT respiration and orientation recovery times.
Brown et al. (2007)	USA	PG‐RCT, d.b.	IG: n = 23 42.1 ± 6.6 years 56.5% M PG: n = 21 40.7 ± 8.0 years 47.6% M	Bipolar disorder and cocaine dependence	Citicoline initial 500 mg/day p.o., increased by 500 mg every 2 weeks (max: 2000 mg/day) Concomitant psychotropic medications	12	RAVLT	IDS‐SR, YMRS, PRD‐III somatic symptom scale, and urine drug screens	Significantly higher RAVLT alternative word list scores with citicoline versus placebo (p = 0.006). No significant differences in RAVLT total words recalled, delayed recall, IDS‐SR, and YMRS. Significant reduction in likelihood of cocaine‐positive urine at exit was observed with citicoline (p = 0.026).Significantly lower PRD‐III somatic symptom scale scores were observed with citicoline versus placebo (p < 0.001).
Vaisman et al. (2008)	Israel	PG‐RCT, d.b.	IG: n = 18 9.17 ± 1.27 years 83.3% M PG: n = 21 9.31 ± 1.28 years 71.4% M	ADHD	PS 300 mg/day p.o., EPA‐DHA 250 mg/day, p.o.	8	TOVA, abbreviated CRS, CBCL	Plasma and erythrocyte phospholipid fatty acid profiles (via gas chromatography)	Significantly improved total TOVA scores and executive functions (commission errors, response time, and variability) with PS‐omega‐3 versus placebo (p < 0.001); no effect observed on omission errors. Significant increases in plasma EPA (p < 0.001), DHA (p = 0.012); erythrocyte DPA (p = 0.022), linoleic acid (p = 0.046); reduced erythrocyte very‐long‐chain saturated fatty acids (C22:0 and C24:0; p = 0.046 and p = 0.045, respectively) with PS‐omega‐3. These changes correlated with improved TOVA scores (p < 0.05). No significant differences observed in abbreviated CRS and CBCL scores.
Brown and Gabrielson (2012)	USA	PG‐ RCT, d.b.	IG: n = 28 41.6 ± 9.9 years 46.4% M PG: n = 20 34.0 ± 7.3 years 65% M	Bipolar‐unipolar depression and methamphetamine dependence	Citicoline initial 500 mg/day p.o., increased by 500 mg every 2 weeks (max: 2000 mg/day) Concomitant psychotropic medications	12	HVLT	IDS‐C, self‐reported and urine drug screens	Significantly greater improvement in IDS‐C scores with citicoline versus placebo (p = 0.05). No significant differences observed in methamphetamine use or memory functions.
Manor et al. (2012)	Israel	PG‐ RCT d.b.	IG: n = 100 9.2 ± 2.0 years 72.0% M PG: n = 47 9.2 ± 1.8 years 68.0% M	ADHD	PS‐omega‐3300 mg/day p.o.	15	CRS‐T CRS‐P	CHQ SDQ	Significant reduction in CRS‐P global restless/Impulsive subscale scores was observed with PS‐omega‐3 versus placebo (p = 0.047). Significant improvements were observed with PS‐omega‐3 in CHQ emotional impact on parents (p = 0.022) and family activities subscale scores (p < 0.05).
Deutsch et al. (2013)	USA	PG‐RCT, d.b., d.d	IG: n = 19 54.37 ± 8.50 years 89.5% M PG: n = 24 52.38 ± 11.04 years 99.9% M	Schizophrenia	Galantamine Initial 500 mg/day p.o. for 3 days, then 1000 mg/day for 4 days (max: 2000 mg/day starting week 2) Citicoline Initial 8 mg/day p.o., increased by 8 mg every week (max: 24 mg/day starting week 3) Concomitant second generation antipsychotics	16	MCCB, USC‐REMT	PANSS, CGI, scale of functioning	Significantly improved overall functioning (p = 0.05) and free verbal recall (USC‐REMT multi‐trial free recall) (p = 0.04) with treatment versus placebo. No significant differences observed in negative symptoms, other PANSS symptom factors, MCCB, or CGI scores.
Hirayama et al. (2014)	Japan	PG‐ RCT, d.b.	IG: n = 19 9.1 ± 1.7 years 99,95% M PG: n = 17 8.7 ± 3 years 99.94% M	ADHD	PS 200 mg/day p.o.	8	Digit span test, Go/No‐Go task, ADHD symptoms	—	Short‐term auditory memory significantly improved (p < 0.05), while working memory showed no significant change via digit span test. Significant reductions in inattention and total error counts were observed during differentiation and reverse differentiation phases of the Go/No‐Go task (p < 0.05). Significant improvements were observed in all ADHD domains compared to baseline (p < 0.01); post‐intervention total ADHD and inattention scores were significantly superior to placebo (p < 0.01).
Gruber et al. (2015)	USA	PG‐RCT, d.b.	IG: n = 10 27.70 ± 6.98 years 80% M PG: n = 9 30.00 ± 7.09 years 100% M	Cannabis abuse or dependence	Citicoline 2000 mg/day p.o.	8	MSIT, stroop color‐word test, BIS‐11	fMRI Participants' cannabis‐use diary	Significantly smaller impulsive speeding in reaction time (MSIT) versus placebo (p = 0.02). Significantly reduced attentional impulsivity (BIS‐11 attention subscale) versus placebo (p = 0.02). Shifted ACC activation from posterior to anterior regions and increased neural efficiency during cognitive control tasks (p < 0.05). Significant decreases in craving and frequency of use observed over time, but no significant difference versus placebo.
Brown et al. (2019)	USA	PG‐ RCT, d.b.	IG: n = 29 49.03 ± 8.36 years 75.9% M PG: n = 26 43.61 ± 9.3 year 80.8% M	Alcohol use disorder	Citicoline initial 500 mg/day p.o., increased by 500 mg every 2 weeks (max: 2000 mg/day)	12	RAVLT, stroop color‐word test, TMT, RMCPT, SMT	TLFB, PACS, IDS‐SR, PRD‐III somatic symptom scale	No significant differences were observed between groups in alcohol use, craving, cognitive function, and depressive symptoms.
Rheims et al. (2024)	France	PG‐RCT, d.b.	IG: n = 44 10.3 ± 2.7 years 79.0% M PG: n = 30 9.6 ± 2.1 years 73.0% M	ADHD + epilepsy	PS‐omega‐3300 mg/day p.o. Concomitant antiseizure medications	12	TOVA, ADHD‐ RS‐IV	CDI, revised Children'sManifest anxiety scale, EFIQUACEE quality of life	ADHD‐RS‐IV (inattentive subscore and total), TOVA, and EFIQUACEE scores did not differ significantly between groups.
Hübner et al. (2024)	Brazil	CO‐RCT d.b.	IG: n = 13 9.0 ± 1.87 years 69.2% M PG: n = 14 9.7 ± 1.63 years 92.8% M	ADHD	Citicoline 250 mg/day p.o. Concomitant psychotropic medications	8	WASI vocabulary/matrix reasoning subtests, digit subtest, CPT	—	No significant differences were observed between the use of citicoline and placebo in cognitive function.

Abbreviations: ADHD‐RS‐IV, ADHD Rating Scale IV; ADHD, Attention‐deficit/hyperactivity disorder; BIS, Barratt Impulsiveness Scale; CBCL, Child Behavior Checklist; CDI, Children's Depression Inventory; CGI, Clinical Global Impression; CHQ, Child Health Questionnaire; CPT, Continuous Performance Test; CRS, Conners' Rating Scale; CRS‐P, CRS‐Parent; CRS‐T, CRS‐Teacher; d.b., double blind; d.d., double dummy; DPA, Docosapentaenoic acid; ECT, Electroconvulsive Therapy; EPA‐DHA, Eicosapentaenoic‐docosahexaenoic acid; fMRI, functional Magnetic Resonance Imaging; HVLT, Hopkins Verbal Learning Test; IDS‐C, IDS‐Clinician version; IDS‐SR, Inventory of Depressive Symptomatology—Self Report; IG, Intervention group; iv, intravenous; M, Males; MCCB, MATRICS Consensus Cognitive Battery; mg, milligram; MSIT, Multi‐Source Interference Task; p.o., peroral; PACS, Penn Alcohol Craving Scale; PANSS, Positive and Negative Syndrome Scale; PG, Placebo group; PG‐RCT, Parallel group—randomized controlled trial; PRD‐III, Psychobiology of Recovery from Depression‐III; PS, Phosphatidylserine; RAVLT, Rey Auditory Verbal Learning Test; RMCPT, Running Memory CPT; SDQ, Strengths and Difficulties Questionnaire; SMT, Sternberg Memory Task; TEA, Test of Everyday Attention; TLFB, Timeline Followback; TMT, Trail Making Test; TOVA, Test of Variables of Attention; USC‐REMT, University of Southern California‐Repeatable Episodic Memory Test; WASI, Wechsler Abbreviated Scale of Intelligence; YMRS, Young Mania Rating Scale.

3. Results

3.1. Study Selection

The flow of study selection is shown in the PRISMA diagram (Figure 1). Initially, a total of 2489 records were identified from electronic databases. After removing duplicates, 1480 records were screened by title and abstract. Of these, 1392 were excluded due to irrelevance. A total of 78 full‐text articles were assessed for eligibility. Subsequently, 67 articles were excluded based on specific criteria (Figure 1). Finally, 11 studies met the inclusion criteria and were included in the systematic review. A list of excluded studies and the specific reasons for their exclusion are provided in the Table S2.

FIGURE 1.

PRISMA flow diagram of the study selection process.

3.2. Study Characteristics

A total of 11 studies were included in the review, all of which were placebo‐controlled and double‐blind. Of these, 10 were randomized controlled trials (RCTs) (Brown and Gabrielson 2012; Brown et al. 2019; Hübner et al. 2024; Manor et al. 2012; Hirayama et al. 2014; Vaisman et al. 2008; Gruber et al. 2015; Deutsch et al. 2013; Brown et al. 2007; Rheims et al. 2024), and one was a non‐randomized controlled trial (Ayuso‐Gutierrez and Saiz‐Ruiz 1982). Regarding the design of the RCTs, nine utilized a parallel‐group design (Brown and Gabrielson 2012; Brown et al. 2019; Manor et al. 2012; Hirayama et al. 2014; Vaisman et al. 2008; Gruber et al. 2015; Deutsch et al. 2013; Brown et al. 2007; Rheims et al. 2024), while one employed a crossover design (Hübner et al. 2024). Notably, the study by (Deutsch et al. 2013) also utilized a double‐dummy design. Five studies were conducted in children and adolescents (Hübner et al. 2024; Manor et al. 2012; Hirayama et al. 2014; Vaisman et al. 2008; Rheims et al. 2024), and six studies involved adult population (Brown and Gabrielson 2012; Brown et al. 2019; Gruber et al. 2015; Deutsch et al. 2013; Brown et al. 2007; Ayuso‐Gutierrez and Saiz‐Ruiz 1982).

Participant ages ranged from 4 (Hirayama et al. 2014) to 16 years (Rheims et al. 2024) in pediatric samples, and from 18 years (Brown and Gabrielson 2012; Brown et al. 2019; Gruber et al. 2015; Deutsch et al. 2013; Brown et al. 2007) to 75 years (Brown et al. 2019) in adults. Mean ages in the treatment groups ranged from 9.1 (Hirayama et al. 2014) to 54.37 years (Deutsch et al. 2013), while in the placebo groups, they ranged from 8.7 (Hirayama et al. 2014) to 52.38 years (Deutsch et al. 2013).

The total number of randomized participants was 701, including 420 children/adolescents and 281 adults. Sample sizes across the included studies varied significantly, ranging from 19 (Gruber et al. 2015) to 200 participants (Manor et al. 2012). In the analyzed sample, the gender distribution was 72.9% male and 27.1% female. Among children and adolescents, 76.5% were male, and 23.5% were female, whereas in the adult population, 68% were male and 32% were female.

3.2.1. Clinical Diagnosis

The included studies comprised participants from various diagnostic groups, primarily focusing on neurodevelopmental disorders. (Vaisman et al. 2008; Hirayama et al. 2014; Manor et al. 2012; Hübner et al. 2024) studied individuals with ADHD, and (Rheims et al. 2024) focused on ADHD with comorbid epilepsy. Following neurodevelopmental disorders, alcohol and substance use disorders were a major area of investigation. (Brown and Gabrielson 2012) evaluated methamphetamine dependence comorbid with bipolar or unipolar depression, while (Brown et al. 2007) focused on bipolar disorder with cocaine dependence. Furthermore, (Brown et al. 2019; Gruber et al. 2015) addressed alcohol use disorder and cannabis abuse/dependence, respectively. Regarding other diagnoses, (Deutsch et al. 2013) included patients with schizophrenia or schizoaffective disorder, and (Ayuso‐Gutierrez and Saiz‐Ruiz 1982) included those with endogenous depression.

3.2.2. Assessment

3.2.2.1. Neurocognitive Tests

To evaluate the efficacy of citicoline and PS on cognitive performance, the included studies utilized a wide range of neuropsychological batteries and computerized tasks. The assessment tools were categorized into six primary domains: memory, attention, working memory, impulsivity, cognitive flexibility, and processing speed. For instruments measuring more than one construct, outcomes were classified according to the authors' original description and their most common neuropsychological use. During reporting, when several outcomes were available for the same domain, the primary cognitive outcome of the study was emphasized. Objective performance‐based measures were reported first, followed by results derived from clinical symptom scales. This approach was adopted to ensure reporting consistency across the included studies and make the findings easier to follow, regardless of the statistical significance of the findings.

Memory performance was among the most frequently assessed domains, evaluated in six studies (Brown and Gabrielson 2012; Brown et al. 2019; Hirayama et al. 2014; Deutsch et al. 2013; Brown et al. 2007; Ayuso‐Gutierrez and Saiz‐Ruiz 1982). Within this domain, verbal memory and learning were measured using the Rey Auditory Verbal Learning Test (RAVLT) (Brown et al. 2019; Brown et al. 2007), Hopkins Verbal Learning Test‐Revised (HVLT‐R) (Brown and Gabrielson 2012; Deutsch et al. 2013), and University of Southern California‐Repeatable Episodic Memory Test (USC‐REMT) (Deutsch et al. 2013). Visual memory performance was evaluated via the Brief Visuospatial Memory Test‐Revised (BVMT‐R) (Deutsch et al. 2013). Additional instruments included the Sternberg Memory Task (SMT) (Brown et al. 2019), the Memory subtest of the Test of Everyday Attention (TEA) (Ayuso‐Gutierrez and Saiz‐Ruiz 1982), and the Digit Span Test of the Wechsler Intelligence Scale for Children (WISC‐III) to assess short‐term auditory memory (Hirayama et al. 2014).

Attention was also commonly assessed in six studies (Brown et al. 2019; Hübner et al. 2024; Hirayama et al. 2014; Vaisman et al. 2008; Deutsch et al. 2013; Rheims et al. 2024). Sustained and selective attention were measured primarily through continuous performance paradigms, specifically the Continuous Performance Test (CPT) (Hübner et al. 2024), Running Memory CPT (RMCPT) (Brown et al. 2019), CPT‐Identical Pairs (CPT‐IP) (Deutsch et al. 2013), and the Test of Variables of Attention (TOVA) (Vaisman et al. 2008) (Rheims et al. 2024). In one study, TOVA scores were additionally used as an inclusion criterion, requiring baseline performance below −1.8 standard deviations (Vaisman et al. 2008). Other measures included the Go/No‐Go Task (Hirayama et al. 2014), Stroop Color‐Word Test (Brown et al. 2019), and the Trail Making Test (TMT) (Brown et al. 2019).

Impulsivity and response inhibition were specifically addressed in three studies (Hirayama et al. 2014; Vaisman et al. 2008; Gruber et al. 2015). Assessment tools employed for these domains included the Go/No‐Go Task (Hirayama et al. 2014), Stroop Color‐Word Test (Gruber et al. 2015), and the Multi‐Source Interference Task (MSIT) (Vaisman et al. 2008). Additionally, commission errors derived from the TOVA were used as an index of impulsivity in one study (Vaisman et al. 2008).

Working memory capacity was examined in two studies (Hirayama et al. 2014; Deutsch et al. 2013) using span‐based tasks across verbal and spatial modalities. These included Letter‐Number Span (Deutsch et al. 2013), the Spatial Span subtest of the Wechsler Memory Scale (WMS‐III) (Deutsch et al. 2013), and the Digit Span Test of the WISC‐III (Hirayama et al. 2014).

Additional executive domains were assessed less frequently. Cognitive flexibility was measured using the Digit Subtest of the Wechsler Scale (Hübner et al. 2024), while processing speed was specifically evaluated via the Trail Making Test Part A (TMT‐A) and Brief Assessment of Cognition in Schizophrenia (BACS) Symbol Coding (Deutsch et al. 2013).

3.2.2.2. Psychometric Scales

ADHD symptoms were measured using the ADHD Rating Scale IV (ADHD‐RS‐IV) (Rheims et al. 2024) and the Conners' Parent and Teacher Rating Scales (CRS‐P and CRS‐T) (Manor et al. 2012). Mood symptoms were assessed with the Inventory of Depressive Symptomatology (IDS) (Brown and Gabrielson 2012; Brown et al. 2019; Brown et al. 2007), and the Young Mania Rating Scale (YMRS) (Brown et al. 2007). Behavioral assessment was conducted using the Abbreviated Conners Rating Scale (Vaisman et al. 2008), the Child Behavior Checklist (CBCL) (Vaisman et al. 2008), and the Strengths and Difficulties Questionnaire (SDQ) (Manor et al. 2012). In one study, the Positive and Negative Syndrome Scale (PANSS) was used to examine schizophrenia symptoms (Deutsch et al. 2013). Impulsivity was assessed by Gruber et al. (2015) (Gruber et al. 2015) via the Barratt Impulsiveness Scale (BIS), a self‐report instrument including attention, motor, non‐planning, and total impulsivity subscales. Additionally, (Rheims et al. 2024) employed the Children's Depression Inventory (CDI) and the Revised Children's Manifest Anxiety Scale to investigate the presence of comorbid mental diagnoses throughout the treatment process. In addition to serving as an inclusion criterion, the Clinical Global Impression (CGI) was also used for clinical assessment in one study (Deutsch et al. 2013). Consequently, (Brown et al. 2019) employed the Timeline Followback (TLFB) to examine whether citicoline could reduce alcohol consumption, and the Penn Alcohol Craving Scale (PACS) to evaluate its potential impact on alcohol craving.

Among the 11 included studies, the impact of citicoline or PS supplementation on quality of life was assessed in two studies. These studies employed the EFIQUACEE quality of life score (Rheims et al. 2024) and the Child Health Questionnaire (CHQ) (Manor et al. 2012) as assessment instruments. Additionally, impairment in everyday functioning was evaluated in one study using the Scale of Functioning (Deutsch et al. 2013). Quality of life and everyday functioning were not pre‐specified outcomes and were included as exploratory outcomes. Treatment‐related safety was assessed in two studies, with side effects measured via the Psychobiology of Recovery from Depression‐III (PRD‐III) Somatic Symptom Scale (Brown et al. 2019; Brown et al. 2007).

Furthermore, specific scales and methodologies were employed to define inclusion criteria in three studies. (Deutsch et al. 2013) recruited participants who scored at least four on one or more of the five negative‐symptom items of the PANSS. Using the TLFB method, (Brown et al. 2019) included individuals who reported consuming at least 28 drinks per week and having at least seven heavy drinking days within the past 28 days (≥ 4 drinks/day for women, and ≥ 5 drinks/day for men). Finally, (Manor et al. 2012) included participants with scores at least 1.5 standard deviations above age‐ and sex‐adjusted normative values on the teacher‐rated ADHD‐RS‐IV, alongside a score of four or higher on the Clinical Global Impression–Severity (CGI‐S) scale.

3.2.3. Interventions

The interventions implemented in the 11 included studies exhibited diversity in terms of supplements, routes of administration, and formulations. Seven of the studies used citicoline as the intervention supplement (Brown and Gabrielson 2012; Brown et al. 2019; Hübner et al. 2024; Gruber et al. 2015; Deutsch et al. 2013; Brown et al. 2007; Ayuso‐Gutierrez and Saiz‐Ruiz 1982). In one of these studies (Deutsch et al. 2013), citicoline was administered in combination with galantamine, whereas in another (Ayuso‐Gutierrez and Saiz‐Ruiz 1982), it was given alongside bilateral electroconvulsive therapy (ECT). In three of the remaining four studies, omega‐3‐enriched PS (PS‐omega‐3) was administered (Manor et al. 2012; Vaisman et al. 2008; Rheims et al. 2024), whereas soy‐derived PS was used in the remaining single study (Hirayama et al. 2014). In five studies, supplements were used as add‐on therapy to psychotropic medications (Hübner et al. 2024; Vaisman et al. 2008; Gruber et al. 2015; Brown et al. 2007; Rheims et al. 2024).

Although their mechanisms differ, phosphatidylserine and citicoline are considered complementary because phosphatidylserine is related to membrane structure and signaling (Kim et al. 2014; Glade and Smith 2015), whereas citicoline contributes to membrane phospholipid synthesis and cholinergic metabolism (Secades and Gareri 2022). Both compounds have been associated with potential effects in overlapping cognitive domains, including memory, attention, and executive function. Therefore, they were considered within the same review framework, but they were not treated as pharmacologically interchangeable. Similarly, soy‐derived phosphatidylserine and omega‐3‐enriched phosphatidylserine were included as clinically evaluated PS‐based interventions; however, they were described separately and were not assumed to be equivalent in composition or expected biological effects.

Regarding the routes of administration, supplements were delivered orally in 10 of the included studies (Brown and Gabrielson 2012; Brown et al. 2019; Hübner et al. 2024; Manor et al. 2012; Hirayama et al. 2014; Vaisman et al. 2008; Gruber et al. 2015; Deutsch et al. 2013; Brown et al. 2007; Rheims et al. 2024). Only one study (Ayuso‐Gutierrez and Saiz‐Ruiz 1982) employed the parenteral route, in which citicoline was administered via injection within 100 cc of physiological saline. Orally administered interventions were carried out in various formulations. In three studies (Hübner et al. 2024; Manor et al. 2012; Rheims et al. 2024), the supplement was administered in capsule form, in one study (Brown et al. 2007) as a tablet, and in another (Hirayama et al. 2014) as cocoa‐flavored chewable forms. Additionally, in one study (Vaisman et al. 2008), the supplements were presented as an emulsified, dairy chocolate‐flavored spread formulation. The remaining four studies did not report specific information regarding the formulation (Brown et al. 2019; Hübner et al. 2024; Gruber et al. 2015; Deutsch et al. 2013), and five of the 11 studies (Manor et al. 2012; Hirayama et al. 2014; Vaisman et al. 2008; Deutsch et al. 2013; Rheims et al. 2024) explicitly stated that the placebos were identical in appearance to the intervention.

Regarding the dosing strategies, seven studies utilized a fixed‐dose regimen (Hübner et al. 2024; Manor et al. 2012; Hirayama et al. 2014; Vaisman et al. 2008; Gruber et al. 2015; Rheims et al. 2024; Ayuso‐Gutierrez and Saiz‐Ruiz 1982), while four studies (Brown and Gabrielson 2012; Brown et al. 2019; Deutsch et al. 2013; Brown et al. 2007) employed dose titration throughout the treatment period.

The duration of the interventions ranged from 8 days (Ayuso‐Gutierrez and Saiz‐Ruiz 1982) to 16 weeks (Deutsch et al. 2013). Regarding specific protocols (Rheims et al. 2024), implemented a 12‐week double‐blind phase followed by a 12‐week open‐label extension, while (Manor et al. 2012) implemented a 15‐week double‐blind phase followed by a 15‐week open‐label extension. In the crossover study by (Hübner et al. 2024), participants completed two 4‐week treatment periods, separated by a 4‐week washout phase.

3.3. Risk of Bias

Risk of bias was assessed using the RoB 2 tool for 10 RCTs and ROBINS‐I for one non‐randomized study. Overall, the RoB 2 assessment indicated that 80.0% (n = 8) of the RCTs were at high risk and 20.0% (n = 2) were at low risk. Regarding bias in the selection of participants, 40.0% (n = 4) of the studies were assessed as low risk, 50.0% (n = 5) as having some concerns, and 10.0% (n = 1) as high risk of bias. For the risk of bias due to deviations from intended interventions, 50.0% (n = 5) of the studies presented a low risk, 10.0% (n = 1) had some concerns, and 40.0% (n = 4) were rated as high risk. Regarding missing outcome data, 40.0% (n = 4) were at low risk and 60.0% (n = 6) at high risk of bias. All included studies (100%, n = 10) were judged to be at low risk of bias in terms of measurement of the outcome. For bias in the selection of the reported results, 30.0% (n = 3) of studies were at low risk, 40.0 (n = 4) had some concerns, and 30.0% (n = 3) were at high risk of bias. The RoB 2 assessment results for each domain are summarized in Figure 2.

FIGURE 2.

Risk of bias assessment of the included RCTs.

The single non‐randomized study (Ayuso‐Gutierrez and Saiz‐Ruiz 1982) was assessed using ROBINS‐I and was found to have a serious overall risk of bias. Domain 1 (confounding) showed a serious risk of bias, whereas the other domains were rated as low risk. The detailed assessment for each domain is summarized in Figure 3.

FIGURE 3.

Risk of bias assessment of the non‐randomized study.

3.4. Description of the Study Results: Summary of Outcomes

The clinical outcomes of the 11 included studies were classified based on diagnostic categories to evaluate the therapeutic impact of citicoline and PS across different mental disorders. Given the substantial heterogeneity in interventions and outcome measures, the narrative synthesis followed a consistent hierarchical approach. Findings were first grouped by diagnostic category and then described according to the intervention examined, namely citicoline, PS or PS‐containing supplementations. Within each diagnostic category, objective performance‐based cognitive outcomes were reported first, followed by clinical symptom ratings and other relevant outcomes, such as functioning, substance use measures, quality of life, or tolerability.

3.4.1. ADHD

The cognitive and clinical efficacy of citicoline and PS‐containing supplementations in ADHD was evaluated across five studies (Hübner et al. 2024; Manor et al. 2012; Hirayama et al. 2014; Vaisman et al. 2008; Rheims et al. 2024) involving pediatric populations, with one study (Rheims et al. 2024) including a cohort with comorbid epilepsy. Three studies evaluated the impact of PS‐omega‐3 supplementation (Manor et al. 2012; Vaisman et al. 2008; Rheims et al. 2024), while one study investigated isolated soy‐derived PS supplementation (Hirayama et al. 2014). The remaining study focused on citicoline (Hübner et al. 2024). Notably, significant improvements in cognitive performance were associated with PS‐containing interventions; however, these effects were domain‐specific and differed between objective neurocognitive tests and clinical symptom scales.

Regarding objective neurocognitive test outcomes studies using PS‐omega‐3 supplementation (Vaisman et al. 2008) and isolated soy‐derived PS supplementation (Hirayama et al. 2014) reported statistically significant improvements in attention and impulsivity, while statistically significant improvements in memory were observed in the isolated soy‐derived PS study (Hirayama et al. 2014). Regarding clinical symptom scales, one study using PS‐omega‐3 supplementation (Manor et al. 2012) reported a statistically significant improvement in parent‐rated restless/impulsive symptoms. However, two of the five studies (Hübner et al. 2024; Rheims et al. 2024) found no significant differences in cognitive function.

In the Vaisman et al. study (2008), a significant improvement was observed in total TOVA score, which assesses visual sustained attention performance and inattention, in the group treated with PS‐omega‐3 supplementation compared to baseline (p < 0.001). Compared to placebo, PS‐omega‐3 supplementation resulted in statistically significant enhancements, specifically reducing errors of commission as a measure of impulsivity (p = 0.050), improving response time as an index of information processing speed (p = 0.031), and decreasing response time variability, which reflects attention consistency (p = 0.003). However, the treatment did not yield a statistically significant difference compared to placebo regarding errors of omission as a measure of inattention (p = 0.125). Furthermore, 61% of the children in the PS‐omega‐3 supplementation group achieved clinical normalization with scores returning to the normal reference range, a rate significantly higher than the 14% observed in the placebo group (p = 0.004).

Regarding subjective behavioral assessments, no significant differences were demonstrated between the treatment groups. While the parental Abbreviated Conners Rating Scale indicated significant improvements from baseline across all groups, no specific treatment effect was found for PS‐omega‐3 supplementation (p > 0.05). Similarly, minimal and insignificant changes from baseline values were noted by the parents in the subscale of the CBCL (p > 0.05) (Vaisman et al. 2008).

Hirayama et al. (2014) reported that, based on the WISC‐III test, no significant difference was observed in working memory (Digit Span Backward) scores between pre‐ and post‐intervention measurements (p = 0.55). In contrast, short‐term auditory memory (Digit Span Forward) performance significantly improved in the isolated soy‐derived group over time (p = 0.03), whereas no similar improvement was observed in the placebo group (p = 0.47).

Regarding the Go/No‐Go task results, the soy‐derived PS group showed significant improvements over time in inattention scores during both the differentiation (p = 0.04) and reverse differentiation (p = 0.03) tests, as well as significant reductions in the total number of errors (inattention plus impulsivity) in the reverse differentiation test (p = 0.03). When aggregating both tasks, the soy‐derived PS group significantly reduced both total inattention errors (p = 0.02) and overall total errors (p = 0.02); in contrast, the placebo group did not demonstrate improvement in any of these measurements over time (p > 0.05; range p = 0.20–0.83) (Hirayama et al. 2014).

In post‐intervention comparisons between groups, the soy‐derived group performed significantly better than the placebo group on inattention (p = 0.03) and total error (p = 0.01) scores in the reverse differentiation task. Similarly, the PS group outperformed the placebo group on both the differentiation and reverse differentiation tasks (p = 0.04 and p = 0.03, respectively) (Hirayama et al. 2014).

Based on the DSM‐IV‐TR diagnostic criteria, the soy‐derived group exhibited a significant reduction in total ADHD, Inattention (AD), and Hyperactivity/Impulsivity (HD) scores over time (p < 0.01 for all subscales). Conversely, no significant changes were observed in the placebo group (ADHD: p = 0.53; AD: p = 1.00; HD: p = 0.56). Post‐intervention, the soy‐derived PS group demonstrated significantly better clinical scores compared to the placebo group in terms of total ADHD (p < 0.01) and Inattention (AD) (p < 0.01) (Hirayama et al. 2014).

In the Manor et al. study (2012), the main outcome was a significant reduction in the CRS‐P Global Restless/Impulsive subscale (p = 0.047).

3.4.2. Comorbid Mood Disorders and Substance Use Disorders

Two studies (Brown and Gabrielson 2012; Brown et al. 2007) examined the effects of citicoline in participants diagnosed with comorbid mood disorders and substance use disorders. Regarding objective neurocognitive test outcomes, the studies examining the effects of citicoline reported differing results on memory tasks. In one study (Brown et al. 2007), citicoline was associated with a significant improvement in specific verbal learning tasks, particularly the RAVLT Alternative Word List; however, this effect did not extend to total words recalled (p = 0.439) or delayed recall (p = 0.105). Another study (Brown and Gabrielson 2012) using the HVLT found no significant differences with citicoline (p > 0.05).

Regarding clinical symptom scales assessing affective symptomatology, citicoline supplementation was associated with a significantly greater improvement in clinician‐rated depressive scores (IDS) compared to placebo in one study (p = 0.05) (Brown et al.); however, no significant differences were observed in another study (p > 0.05) (Brown et al. 2007).

3.4.3. Alcohol Use Disorder and Substance Use Disorder

Of the two studies included in this section, one investigated the effects of citicoline in participants diagnosed with alcohol use disorder (Brown et al. 2019), while the other examined its effects in participants diagnosed with cannabis abuse/dependence (Gruber et al. 2015). The outcomes of these studies demonstrated a significant impact on behavioral impulsivity in cannabis users, whereas findings related to cognitive performance in the alcohol‐dependent population were not statistically significant.

The administration of citicoline in alcohol use disorder did not yield significant cognitive or clinical improvements (Brown et al. 2019). Specifically, no significant differences were observed between the citicoline and placebo groups in alcohol consumption, craving, cognitive function, or depressive symptoms (Brown et al. 2019).

In the study examining cannabis abuse/dependence, objective cognitive performance on the MSIT differed between groups. The placebo group demonstrated a greater change with significantly faster reaction times, suggesting impulsivity (p = 0.02). In contrast, the citicoline group exhibited a smaller change, performing slightly more slowly. Additionally, citicoline showed a greater percent improvement in accuracy on both the MSIT interference task and the interference condition of the Stroop Color‐Word Test, although these findings did not reach statistical significance (Gruber et al. 2015). Regarding clinical symptom scales, citicoline significantly reduced attentional impulsivity (measured via the BIS‐11 Attention subscale) compared to placebo (p = 0.02), although improvements in total impulsivity and other subscores did not reach statistical significance (Gruber et al. 2015).

Regarding other substance use‐related outcomes, Gruber et al. also (2015) noted that while both the citicoline and placebo groups reported significant decreases in marijuana craving, frequency of use, and amount consumed from baseline to study exit, there were no statistically significant differences between the groups.

Two studies included participants with comorbid mood disorders and substance use disorders: bipolar or unipolar depression with methamphetamine dependence (Brown et al.), and bipolar disorder with cocaine dependence (Brown et al. 2007). In these comorbid samples, the effects of citicoline on substance use outcomes were mixed. One study reported no significant between‐group differences in methamphetamine use outcomes (p = 0.23) (Brown and Gabrielson 2012), whereas another study found a significant reduction in the likelihood of cocaine‐positive urine at study exit with citicoline (p = 0.026) (Brown et al. 2007).

3.4.4. Mood Disorders

Only one study evaluated the effects of citicoline in patients with mood disorders (Ayuso‐Gutierrez and Saiz‐Ruiz 1982). Regarding objective neurocognitive outcomes, (Ayuso‐Gutierrez and Saiz‐Ruiz 1982) showed that citicoline was not effective on cognitive functions, including memory scores, and orientation in patients diagnosed with endogenous depression undergoing ECT. According to the TEA memory test, the numerical memory score after the second ECT session was the same as before treatment, while a slight decrease was observed in both groups after the fourth session. The reduced capacity to remember digits was practically identical in both groups. Regarding associative memory, scores were higher in both groups after the second ECT but noticeably lower after the fourth session. Although losses in the associative memory tests were numerically higher in the placebo group (6.29) compared to the citicoline group (3.08), this difference was not statistically significant.

3.4.5. Schizophrenia

This systematic review included one study evaluating patients with schizophrenia (Deutsch et al. 2013). In the study by (Deutsch et al. 2013), in which citicoline was investigated, no significant treatment‐by‐time effects were observed in any cognitive domains assessed by the MATRICS Consensus Cognitive Battery (MCCB), including symbol coding, attention/vigilance, working memory, and verbal or visuospatial learning. Furthermore, while yes‐no recognition parameters showed no change, a significant treatment effect was observed for total recall across trials in the multi‐trial free recall measure of the USC‐REMT (p = 0.04), with a trend toward significance for forced‐choice recognition memory (p = 0.08).

Regarding clinical symptom scales, no statistically significant differences were observed between groups regarding the PANSS negative symptom factor or other clinical measures, including PANSS positive scores, total scores, and CGI. In terms of functioning, the treatment group exhibited a significant increase in the global score of the Scale of Functioning (p = 0.05).

4. Discussion

This systematic review examined the effects of citicoline and PS‐containing interventions in children, adolescents, and adults with various mental disorders. Although the included studies encompassed a broad spectrum of mental disorders, the effects of these interventions were synthesized within a transdiagnostic framework, focusing on shared cognitive domains. Nevertheless, recognizing the heterogeneity of mental disorders, the findings were also discussed within specific diagnostic categories to preserve disorder‐specific context. These agents were evaluated as monotherapy or adjunctive treatments with respect to cognitive outcomes, psychiatric symptom severity, and tolerability.

Overall, the available evidence suggests that the cognitive effects of citicoline and PS‐containing interventions are variable and context‐dependent. While some studies reported modest improvements in specific domains such as attention or impulsivity, these findings were not consistently replicated across trials or cognitive domains. In several cases, statistically significant results emerged primarily from secondary analyses or restricted subgroups, limiting their generalizability. Additionally, two controlled trials did not demonstrate cognitive benefits relative to placebo. Taken together, these findings indicate that current evidence for cognitive efficacy remains limited and heterogeneous.

4.1. ADHD

In this systematic review, the effects of citicoline and PS‐containing supplemantations on cognitive and clinical outcomes in ADHD samples were evaluated across five controlled studies. Three studies examined PS‐omega‐3, one study examined soy‐derived PS, and one study examined citicoline. Overall, significant findings in cognitive performance domains, particularly attention and impulsivity, were mainly driven by the PS‐containing studies, whereas two studies, one in a comorbid epilepsy sample and the other evaluating citicoline, did not demonstrate significant effect on cognitive abilities, despite the use of an appropriate dosage. However, the significant findings regarding attention and impulsivity reported in three studies (Manor et al. 2012; Hirayama et al. 2014; Vaisman et al. 2008) should be interpreted with caution because these studies were judged to have a high overall risk of bias, reducing confidence in the conclusions. In addition, because most ADHD studies used PS‐omega‐3 formulations, the individual contribution of PS and omega‐3 fatty acids could not be disentangled (Verche et al. 2018; Zhang et al. 2020).

PS is a major phospholipid in the brain, predominantly localized within the neuronal plasma membrane (Lewis et al. 2021). It plays a crucial role in intracellular signaling, neuronal plasticity, synaptic transmission, and neurotransmission (Ma et al. 2022). Instead of being a direct structural component, citicoline serves as a precursor to phosphatidylcholine, a primary constituent of the neuronal membrane. These phospholipid‐related processes are relevant in conditions involving neurodegeneration, chronic stress exposure, or aging‐related decline (Long et al. 2026), but ADHD is generally conceptualized as a neurodevelopmental condition in which the underlying pathophysiology may be less closely related to primary membrane disruption or neurotoxic injury. Therefore, one possible explanation for the inconsistent findings is that improving phospholipid availability alone may be insufficient to substantially affect executive network functioning in ADHD, a condition in which neurodevelopment is influenced by multiple genetic and activity‐dependent mechanisms (Armide and Babaei 2025).

Improvements were detected on performance‐based measures and task indices of impulsivity and attention consistency, whereas memory effects were more selective and less consistently replicated. As a promising objective cognitive finding, the study by (Vaisman et al. 2008) showed improvements in several TOVA outcomes, with a high rate of clinical normalization following PS–omega‐3 supplementation. However, these findings should be interpreted with caution because the study had a relatively small sample size and evaluated PS in combination with omega‐3 fatty acids, making it difficult to attribute the observed effects solely to PS.

When considered collectively, the included studies suggest a discrepancy between objective cognitive gains and more variable changes in subjective behavior ratings. In some trials, standardized symptom scores showed clinically meaningful improvement relative to placebo, while in others, parent‐reported scales improved non‐specifically across groups without a clear PS‐related effect. Taken together, the inconsistency across outcomes, methodological limitations, and the limited use of isolated PS formulations preclude firm conclusions regarding the cognitive and clinical efficacy of PS‐containing supplements. Consequently, larger, well‐designed randomized controlled trials with a low risk of bias and standardized outcome measures are needed to better characterize PS‐related cognitive and behavioral effects.

4.2. Alcohol Use Disorder and Substance Use Disorder

Studies involving alcohol and substance use disorders (n = 4) exclusively examined the effects of citicoline. Two studies included samples with comorbid mood disorders, and one specifically focused on alcohol use disorder. Overall, cognitive findings were heterogeneous and appeared to vary across different substance examined.

With respect to cognitive outcomes, improvements were observed in citicoline‐treated individuals using cocaine (Brown et al. 2007) and cannabis (Gruber et al. 2015), particularly in domains such as attention and verbal memory. However, these improvements should be interpreted with caution, as both studies were judged to have a high overall risk of bias, primarily driven by missing outcome data. In contrast, no cognitive benefits were detected in individuals with methamphetamine dependence (Brown and Gabrielson 2012). Given that methamphetamine is substantially more neurotoxic than cocaine or cannabis (Fowler et al. 2008; Rogers et al. 2024), these findings may indicate that the cognitive effects of citicoline could vary according to the type and severity of substance‐related neurobiological impairment. Nevertheless, future studies directly comparing different substance use disorders are needed before firm conclusions can be drawn regarding substance‐specific effects.

Despite these possible substance‐related cognitive effects, improvements did not consistently translate into meaningful clinical outcomes. While reduced attentional impulsivity was reported in cannabis users and depressive symptoms improved in methamphetamine dependence, citicoline did not demonstrate superiority over placebo in reducing substance use, craving, or drinking outcomes across studies (Brown and Gabrielson 2012; Brown et al. 2019; Gruber et al. 2015). In alcohol use disorder specifically, citicoline was ineffective across measures of cognition, mood, and alcohol consumption (Brown et al. 2019).

Some evidence suggesting potential benefit was observed in individuals with comorbid bipolar disorder and cocaine dependence. Among these individuals, citicoline treatment was associated with reduced cocaine use alongside improvements in verbal memory and somatic symptoms, although no broader effects on global memory or mood stability were observed (Brown et al. 2007). However, the high overall risk of bias in this study further reduces confidence in these subgroup findings. Collectively, these findings suggest that citicoline's cognitive benefits in substance use disorders are limited, inconsistent, and potentially restricted to narrowly defined clinical subgroups rather than representing a broadly effective cognitive intervention in addiction. Therefore, future randomized controlled trials with a low risk of bias are needed to confirm these findings.

4.3. Mood Disorders

Evidence regarding citicoline supplementation in mood disorders shows a heterogeneous pattern, particularly across cognitive outcomes. In patients with major depression undergoing ECT, citicoline administered to prevent post‐ECT confusion did not demonstrate protective effects on memory or orientation, with no evidence of benefit for acute cognitive recovery in this context (Ayuso‐Gutierrez and Saiz‐Ruiz 1982). Similarly, in bipolar disorder with comorbid cocaine dependence, citicoline did not produce global memory improvements; however, a selective benefit in verbal learning was observed, reflected in improved performance on alternative word list tasks (Brown et al. 2007).

Findings related to mood symptoms should be interpreted cautiously because the available evidence was largely derived from samples with comorbid substance use disorders. In one study, greater improvement in depressive symptoms was reported among citicoline‐treated participants with bipolar or unipolar depression and methamphetamine dependence (Brown and Gabrielson 2012). However, in another study involving bipolar disorder with comorbid cocaine dependence, the findings did not support a broader or consistent mood‐related effect (Brown et al. 2007). Thus, the current evidence is insufficient to determine whether citicoline has an independent effect on mood symptoms in mood disorders beyond the context of comorbid substance use.

Notably, the studies mentioned above reporting a selective verbal learning benefit (Brown et al. 2007) and greater improvement in depressive symptoms (Brown and Gabrielson 2012) were both assessed as having a high overall risk of bias. This limitation should be taken into account when interpreting these findings, as it reduces the certainty of the evidence.

Overall, evidence regarding its effects on mood symptoms is also limited and confounded by comorbid substance use in the available studies. These findings warrant further investigation in larger, diagnosis‐specific trials employing standardized cognitive and clinical assessments.

4.4. Schizophrenia

In this systematic review, only one study (Deutsch et al. 2013) examined the effects of citicoline in individuals with schizophrenia, highlighting the limited evidence base in this population. Citicoline has been considered a potentially relevant intervention in schizophrenia due to its cholinergic modulation properties and the hypothesis that α7 nicotinic acetylcholine receptor hypofunction contributes to cognitive deficits and negative symptoms (Ghajar et al. 2018; Deutsch et al. 2005; Jones et al. 2012; Martin and Freedman 2007). However, clinical evidence supporting this rationale remains sparse. A recent meta‐analysis (Fornaro et al. 2025) suggested that citicoline may have adjunctive potential in schizophrenia, although the overall certainty of evidence was low.

In the identified trial, citicoline did not produce significant improvements in core schizophrenia symptoms, and cognitive outcomes were mixed. No significant differences were found in MCCB composite scores reflecting global cognition; however, a significant improvement in verbal free recall was observed on the USC‐REMT memory task. This pattern may point to a domain‐specific cognitive signal rather than generalized cognitive enhancement.

Unlike the other included studies, this study (Deutsch et al. 2013) was assessed as having a low overall risk of bias, reflecting its favorable methodological quality. However, since these findings are derived from a single study, they should be interpreted cautiously. Additional well‐powered trials are necessary to determine whether citicoline has meaningful cognitive or clinical benefits in schizophrenia.

4.5. Tolerability

Across the 11 included studies, both citicoline and PS were generally well tolerated, and no serious adverse events were reported. The most commonly reported adverse events were mild gastrointestinal complaints (e.g., abdominal pain, dyspepsia), consistent with broader safety data in the literature (Bruton et al. 2021; Zhao et al. 2025). Overall, available evidence suggests a favorable tolerability profile; however, the relatively short duration and modest sample sizes of several trials limit firm conclusions regarding long‐term safety.

4.6. Limitations

The findings of this systematic review should be interpreted in light of several methodological limitations. Many included studies were characterized by a high risk of bias, frequently related to missing outcome data, selective reporting, uncertainties in the randomization process, deviations from intended interventions, and insufficient blinding procedures. Notably, even studies conducted with comparatively stronger methodological rigor did not consistently demonstrate statistically significant effects, underscoring the need for larger, well‐blinded trials designed to minimize risk of bias and employing standardized, performance‐based cognitive outcomes.

Publication bias should be considered when interpreting the findings of this review. Because only 11 studies met the inclusion criteria and no quantitative meta‐analysis was conducted, a formal assessment of publication bias was not feasible. Studies with statistically significant or positive findings are generally more likely to be published than those reporting null or negative results, which may have influenced the available evidence and the conclusions of this review.

Considerable heterogeneity across studies further complicates interpretation. Variability in diagnostic groups, age ranges, sex distribution—often with a marked male predominance—outcome measures, dosages, formulations, and treatment durations limited cross‐study comparability and precluded quantitative synthesis. Notably, the overall sample across the included studies was predominantly male (72.9%), which limits the generalizability of the findings to female populations. This limitation is particularly important in ADHD, as clinical presentation, symptom expression, and neurocognitive profiles can differ between males and females. Therefore, future research should include more sex‐balanced samples and examine potential sex‐specific effects to determine whether the observed findings are consistent across males and females.

Sample sizes were frequently small and intervention periods relatively short, raising concerns about insufficient statistical power and limited generalizability. The assessment of cognition also varied widely, with some studies relying on self‐report Likert‐type scales rather than objective neuropsychological testing. Moreover, overlapping constructs across cognitive instruments complicated the interpretation of domain‐specific effects.

The frequent use of add‐on designs made it difficult to isolate the independent effects of citicoline or PS from concomitant psychotropic treatments. Similarly, the use of omega‐3‐enriched PS formulations in several studies limits the ability to determine whether the observed effects were attributable to PS itself, omega‐3 components, or their combined action. Since citicoline and PS act through different pathways, the findings should be interpreted according to the specific agent and formulation rather than as a unified treatment effect. In addition, the restriction to English‐language publications introduces a potential risk of language bias.

5. Conclusion

The evidence synthesized in this systematic review suggests that the potential cognitive benefits of citicoline and PS‐containing supplementations in mental disorders are limited and inconsistent, which may be partly related to methodological constraints, risk of bias, and modest effect sizes across studies. Current findings do not provide sufficient support for their use as stand‐alone cognitive interventions. Rather, these agents may be more appropriately considered as adjunctive or investigational options alongside established treatments.

Future research employing rigorous methodology, larger samples, and objective biomarkers such as electrophysiological or neuroimaging measures is needed to better clarify their potential role. Until more definitive evidence emerges, claims regarding the cognitive benefits of membrane‐targeting supplements in ADHD and other mental disorders should be interpreted with appropriate caution.

Author Contributions

Remzi Ogulcan Ciray (R.O.C.) was responsible for the conceptualization, design, and methodology of the study. Literature search and screening were conducted by Nazan Gundogan Kucuksahin (N.G.K.), Tugce Turan Kaya (T.T.K.), and Ekin Sut (E.S.). The data extraction and risk of bias assessment were performed independently by N.G.K. and T.T.K. Visualization of the data was handled by N.G.K., T.T.K., and E.S. The original draft of the manuscript was prepared by N.G.K., T.T.K., Eren Halac (E.H.). All authors contributed to the writing – review and editing process. R.O.C. provided overall supervision for the study. All authors have read and agreed to the published version of the manuscript.

Funding

The authors have nothing to report.

Ethics Statement

This study utilized publicly available and/or anonymized data. As no identifiable human subjects were involved, ethical approval and informed consent were not required in accordance with institutional guidelines. All data were handled in compliance with applicable data protection and privacy regulations.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Table S1: Prisma checklist for reporting the systematic review.

Table S2: Excluded records with their reasons.

Data Availability Statement

Study characteristics are summarized in Table 1. Excluded records with their reasons are provided in Table S2. The data extraction template (pre‐piloted Excel form) and additional extracted data are available from the corresponding author upon reasonable request. No analytic code is available because no quantitative synthesis was conducted.