Autism spectrum disorder (ASD) spans many symptoms. We focus on those most relevant to a social circuit account: reduced social motivation and initiation, cognitive inflexibility, and heightened sensory reactivity. We examined three interacting systems: a fronto-limbic cortico-striatal–thalamic loop, cerebellar modulation of mesolimbic dopamine, and a rapid subcortical visual route to the amygdala, alongside insular salience. These features are key in functional outcome and caregiver burden, and their improvement is a central target of clinical care. Prior work has connected these systems to motivation, cognitive control, and sensory regulation relevant to these domains. Building on this, we evaluate whether model-specific and convergent alterations at the level of network organization and connectivity strength account for symptom-relevant variance.

Build a 3D connectomics dataset of the target areas and test model-specific and convergent circuit alterations relevant to social motivation and salience in ASD.

We analyzed three mouse models: Fmr1 KO, Shank3b KO, and Tsc1 cKO, with controls. Brains were processed for immunofluorescence with antibody labeling of neuronal classes and substructures, and imaged by confocal microscopy. Tiles were stitched with BigStitcher. Arivis Pro U-Net segmented soma, axon hillock, dendrites, and nuclei to generate voxel masks. Automated 3D reconstruction used Vaa3D APP2 with export as SWC. Traces and volumes were registered to a reference atlas covering the relevant target areas. A NetworkX-based data analysis pipeline converted SWC data to region-level directed graphs and performed group comparisons between each model and matched controls as well as across models. Simulations in MATLAB provided simplified models of axonal excitability, dendritic integration, and synaptic currents to contextualize the network findings.

We assembled a circuit-resolution connectomics dataset across all models. Network analysis showed weaker top-down drive within the fronto-limbic loop with fewer relay connections, reduced cerebellar modulation of mesolimbic control centers (lower hub centrality), and greater fragmentation along a rapid subcortical visual–affective route. The insular network also decoupled from frontal–striatal nodes, yielding a sparser, less integrated social-circuit backbone.

Taken together, the network that supports social valuation, control, and orienting appears less able to drive approach, to adapt when context changes, and to stabilize sensory responses. The transmitter pattern suggests damped catecholaminergic drive and greater inhibitory tone in key targets. This open access 3D, projection-defined map converts those tendencies into quantitative region-to-region measures suited for replication, cross-model comparison, and testing circuit-focused interventions.

None Declared

Attention-deficit/hyperactivity disorder (ADHD) and borderline personality disorder (BPD) are frequently comorbid and share overlapping clinical features. While psychotherapy remains the primary treatment for BPD, there is no established pharmacological intervention. Methylphenidate and atomoxetine are effective in managing ADHD symptoms.

This study aimed to evaluate the impact of methylphenidate and atomoxetine on borderline personality traits in adolescents diagnosed with ADHD.

Adolescents with ADHD who were initiated on methylphenidate or atomoxetine were followed for 12 weeks. Prior to and following treatment, participants completed the Borderline Personality Features Scale for Children (BPFS-C), the Difficulties in Emotion Regulation Scale (DERS), the Barratt Impulsiveness Scale–Short Form (BIS-S), and the Adolescent Dissociative Experiences Scale (ADES). Parents completed the Turgay DSM-IV–Based Disruptive Behavior Disorders Rating Scale (T-DSM-IV-S). Due to the small number of participants receiving atomoxetine (n = 5), only data from those treated with methylphenidate (n = 45) were included in the final analyses.

Post-treatment assessments revealed significant reductions in BPFS-C subscales and total scores, as well as in the total scores of DERS, BIS-S, and ADES. Improvements were more pronounced in female participants, particularly in the BPFS-C self-harm subscale, the total BPFS-C score, and BIS-S scores. After controlling for ADHD symptom change and sex, significant reductions remained in the BPFS-C emotional dysregulation subscale, total BPFS-C score, and total scores of DERS, BIS-S, and ADES.

Methylphenidate treatment led to significant reductions in borderline personality traits among adolescents with ADHD. These improvements were independent of ADHD symptom reduction and were especially evident in female participants.

None Declared

Youth mental health is a growing concern, with reports of psychiatric diagnoses becoming increasingly prevalent. In the Czech Republic, the prevalence of adolescent psychiatric diagnoses has increased by approximately 40% over the past decade.

While clinical guidelines recommend psychotherapy as the first-line treatment, our nationwide registry study assessed how many pediatric patients received psychotherapy before starting pharmacotherapy.

We analyzed data from the National Registry of Reimbursed Health Services which covers inpatient and outpatient services as well as prescription medications. We selected patients under 18 years old who were prescribed antipsychotics (ATC group N05A), antidepressants (N06A), anxiolytics (N05B), between 2015 and 2024, separately for every year and every ATC group. We defined the psychopharmaceutical as being prescribed for the first time if the patient had no record of the respective prescription within the previous five years. We then assessed whether patients had received psychotherapy, paid by a health insurance company, within three months prior to the first prescription.

We found an increase in the number of patients under 18 years prescribed antipsychotics and of those prescribed an antidepressant over the study period (14 636 in 2015 vs. 19 648 in 2024 for antipsychotics and 11 145 vs. 21 322 for antidepressants). The number of children and adolescents prescribed these medications for the first time also increased (5 044 in 2015 vs. 5 987 in 2024 for antipsychotics and 5 348 vs. 8 645 for antidepressants). The proportion of patients who attended at least four sessions of psychotherapy, covered by health insurance, within three months prior to the first prescription increased slightly over time from 1.1 % in 2015 to 3.7 % in 2024 for those prescribed antipsychotics, and from 1.6% to 3.9 for those prescribed antidepressants. In contrast, we observed the decrease in the number of children and adolescents prescribed an anxiolytic, decreasing from 36 034 in 2015 to 14 068 in 2024. We observed a similar proportion of decrease in the number of those prescribed an anxiolytic for the first time, from 26 649 patients to 8 384. The proportions of children and adolescents who attended at least 4 sessions of psychotherapy, covered by health insurance, within 3 months prior the first prescription of anxiolytics increased from 0.1 % in 2015 to 1.0 % in 2024.

Our findings indicate insufficient use of psychotherapy. Pharmacotherapy, which for many pediatric diagnoses should be reserved for second-line treatment, was in the vast majority of cases prescribed as the first-line option. These results highlight a critical gap between evidence-based guidelines and real-world clinical practice, underscoring the need for systemic changes to improve access to and utilization of psychotherapeutic interventions in child and adolescent psychiatry.

None Declared

*These authors contributed equally. Parameters of wellness in youth include physical activity, sleep, and screen time. While these appear to be dysregulated in disruptive disorders such as attention-deficit hyperactivity disorder (ADHD) and behavior problems, the effect of ADHD on behavior problems and wellness parameters is not clearly understood.

To examine the effect of ADHD on wellness parameters in youth with behavior problems measured on a gradient (low, moderate, severe).

The National Survey of Children’s Health (NSCH), a nationally representative sample of youth in the US for years 2022-2023, constituted the sample frame. Inclusion criteria were: 1) ages 5-17 years; 2) positive behavior problems (mild, moderate, severe); 3) ADHD data present (no/yes). Exclusion criteria were: 1) autism; 2) intellectual disability/developmental delay; 3) organic brain disease (epilepsy, cerebral palsy, Tourette syndrome). Six predictor groups were formed based on the severity of the behavior problems (3 levels) by presence of ADHD (yes/no). Multinomial logistic regressions were used to compare the predictive value of each of the five remaining groups using No ADHD, Mild Behavior Problems as a baseline group. The normative (most common) sleep, screen time and exercise values were also used as the baseline subgroup within the baseline group (No ADHD, Mild Behavior Problems). Chi-square tests were used for frequency data, and logistic regressions adjusted for socioeconomic status, caregiver mental health, and adversity measures.

From 109,265 households, only 2,690 children met inclusion criteria (10.77 +/- 3.31 years; 67% male). The severe behavior problems group was significantly less white (p < 0.001) and lower income (p < 0.001). Unexpectedly, the group with Severe Behavior Problems and No ADHD was the most discrepant from baseline, showing the worst wellness parameters. This group was ~3× more likely to report no physical activity (RRR=3.25 [1.34-7.88]) and 5× more likely to sleep less than 6 hours (RRR=5.00 [1.14-21.98]. Interestingly, screentime use depended on having ADHD compared to baseline (RRR= 1.80 [1.06–3.03] p < 0.05). Adjusting for confounders reduced, but did not eliminate, these effects.

Behavior problems, especially in the severe range, have the greatest effect on wellness parameters in youth. ADHD modifies, but does not determine, worse wellness parameters. This finding highlights the immense impact of behavioral problems beyond ADHD, suggesting that targeted approaches to address family dysfunction, implement behavioral management and engage youth in productive activities early in life will diminish poor wellness beyond the effect of ADHD.

None Declared

Esketamine nasal spray (ESK) was approved in 2019 in Europe for adults with Treatment-Resistant Depression (TRD), defined as inadequate response to at least two different antidepressants during the current moderate to severe depressive episode

To assess patient-reported outcomes related to symptomatology, quality of life, and occupational impairment over 12 months following ESK initiation in patients with Major Depressive Disorder (MDD)

ELLIPSE is a French, prospective, multicenter, non-interventional study describing MDD patients treated with ESK. Patient-reported outcomes were collected using QIDS-SR16, QLDS, and WPAI:D questionnaires throughout the 12-month follow-up (Baseline, D8, M1, M2, M3, M6, M9, M12).

A total of 200 MDD patients (including 195 [97.5%] treatment resistant patients) were enrolled across 31 sites.

Among them, 108 (54.0%) were employed at baseline, including 79 on sick leave. The remaining 92 patients had no occupation at baseline, and 50 (25.0%) cited the current depressive episode as the reason, while 30 (15.0%) were retired.

QIDS response rates (improvement from baseline ≥ 50% or score ≤ 5) in patients treated with ESK increased to 37.7% at Month 2 (M2) and 45.0% at Month 3 (M3), then plateaued (Figure 1). QLDS response (improvement from baseline ≥ 8, (Rozjabek et al. 2022)) followed a similar trend, reaching 26.5% at M2 in patients undertreatment, with sustained response over the follow-up in both the total population and the patients still under treatment (Figure 2). WPAI:D scores showed consistent improvement across all four domains in patients still under treatment at the visit (Figure 3).Image 1:

ELLIPSE is the first large-scale, real-world study in France to evaluate esketamine in MDD, offering a 360-degree view of the patient evaluation. By integrating patient-reported outcomes on depressive symptoms, quality of life, and occupational functioning, the study highlights the central role of the patient perspective in assessing treatment effectiveness. These findings reinforce the relevance of ESK in the therapeutic arsenal for TRD.

P.-M. Llorca Consultant of: Johnson&Johnson consultant, L. Samalin Consultant of: Johnson&Johnson consultant, E. Olié Consultant of: Johnson&Johnson consultant, I. Rodriguez Employee of: Johnson&Johnson employee, J. Dupin Employee of: Johnson&Johnson employee, C. Wicart Employee of: Johnson&Johnson employee, P. Demaricourt Consultant of: Johnson&Johnson consultant

Psychotic symptoms are frequent during acute episodes of bipolar disorder (BD), particularly in mania, and are traditionally considered a marker of greater severity and worse prognosis. However, data comparing psychotic and non-psychotic mania remain limited and inconsistent.

This study aimed to retrospectively examine clinical, therapeutical and 3-year outcome differences between patients hospitalized for mania with and without psychotic symptoms.

We included all patients admitted for a manic episode to the acute psychiatry unit at Hospital Clínic of Barcelona between 2015 and 2019 (n = 277). Data were extracted from medical records. Patients were followed for 3 years to assess psychiatric emergency department (PED) visits and readmissions. Functional outcomes were also analysed in the BD subgroup (n = 234). Descriptive statistics and survival analyses were used.

Psychotic symptoms were present in 73.6% of the patients and were associated with younger age (p<0.001), earlier onset (p=0.020), poorer insight (p<0.001), higher cannabis use (p=0.018) and manic predominant polarity (p=0006). Non-psychotic patients had more previous admissions for depressive episodes (p=0.003), more previous mixed episodes (p=0.006) and suicide attempts (p=0.002). No significant differences were found in PED visits or readmissions in the next 3 years. Logistic regression identified the number of previous mixed-episode-related admissions as a significant predictor of readmissions (p=0.041), while psychotic symptoms were not associated with either outcome.

While psychotic and non-psychotic mania show clinical and therapeutical differences, psychosis may not predict short-to-medium term outcomes. Non-psychotic mania - which is associated with depressive polarity, mixed features and suicidal behaviour - may indicate a clinically relevant and therapeutically challenging but underrecognized subgroup. These findings call for a nuanced comprehension of the impact of psychosis in mania and multidimensional approaches.

None Declared

Lithium remains the most effective prophylactic agent for bipolar disorder (BD), yet reports suggest that clinical practice diverges from guideline recommendations. Understanding the prescribing practices is critical to inform treatment policy.

We aimed to characterise prescribing patterns for BD in Greece, focusing on drug category prevalence, adherence, polypharmacy, and diagnostic coding practices.

We analysed nationwide prescription records for 59,035 BD patients (ICD-10 F30.x, F31.x) covering >1.2 million prescriptions over one year (2020) n the e-Government Center for Social Security (IDIKA) database. The analyses included drug prevalence, medication possession ratio (MPR) to proxy adherence, and a range of subgroup and sensitivity analyses for major psychiatric comorbidities. The derived polypharmacy and MPR phenotypes were statistically modelled with fixed- and mixed-effects models to identify predicting variables.

Sex and age distributions diverged: BD was four times more likely in boys than girls, but prevalence in adults skewed female, peaking at ages 50–59. Lithium was prescribed to 7.6% of BD patients but showed one of the highest adherence scores (median MPR 0.67). Antipsychotics dominated, with 61% of the cohort receiving olanzapine or quetiapine at least once. Polypharmacy occurred in 22% of patients, was as rule short-term, and was strongly predicted by specialist involvement and unspecified severity codes. Mixed-effects modelling showed a substantial influence of unaccounted individual-level factors (R2 61% vs R2 4% in the fixed effects model). In the sensitivity analyses, diagnostic volatility was marked, with 64.4% of patients coded under general F31/31.9, and 37% showing comorbid BD+MDD diagnoses.

In this nationwide cohort study, an underuse of lithium is apparent despite high adherence, and antipsychotics dominate. Polypharmacy may reflect specialist involvement in diagnostically uncertain cases. Previous epidemiological findings of BD sex dimorphism are validated and this is more obvious in older ages, possibly due to more accurate diagnostic processes. This misalignment of guidelines and practice suggests that prescriber habits are partly shaping BD management.

None Declared

Antidepressants (ADs) are widely prescribed for a variety of depressive disorders, but may precipitate manic episodes (antidepressant-associated mania or hypomania, AAMH). Identifying predictors and rates of AAMH in real-world settings is crucial to stratify clinical risk, especially in the absence of prophylactic mood stabilizers.

We aim to identify AD that carry the highest risk of AAMH via a triangulation of evidence framework that combines nationwide prescription datasets, large pharmacovigilance databases and genetics-informed pharmacodynamic data.

We analyzed the Electronic Governance of Social Security (IDIKA) database encompassing patients with diagnoses of unipolar and manic episodes/bipolar depression. Incident AD users were identified based on a minimum of six observed months with no recorded AD prescriptions. Patients with prior manic or bipolar diagnoses were excluded. AAMH was defined as a new manic diagnosis within a maximum of 180 days post-AD initiation. Cox proportional hazards models stratified by AD were used to evaluate predictors of switch. The FDA Adverse Event Reporting System database was used for the extraction of manic or hypomanic episodes that follow an AD prescription. To further substantiate mechanistically any observed differences in those two sources, publicly available pharmacodynamic data on the affinities of each drug were used and corresponding variants were assessed for their strength of association with various mood-related phenotypes.

Among incident AD users without prior recorded mania (n=134,353, mean (SD) age: 64 (17) years, 66% female), a small proportion experienced manic switch during follow-up (992 events, switch rate = 0.7%). Adjusted logistic and survival models revealed heterogeneity in risk between ADs. Venlafaxine (adjusted hazard ratio (aHR) 1.74, 95% CI: 1.37–2.22, p<10⁻⁶) and mirtazapine (1.65, 95% CI: 1.28–2.12, p=9×10⁻⁵) had the highest rates, whereas citalopram (1.04, 95% CI: 0.78–1.40, p=0.69) and escitalopram (baseline switch rate 0.58%) had the lowest. Moclobemide, although rarely prescribed, was a notable outlier (aHR 21.1, 95% CI: 6.7–66.3, p<10⁻⁶). Age and female gender were modest predictors, with higher risk in younger female patients. The FAERS database showed largely concordant results, with vortioxetine ranking first and citalopram last. Few differences were observed with respect to mapping of the drug target genes to mood-related phenotypes.

There are important differences in the AAMH risk profile of ADs. While overall switch rates were low, notable differences emerged by drug class, with monoamine oxidase inhibitors, SNRIs, and mirtazapine showing high rates. The largest predictor was prior BD history. Although absolute risks were lower than prior evidence suggests, clinicians should be cautious of prescribing higher-risk ADs in individuals with latent bipolarity.

None Declared

Antenatal depression (AND) is considered one of the most prevalent medical complications during pregnancy. In Malaysia, the prevalence of AND varies across different regions, ranging from 8.4% to 20.3% in recent studies [1,2]. Genetic variations in the vitamin D receptor (VDR) gene have been associated with the development of depressive symptoms [3,4]. However, its relationship with AND is yet to be explored.

This study aimed to investigate the prevalence of AND and the role of genetic variations in the VDR gene in the development of this disorder

This is a prospective cohort study that includes subjects who were recruited in their 36 – 38 weeks of gestation and attending the Obstetrics & Gynecology Clinic, Hospital Pakar Universiti Sains Malaysia (HPUSM). VDR polymorphisms (ApaI, TaqI, FokI, and BsmI) were genotyped using the Amplification Refractory Mutation System (ARMS-PCR) technique.AND was determined using the Malay version of the Edinburgh Postnatal Depression Scale (EPDS) with a cut-off point of 12. Statistical analysis for risk factors of AND was determined using the Chi-square test and logistic regression in IBM SPSS Statistics version 27.0.

A total of 207 subjects were recruited in this study, with mean years of 33 (IQR: 29–36 years). The prevalence of AND symptoms is 7.5% (n=15). Results showed that subjects who felt anxious during their current pregnancy have nine times higher odds of experiencing AND symptoms compared to those who did not have anxiety ( OR = 9.56, 95% CI: 2.92 - 31.32, P= < 0.001). A higher percentage of subjects with VDR-ApaI allelic variant (A allele) were found to have AND according to the Chi-square test (P= 0.043), while further analysis using the logistic regression test showed a trend of increased risk, although not significant (OR = 4.27, 95% CI: 0.93 - 19.5, P= 0.061). On the other hand, our results found that there is no association between the presence of each of VDR-SNPs (TaqI, BsmI and FokI) and AND among subjects.

The prevalence of AND among our study population is 7.5%. Our findings suggest that ApaI could be considered a promising marker for AND development, which requires further investigation in a larger population.

References:

Arifin et al. IIUM Med J Malaysia 2021; 20(1).

Narayanan et al. Front Psychiatry 2024; 15:1466074

Glocke et al. Kidney Blood Press Res 2013; 37:311–22.

da Silva Sabião et al. Sci Rep 2024; 14:1–11.

None Declared

Differentiating bipolar disorder (BD) from major depressive disorder (MDD) remains a major challenge in psychiatry. Although both present with depressive episodes, their underlying mechanisms and treatment pathways differ significantly. As BD often begins with depression, diagnostic approaches relying on symptoms and interviews lack precision, leading to the misdiagnosis of BD patients as MDD. The average diagnostic delay (about 10 years) has serious clinical consequences, including inappropriate treatment, risk of manic switching and increased suicide rates. A-to-I RNA editing is a highly dynamic process implicated in regulatory mechanisms relevant to psychiatry. Recent studies used circulating RNA editing biomarkers to generate AI-based algorithms capable to differentiate BD from MDD patients (Salvetat et al., 2024, J. Affect. Disord. 356, 385–393).

This study aimed to externally validate and compare the diagnostic performance of three previously developed algorithms for distinguishing BD from MDD in patients experiencing a major depressive episode in a multicentric cohort.

In this multicenter cross-sectional trial (NCT05603819), adults with BD or MDD presenting a current depressive episode (N=393) were recruited across four sites in Barcelona, Paris, and Copenhagen. Blood samples were collected, RNA was extracted, and sequencing data were processed via a secure HDS-certified platform. Algorithm outputs were compared with physician diagnoses (MINI, MADRS, YMRS and EQ-5D-5L).

The three algorithms were evaluated and compared using ROC analysis. Observed AUC were 0.834 (CI 95%: [0.793 - 0.875]) for Algorithm A1, 0.873 (CI 95%: [0.837 - 0.909]) for Algorithm A2, and 0.741 (CI 95%: [0.690 - 0.792]) for Algorithm A3. Algorithm A2, exhibited the strongest results and highest performance, with high sensitivity (82.6%) and specificity (80.3%). These results confirm our previous findings and reproduce the performances obtained on a previous independent cohort (Salvetat et al., 2024, J. Affect. Disord. 356, 385–393).

The study confirms the diagnostic potential of RNA editing signatures combined with AI for differentiating BD from MDD. Algorithm A2 demonstrated robust and reproducible accuracy in three independent cohorts, corroborating its value as an innovative complementary diagnostic tool in clinical practice.

D. Weissmann Grant / Research support from: Part of the funding for this study was provided by EIT Health., N. Salvetat: None Declared, F. ROBLES: None Declared, C. Cayzac: None Declared, F. SCHNEIDER: None Declared, M. MENHEM: None Declared, D. Vetter: None Declared, J. Haro: None Declared, C. Henry: None Declared, L. Kessing: None Declared, E. VIETA: None Declared

Pathways to homelessness for Canadian Veterans are complex and influenced by pre-existing health conditions and experiences related to military service. Conservative estimates suggest there are 2,400 Canadian Veterans experiencing homelessness with 30% identifying as women. Although 52 studies have examined homelessness among Women Veterans, none have been conducted outside the US.

The objectives of this national study are twofold: (1) to co-create solutions to homelessness among Canadian Women Veterans; and, (2) to address the knowledge gaps in the literature. This study aimed to deepen the understanding of how gender intersects with psychosocial factors; explore the service barriers faced by Women Veterans; and generate strategies to reduce their risk of homelessness.

While recruitment is currently ongoing, up to 160 Women Veterans with current or previous experience of homelessness will participate in individual, semi-structured interviews that combine open ended questions with a brief set of closed-ended items. The qualitative component is gathering data regarding lived experiences and suggestions for an ideal program to support Women Veterans. A thematic analysis informed by ethnography is being utilized for qualitative analyses to ascertain cultural intersections within this unique population. Quantitative data includes demographics, housing history and quality of life.

This study aims to deepen understandings of how gender intersects with psychosocial factors, explore the service barriers faced by Women Veterans, and generate strategies to reduce their risk of homelessness. Preliminary emergent findings uncovered the following concepts: (1) trauma and mental health challenges are widespread, requiring additional support; (2) there is a need for greater awareness of supportive Veteran services; (3) transitional housing often requires living alongside men; and, (4) there is a need for gender-specific housing and outreach services. Perceptions of sexism were also widely reported. When asked what an ideal program would be, the participants largely suggested the following: a) Housing specifically for women Veterans only; b) greater outreach to find women Veterans; c) peer support d) financial and employment support; e) more mental health supports made more available and accessible.

The preliminary findings revealed that more mental health supports and gender-specific models need to be developed. This data and analyses generated from this study will contribute a significant step towards laying the foundations for future Veteran research across Canada and the international community. It is hoped that this study will accelerate population-level functional zero homelessness in communities across all of Canada through the development of gender-specific recommendations for practice, policy and housing.

None Declared

Lithium is a first-line mood stabilizer utilized in the management of perinatal bipolar disorder. Lithium is primarily excreted renally. Pregnancy induces significant physiological adaptations, including expanded plasma volume and altered renal function, which can impact lithium disposition. Polyhydramnios, a pathological increase in amniotic fluid volume, may further modulate maternal fluid dynamics and renal perfusion, thereby influencing lithium pharmacokinetics.

To characterize the effect of polyhydramnios on maternal serum lithium levels and to evaluate the potential clinical implications for therapeutic monitoring and maternal and fetal safety.

Women attended in a perinatal psychiatric outpatient clinic of a single tertiary university hospital and treated with lithium (in monotherapy or politherapy) in the perinatal period were evaluated for eligibility to participate in this retrospective observational cohort study (November 2006-December). HCB/2020/1305). Study included demographic, psychiatric, obstetric, and serum lithium concentrations data for each pregnancie obtained from the hospital medical records.Serum lithium concentrations were determined by means of an AVL 9180 electrolyte analyser based on the ion- selective electrode (ISE) measurement principle: Limit of quantification (LoQ) was 0.20 mEq/L and detection limit (LoD) was 0.10 mEq/L. Longitudinal pattern of serum lithium concentrations during perinatal period were studied using linear mixed models.

A total of 1,260 serum lithium concentration measurements were collected from 109 pregnancies in 95 women who received lithium treatment during the perinatal period. Polyhydramnios was observed in 11 pregnancies (10%) between gestational weeks 28 and 36, with two cases classified as severe (amniotic fluid index > 35 cm). One of these cases required amnioreduction.Screening for potential underlying etiologies was negative in all cases of polyhydramnios. Given a dose of 1000 mg, serum lithium concentrations decrease in women without polyhydramnios (n=98) on average 32.9 % in the first trimester of pregnancy, 32.2 % in the second, 21.0 % in the third and increase by 0.3 % in the first trimester of postpartum. On average, mothers with polyhydramnios (n=11) exhibit higher serum lithium levels compared to those without polyhydramnios, while the relative changes are similar in both groups. Based on the linear mixed model, controlling for period and lithium dose, the differences in serum lithium concentrations between pregnant women with (n=11) and without polyhydramnios (n=98) were not statistically significant (p=0.144).

Our findings suggest that polyhydramnios does not affect serum lithium concentrations during pregnancy. Further prospective studies are warranted to develop individualized lithium monitoring protocols throughout the perinatal period.

None Declared

Extreme prematurity (<28 weeks) exposes parents to high stress levels in the Neonatal Intensive Care Unit (NICU), increasing risks of anxiety, depression, and impaired bonding. The interplay between family structure, cultural background, and parental mental health trajectories remains insufficiently studied.

To examine longitudinal changes in anxiety, depression, and bonding among parents of extremely preterm infants (EPTI) versus term infants (TI), assessing the impact of family size, race/ethnicity, and social support on parental mental health outcomes.

We conducted a prospective observational study including 150 EPTI and 50 TI. Parental anxiety (State-Trait Anxiety Inventory, STAI), depression (Beck Depression Inventory-II, BDI-II), NICU stress (PSS:NICU), bonding (Postpartum Bonding Questionnaire), and perceived social support were assessed at birth and at 40 postmenstrual weeks (PMW). Analyses included Kruskal–Wallis tests, Spearman correlations, hierarchical clustering, and mixed-effects regression models.

Anxiety and depression scores decreased significantly from birth to 40 PMW across most family structures (STAI: β = −19 to −25, p < 0.001; BDI-II: β = −3.5 to −8.1, p ≤ 0.043). Families with 2–3 cohabitants showed the greatest symptom reduction, while larger households (>5) exhibited persistently higher scores. Non-Caucasian ethnicity and lower maternal education were associated with elevated distress and weaker bonding. Heatmaps and clustering revealed distinct psychosocial profiles linked to family size and cultural background. Longitudinal models confirmed time as a significant predictor of symptom reduction in most subgroups.

Parental mental health trajectories after extreme prematurity are shaped by family structure, cultural background, and social support. Culturally sensitive, family-focused interventions in NICUs may mitigate long-term psychological burden and improve bonding outcomes.

None Declared

By entering collecting duct principal cells via the epithelial sodium channel (ENaC), lithium is capable of inducing vasopressin insensitivity, resulting in excessive urine production, nephrogenic diabetes insipidus (NDI) and potential for other long-term forms of renal dysfunction. ENaC inhibitors (ENaC-I) such as amiloride have been shown in animal models to minimise this adverse effect, and while ENaC-I are often considered an effective strategy, the literature on ENaC-I for lithium-related polyuria has not yet been synthesised despite the importance of this topic.

This review aimed to identify all published evidence for adjunctive use of an ENaC-I for lithium-related polyuria to estimate its effectiveness while also exploring potential moderators of effectiveness.

The systematic search covered databases MEDLINE, EMBASE and PsycINFO complemented by handsearches, aiming to identify all studies of ENaC-I interventions in lithium-treated patients with pre- and post-ENaC-I polyuria as outcomes.

10 studies totalling 25 participants were eligible for inclusion and were synthesised narratively. Amiloride was the ENaC-I used in 24/25 participants, and triamterene in the other. 8/10 publications were single case reports, 4 of which presented substantial confounding issues. Clear improvements to polyuria were demonstrated in most papers, including the two larger studies.

Although it appears very likely that ENaC inhibitors help ameliorate polyuria in lithium-treated patients, the quantity and quality of evidence is low. Heterogeneity in patient characteristics, intervention characteristics and study designs limit conclusions regarding the contribution of factors likely to influence ENaC-I effectiveness for lithium-induced polyuria. Besides, adverse effects require further exploration.

None Declared

People with schizophrenia have a markedly reduced life expectancy, largely due to cardiometabolic disease. Clozapine, the most effective antipsychotic for treatment-resistant schizophrenia, carries the highest risk of weight gain and metabolic dysfunction. Glucagon-like peptide-1 receptor agonists (GLP-1RAs), such as semaglutide, have demonstrated substantial weight loss in the general population, but their efficacy and safety in people with schizophrenia remain unknown.

This study aimed to evaluate the efficacy and safety of semaglutide for weight reduction in individuals with schizophrenia treated with clozapine.

The COaST trial was a phase 2, multi-centre, randomised, placebo-controlled, participant and assessor-blinded trial conducted across six Australian sites. Adults (≥18 years) with schizophrenia or schizoaffective disorder, BMI ≥26 kg/m², and on clozapine for ≥18 weeks were randomised 1:1 to receive once-weekly subcutaneous semaglutide (titrated to 2.0 mg) or placebo for 36 weeks. The primary outcome was percentage body weight change. Secondary outcomes included BMI, metabolic parameters, psychosis symptoms (PANSS), and clozapine/norclozapine levels. Adverse events were systematically recorded.

A total of 31 participants were randomised (semaglutide n=15; placebo n=16). At 36 weeks, semaglutide resulted in a mean 13.88% (SE=0.90) reduction in body weight compared to 0.42% (SE=0.93) with placebo (between-group difference −13.46%; p<0.001). Additionally, 66.7% of semaglutide-treated participants achieved ≥10% weight loss versus 0% in placebo. There were no significant differences in clozapine/norclozapine levels or psychotic symptom severity. Semaglutide was generally well tolerated, with gastrointestinal adverse events (nausea, vomiting, diarrhoea) being the most reported but not leading to study withdrawal in most cases.Image 1:

In this small trial, semaglutide demonstrated substantial weight loss without impacting psychotic symptoms or clozapine levels. These findings provide preliminary evidence supporting semaglutide as a potential intervention for antipsychotic-associated obesity in schizophrenia and warrant larger trials to confirm efficacy and long-term safety.

None Declared

Herpes simplex virus 1 (HSV1) is a neurotropic virus capable of establishing latent, persistent infection after initial exposure. HSV1 has been associated with cognitive impairments in adult patients with psychosis; however, this association has not been examined in adolescent psychosis.

We assessed HSV1 status (seropositive vs. seronegative) and cognition in adolescent patients with psychosis (early-onset psychosis; EOP) and healthy controls. We hypothesized that seropositive participants would show poorer cognitive performance than seronegative participants, and that the association would be stronger in or restricted to patients.

We included a total of 109 participants: 37 with EOP (mean age=16 years, 73% females, 27% seropositive), and 72 healthy controls (mean age=16, 53% females, 19% seropositive). In the EOP-group, 34 had non-affective psychosis (19 schizophrenia spectrum disorders and 15 other non-affective psychoses). Three participants had an affective psychosis (one bipolar I disorder and two major depressive disorder). We utilized the Measurement and Treatment Research to Improve Cognition in Schizophrenia (MATRICS) Consensus Cognitive Battery. We evaluated circulatory immunoglobulin G concentrations expressed as a dichotomous measure. In our main analyses, we applied sex- and age-adjusted models to test associations between HSV1 status and MATRICS cognitive domain scores.

In the whole group of patients and healthy controls, HSV1 seropositivity was associated with impaired speed of processing (p=0.012), working memory (p=0.006), verbal learning (p=0.002), visual learning (p=0.049) and reasoning/problem solving (p=0.005), with medium to large effect sizes (partial eta-squared=0.06-0.09). For speed of processing, working memory, verbal learning and reasoning/problem solving, the associations were driven by the patient group, with large effect sizes (partial eta-squared=0.12-0.23).

HSV1 exposure in adolescents was linked to reduced cognitive performance, an effect that was more pronounced among those with psychosis.

D. Andreou: None Declared, N. E. Steen: None Declared, A. Myhre: None Declared, O. Andreassen Consultant of: CorTechs and Precision Health, Speakers bureau of: Janssen, Lundbeck and Sunovion, R. Yolken: None Declared, I. Agartz Speakers bureau of: Lundbeck

First-episode psychosis (FEP) can occur in substance-free individuals or emerge in the context of methamphetamine use. Inflammation has been implicated in the pathophysiology of psychosis, but comparative data on systemic inflammatory markers between these two FEP populations remain limited.

We aimed to compare blood inflammatory and metabolic parameters between patients with substance-free FEP and patients with first-episode methamphetamine-induced psychosis.

This cross-sectional study included 211 patients with substance-free FEP, 199 patients with methamphetamine-induced FEP, and 101 healthy controls. Laboratory measurements comprised white blood cell count (WBC), neutrophils, lymphocytes, monocytes, platelets, triglycerides, low-density lipoprotein (LDL), and high-density lipoprotein (HDL). In addition, several derived inflammatory ratios were calculated: neutrophil-to-lymphocyte ratio (NLR), platelet-to-lymphocyte ratio (PLR), monocyte-to-lymphocyte ratio (MLR), monocyte-to-HDL cholesterol ratio (MHR), systemic immune-inflammation index (SII), systemic inflammation response index (SIRI), and pan-immune-inflammation value (PIV). Group comparisons were performed using ANOVA and post-hoc analyses, with statistical significance set at p<0.05.

WBC and neutrophil counts were elevated in both substance-free and methamphetamine-induced first-episode psychosis groups compared to healthy controls (p<0.001). Lymphocyte counts were significantly higher in the methamphetamine-induced group than in the substance-free group (p=0.001).Triglyceride levels were increased and HDL levels were decreased in the methamphetamine-induced group relative to both substance-free psychosis patients and controls (p<0.001), whereas LDL levels were reduced in both psychosis groups compared to controls (p<0.001). Regarding inflammatory indices, substance-free FIP patients demonstrated higher NLR, MLR, SII, SIRI, and PIV values compared to methamphetamine-induced psychosis patients (all p<0.05). In contrast, MHR was significantly higher in the methamphetamine-induced psychosis group (p=0.032). In substance-free FEP group, the NLR was significantly higher in those with a disease duration of 1–5 years compared to first-year patients (p=0.020).

Methamphetamine-induced FEP was associated with metabolic dysregulation (elevated triglycerides, low HDL, higher MHR), whereas substance-free FEP was characterized by a stronger systemic inflammatory response (elevated NLR, MLR, SII, SIRI, PIV). As a recognized marker of systemic inflammation, elevated NLR may indicate an increased inflammatory burden with longer illness duration These differential biomarker patterns suggest distinct underlying mechanisms and may aid in the clinical differentiation and prognostic evaluation of first-episode psychosis.

None Declared

Schizophrenia is a mental disorder whose pathophysiology is still not fully understood. Electroencephalographic (EEG) microstates (MS), which can be defined as instantaneous representations of scalp electrical activity, have been employed to investigate alterations in neural synchrony associated with this disorder. After being first characterized using resting-state EEG data, MS analysis has gained popularity as a useful tool for investigating spatial and temporal properties of event-related potentials (ERPs).

This study aims to investigate the alterations in parameters of MS in a sample of subjects with SCZ (SCZs), as compared to a group of healthy controls (HCs), and to explore their associations with negative symptoms and cognitive impairment.

The analysis included EEG data from 132 SCZs and 65 HCs recorded during an auditory oddball task (320 standard tones, ST; and 80 target tones, TA). MS analysis was conducted using Cartool, and five parameters [Global explained variance (GEV), mean duration, time coverage, mean global field power (GFP), segment density] were extracted for each MS. Repeated-measures ANOVAs were performed to test group and group × condition interaction effects in each MS class. Correlation analyses of the extracted parameters with negative symptoms and cognitive impairment were performed.

Five MS maps (Image 1) were detected across the two study groups (SCZ and HC) and conditions (ST and TA). Visualization of the five maps showed that Map 1 had the typical topography of the N100 component (frontocentral negativity), while Map 2 was associated with the P3b component (parietal positivity).

The ANOVA showed a significant group effect for Map 1 and Map 2. For both maps, higher time coverage was observed in SCZ (p < 0.01), as compared to HC. Furthermore, GEV and mean duration of Map 3 (characterized by a central positivity) showed significant group and group x condition interaction effects (p < 0.01). Post hoc analysis showed lower GEV (only during the TA) and lower mean duration (during the ST and TA) in SCZ, as compared to HC (p < 0.01).

Correlation analyses showed that the expressive deficit domain of negative symptoms was negatively correlated (p < 0.05) with GEV of Map 1 in both ST (r = -0.22) and TA (r = -0.23) conditions and with mean duration of Map 1 during the TA condition (r = -0.27). Furthermore, significant positive correlations (p < 0.05) were observed between the attention-vigilance domain scores and GEV (r = 0.24), mean duration (r = 0.24) and time coverage (r = 0.23) of Map 3 during the ST condition within SCZ sample.Image 1:

The present study shows that MS analysis can highlight alterations in the temporal dynamics of scalp activity in subjects with schizophrenia during different stages of sensory and cognitive processing. Furthermore, specific parameters of MS are associated with psychopathological features and cognitive impairment.

None Declared

Tardive dyskinesia, or TD, is a drug-induced movement disorder, linked with long-term use of dopamine receptor blockers. TD imposes a substantial healthcare burden, with higher hospitalization and emergency visits, higher medical costs, and increased psychiatric and physical comorbidities.

This study aims to identify drugs associated with TD risks beyond established agents such as metoclopramide using the FDA Adverse Event Reporting System (FAERS) database, which is a public, quarterly-updated, spontaneous reporting system for adverse drug events.

TD reports were extracted from the FAERS database (Q1 2004 to Q4 2024). Four disproportionality methods were appllied to identify statistically significant TD risk signals. Sensitivity analyses utilizing adjusted RORs (aRORs) were conducted for drugs with ≥20 reports and a significant signal. Multivariable logistic regression analysis was employed to identify risk factors influencing reporting probability. Time-to-onset (TTO) was calculated from treatment initiation to event onset and analysed descriptively with median and Weibull shape parameters (WSP), and presented by Kaplan–Meier cumulative-incidence curves.

We identified 9,390 TD reports involving 399 drugs. Antipsychotics dominated the top 3 most frequently reported drugs: aripiprazole (n=2099), quetiapine (n=1121), and risperidone (n=910). Disproportionality analysis identified statistically significant TD signals for 57 drugs: antipsychotics (n=27, 47.4%), antidepressants (n=13, 22.8%), psychostimulants (n=4, 7.0%), and others (n=13, 22.8%). Sensitivity analyses confirmed significant positive aRORs for all evaluated agents. The strongest signals were observed for brexpiprazole (aROR=899.7) and cariprazine (aROR=732.6) among antipsychotics, fluoxetine (aROR=43.3) among antidepressants, and benztropine (aROR=467.3) among other classes. Age significantly influenced reporting risk; the effect of sex varied by drug, with higher risk in females for duloxetine and benztropine, and in males for citalopram and fluoxetine. TTO was shortest in patients aged<19 years (median=155.5 d). Second-generation antipsychotics (median = 421.0 d) and first-generation antipsychotics (626.0 d) exhibited longer TTOs than antidepressants (median=246.0 d). Weibull analysis indicated early-failure patterns for risperidone, olanzapine, lurasidone, haloperidol, and fluoxetine, while all other analyzed agents followed a random failure pattern over time.

Our findings underscore significant TD risk associated with a broad pharmacologic spectrum, particularly with antipsychotics and antidepressants, and reveal distinct TTO profiles by drug class. These data support the need for individualized, long-term monitoring of TD risk during neuropsychiatric treatment.

None Declared

Large language models (LLMs) are becoming ubiquitous, offering scalable, responsive and seemingly empathic dialogue. While their therapeutic potential has attracted attention, emerging reports highlight risks of “AI-associated delusions,” where interactions with conversational agents appear to amplify psychotic thinking. These interactions may validate persecutory or grandiose content, blur boundaries of reality, and reinforce epistemic instability. Given the global burden of psychosis and the accelerating adoption of generative AI, there is an urgent need to characterise risks and outline safeguards.

This study aimed to:

1. 1. Characterise emerging reports of AI-associated delusions.

2. 2. Identify cognitive mechanisms through which LLMs may reinforce psychotic symptoms.

3. 3. Propose clinical and technological strategies for safe integration of AI in psychiatric contexts.

We conducted a narrative review and synthesis of published case reports, media accounts, and preliminary research on AI–psychosis interactions (Hill 2025; Dupré 2025; Moore 2025). Cases were thematically analysed for recurrent delusional patterns. Cognitive and phenomenological theories of psychosis were integrated with AI safety literature to develop a framework for risk mechanisms and safeguarding strategies.

Case synthesis (n≈15, see Table 1) identified three recurring delusional trajectories: (i) spiritual or messianic revelations; (ii) beliefs in sentient/god-like AI; (iii) attachment or romantic delusions towards AI agents. Interactions frequently escalated from pragmatic use into fixation, with conversational “sycophancy” and memory features reinforcing conviction. Theoretical integration suggests heightened agency-detection biases in psychosis may be particularly vulnerable to AI’s anthropomorphic design. We outline potential mechanisms of AI-amplified delusion (Table 2) and a flow model of recursive reinforcement (Figure 1). Proposed safeguards include personalised instruction protocols, reflective check-ins, and digital advance statements embedding relapse-prevention strategies into AI behaviour.

LLMs may serve as both coping aids and destabilising agents for individuals vulnerable to psychosis. Current evidence indicates reinforcement of delusional beliefs rather than induction of de novo schizophrenia, but prevalence and long-term trajectories remain unknown. Clinical teams should routinely assess AI use in patients and co-develop digital safety plans. Developers must embed safeguards (non-human disclaimers, pattern recognition of psychotic content, escalation protocols) prior to widespread deployment. AI literacy should become a psychiatric competency. Urgent interdisciplinary research is required to map prevalence, mechanisms, and protective strategies.

None Declared