EFEvgeniia Frei

Evgeniia Frei

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article

Previous Organizations / Institutes

The organizations / institutes from the scholar's other articles in the past ten years


Evgeniia Frei has a total of 72 co-authors. Here are the top co-authors by number of shared publications:

  • Ole Andreas Andreassen

    Institute of Clinical Medicine, University of Oslo, Oslo, Norway

    6 shared publications
  • Viktoria Birkenæs

    Affiliations Centre for Precision Psychiatry, Division of Mental Health and Addiction, University of Oslo and Oslo University Hospital, Oslo, Norway, PsychGen Center for Genetic Epidemiology and Mental Health, Norwegian Institute of Public Health, Oslo, Norway

    6 shared publications
  • Alexey A. Shadrin

    Affiliation Centre for Precision Psychiatry, Division of Mental Health and Addiction, University of Oslo and Oslo University Hospital, Oslo, Norway

    5 shared publications
  • Pravesh Parekh

    Centre for Precision Psychiatry, Division of Mental Health and Addiction, Oslo University Hospital and Institute of Clinical Medicine, University of Oslo, Oslo, Norway

    5 shared publications
  • Helga Ask

    Norwegian Institute of Public Health

    4 shared publications
  • Nora Refsum Bakken

    Centre for Precision Psychiatry, Division of Mental Health and Addiction, Oslo University Hospital & Institute of Clinical Medicine, University of Oslo, Oslo, Norway

    4 shared publications

Latest Publications

  • Open Access

    FEMA-Long: Modeling unstructured covariances for discovery of time-dependent effects in large-scale longitudinal datasets

    by Pravesh Parekh, Nadine Parker, Diliana Pecheva, Evgeniia Frei, Marc Vaudel, Diana M. Smith, Alison Rigby, Piotr Jahołkowski, Ida Elken Sønderby, Viktoria Birkenæs, et al.

    While linear mixed-effects (LME) models are common for analyzing longitudinal data, most users rely on random intercepts or simple stationary covariance, due to unavailability of computationally tractable solutions. Here, we extend the Fast and Efficient Mixed-Effects Algorithm (FEMA) and present FEMA-Long, a computationally tractable approach to flexibly modeling longitudinal covariance suitable for high-dimensional data. FEMA-Long can: i) model unstructured covariance, ii) model covariates as smooth functions using splines, iii) discover time-dependent effects of covariates with spline interactions, and iv) use these flexible longitudinal modeling strategies to perform longitudinal genome-wide association studies and discover time-dependent genetic effects, in a computationally scalable manner, suitable for high-dimensional data. Through extensive simulations, we show that estimates from FEMA-Long are accurate, while being up to several thousand times faster and with minimal carbon footprint. To show the utility of FEMA-Long for discovering novel biological signal, using data from the Norwegian Mother, Father and Child Cohort Study (MoBa), we performed a longitudinal genome-wide association study with non-linear SNP-by-time interaction on length, weight, and BMI of 68,273 infants with up to six measurements in the first year of life. We found dynamic patterns of random effects including time-varying heritability and genetic correlations, as well as several genetic variants showing time-dependent effects, highlighting the applicability of FEMA-Long to enable novel discoveries.


  • Open Access

    Prediction of Adolescent Internalizing Disorder Risk: Evidence from the Norwegian Mother, Father, and Child Cohort Study

    by Evgeniia Frei, Oleksandr Frei, Espen Hagen, Alexey A. Shadrin, Nora R. Bakken, Viktoria Birkenæs, Helga Ask, Ole A. Andreassen, Olav Bjerkehagen Smeland

    Background: Internalizing disorders are among the most common psychiatric conditions in adolescence, often associated with long-term adverse outcomes. Early identification of at-risk youth is important for effective intervention, though it remains challenging due to the multifactorial nature of risk. Machine learning (ML) offers opportunities to integrate multiple data sources and improve risk prediction for internalizing disorders. Methods: We used data from 13,743 adolescents (mean age 14.45 years; 52.7% female) participating in the Norwegian Mother, Father and Child Cohort Study (MoBa), linked to national health registries. Logistic regression with elastic net regularization was applied to predict the risk of an internalizing disorder (mood, anxiety or stress-related) occurring within one to five years after assessment. Nested models of increasing complexity incorporated sociodemographic, clinical, lifestyle, mental health, psychosocial, and genetic predictors. Model performance was evaluated in a hold-out test set. Simplified models combining three questionnaire scales were also evaluated. Results: Test-set performance increased with model complexity, reaching area under the receiver operating characteristic curve (AUC) of 0.732 for the full model. Mental health self-reported symptoms and psychosocial predictors contributed most to the discrimination. Simplified models using three questionnaire scales, alongside age and sex, achieved AUCs up to 0.715 and effectively stratified adolescents into high- and low-risk groups (OR80/20 ranged 6.39-10.60). Conclusion: Multimodal ML models integrating registry information, mental health symptoms, psychosocial factors, and genetic data demonstrated moderate predictive performance. Simplified models with three questionnaire items reached comparable performance, highlighting their potential utility in the early identification of adolescents at elevated internalizing disorder risk.

  • Open Access

    The phenotypic and genetic relationship between adolescent mental health and time spent on social media, gaming, and TV

    by Evgeniia Frei, Oleksandr Frei, Piotr Jaholkowski, Nadine Parker, Pravesh Parekh, Alexey A. Shadrin, Espen Hagen, Nora R. Bakken, Viktoria Birkenæs, Helga Ask, et al.


  • Open Access

    FEMA-Long: Modeling unstructured covariances for discovery of time-dependent effects in large-scale longitudinal datasets

    by Pravesh Parekh, Nadine Parker, Diliana Pecheva, Evgeniia Frei, Marc Vaudel, Diana M. Smith, Alison Rigby, Piotr Jahołkowski, Ida Elken Sønderby, Viktoria Birkenæs, et al.

    While linear mixed-effects (LME) models are common for analyzing longitudinal data, most users rely on random intercepts or simple stationary covariance, due to unavailability of computationally tractable solutions. Here, we extend the Fast and Efficient Mixed-Effects Algorithm (FEMA) and present FEMA-Long, a computationally tractable approach to flexibly modeling longitudinal covariance suitable for high-dimensional data. FEMA-Long can: i) model unstructured covariance, ii) model covariates as smooth functions using splines, iii) discover time-dependent effects of covariates with spline interactions, and iv) use these flexible longitudinal modeling strategies to perform longitudinal genome-wide association studies and discover time-dependent genetic effects, in a computationally scalable manner, suitable for high-dimensional data. Through extensive simulations, we show that estimates from FEMA-Long are accurate, while being up to several thousand times faster and with minimal carbon footprint. To show the utility of FEMA-Long for discovering novel biological signal, using data from the Norwegian Mother, Father and Child Cohort Study (MoBa), we performed a longitudinal genome-wide association study with non-linear SNP-by-time interaction on length, weight, and BMI of 68,273 infants with up to six measurements in the first year of life. We found dynamic patterns of random effects including time-varying heritability and genetic correlations, as well as several genetic variants showing time-dependent effects, highlighting the applicability of FEMA-Long to enable novel discoveries. FEMA-Long is available at: https://github.com/cmig-research-group/cmig_tools . Author summary: Most large-scale datasets have complexities such as repeated measures, related individuals, or other dependencies across samples, preventing the use of standard regression approaches for analysis. In such circumstances, linear mixed-effects modeling is often employed. However, for high-dimensional datasets, fitting these models is quite challenging. Further, most standard uses of linear mixed-effects modeling focus on simpler covariance models, which may not hold. Here, we introduce FEMA-Long, a novel computationally efficient analytical framework for fitting linear mixed-effects models with time-varying random effects, as well as allowing the effect of the covariates to change smoothly over time by using splines. This is particularly relevant when, for example, studying the effect of genetic variants on phenotypes, where the effects could be non-linear over time. The FEMA-Long framework allows time-varying heritability as well as discovery of genetic variants that show time-dependent effects. By performing a genome-wide association study on data from the Norwegian Mother, Father and Child Cohort Study (MoBa) using FEMA-Long, we show the discovery of genetic variants with time-dependent effects on infant length, weight, and BMI during the first year of life. Our results highlight the potential of using FEMA-Long to make novel discoveries that can lead to biological insights on the genetics of complex traits as well as improve the potential of using genetics for personalized prediction.


  • Open Access

    Genome-wide analysis of screen behaviors among adolescents identifies novel loci and overlap with educational attainment and mental disorders

    by Evgeniia Frei, Tahir Tekin Filiz, Oleksandr Frei, Robert Loughnan, Piotr Jaholkowski, Nora R. Bakken, Viktoria Birkenæs, Alexey A. Shadrin, Helga Ask, Ole A. Andreassen, et al.

    Technological devices play a central role in adolescents’ life. Despite concerns about negative effects of excessive screen time on mental health and development, there is little knowledge of fundamental features of screen behaviours and underlying genetic architecture. Using self-reports from adolescents (14-16 years old) in the Norwegian Mother, Father, and Child Cohort Study (MoBa, n = 18 490), we performed genome-wide association analysis for four screen behaviors: time spent 1) watching movies/series/TV; 2) gaming; 3) sitting/lying down with PC, mobile or tablet; and 4) communicating with friends on social media. The resulting summary statistics were analysed using the conditional false discovery rate (condFDR) approach to increase genetic discovery. We also estimated SNP-based heritabilities of the screen behaviours and the genetic correlations with six major psychiatric disorders (schizophrenia, bipolar disorder, major depressive disorder, autism spectrum disorder, attention-deficit hyperactivity disorder, alcohol use disorder), and educational attainment. The screen-based phenotypes displayed significant SNP-based heritabilities (0.048–0.12). We also observed significant genetic correlations between screen behaviours and psychiatric disorders (r g range: 0.21–0.42). Educational attainment demonstrated significant negative genetic correlation with screen behaviours, most strongly with social media use (r g = −0.69). CondFDR analysis identified three novel loci associated with social media use. Thus, we show that screen behaviors are heritable, polygenic traits that partly share genetic signal with mental disorders and educational attainment. Future studies and larger samples are required to clarify causal relationships between these traits and disorders, and to validate the identified genetic loci associated with social media use.

  • Open Access

    Childhood trajectories of emotional and behavioral difficulties are related to polygenic liability for mood and anxiety disorders

    by Nora R. Bakken, Nadine Parker, Laurie J. Hannigan, Espen Hagen, Pravesh Parekh, Alexey Shadrin, Piotr Jaholkowski, Evgeniia Frei, Viktoria Birkenæs, Guy Hindley, et al.

    Background: Symptoms related to mood and anxiety disorders (emotional disorders) often present in childhood and adolescence. Some of the genetic liability for mental disorders, and emotional and behavioral difficulties seems to be shared. Yet, it is unclear how genetic liability for emotional disorders and related traits influence trajectories of childhood behavioral and emotional difficulties, and if specific developmental patterns are associated with higher genetic liability for these disorders. Methods: This study uses data from a genotyped sample of children ( n = 54,839) from the Norwegian Mother, Father, and Child Cohort Study (MoBa). We use latent growth models (1.5–5 years) and latent profile analyses (1.5–8 years) to quantify childhood trajectories and profiles of emotional and behavioral difficulties and diagnoses. We examine associations between these trajectories and profiles with polygenic scores for bipolar disorder (PGS BD ), anxiety (PGS ANX ), depression (PGS DEP ), and neuroticism (PGS NEUR ). Results: Associations between PGS DEP , PGS ANX , and PGS NEUR , and emotional and behavioral difficulties in childhood were more persistent than age‐specific across early childhood (1.5–5 years). Higher PGS ANX and PGS DEP were associated with steeper increases in behavioral difficulties across early childhood. Latent profile analyses identified five profiles with different associations with emotional disorder diagnosis. All PGS were associated with the probability of classification into profiles characterized by some form of difficulties (vs. a normative reference profile), but only PGS BD was uniquely associated with a single developmental profile. Conclusions: Genetic risk for mood disorders and related traits contribute to both a higher baseline level of, and a more rapid increase in, emotional and behavioral difficulties across early and middle childhood, with some indications for disorder‐specific profiles. Our findings may inform research on developmental pathways to emotional disorders and the improvement of initiatives for early identification and targeted intervention.


  • Open Access

    Youth Psychotic Experiences: Diagnostic Associations and Evaluation of the CAPE-16

    by Viktoria Birkenæs, Pravesh Parekh, Laura Hegemann, Nora R. Bakken, Evgeniia Frei, Piotr Jaholkowski, Olav B. Smeland, Ezra Susser, Katrina M. Rodriguez, Markos Tesfaye, et al.

    Background: Adolescent self-reported psychotic experiences are associated with mental illness and could help guide prevention strategies. The Community Assessment of Psychic Experiences (CAPE) was developed over 20 years ago. In a rapidly changing society, where new generations of adolescents are growing up in an increasingly digital world, it is crucial to ensure high reliability and validity of the questionnaire. Methods: In this observational validation study, we used unique transgenerational questionnaire and health registry data from the Norwegian Mother, Father, and Child Cohort, a population-based pregnancy cohort. Adolescents, aged ∼14 years, responded to the CAPE-16 ( n = 18,835) and fathers to the CAPE-9 questionnaire ( n = 28,793). We investigated the psychometric properties of CAPE-16 through factor analyses, measurement invariance testing across biological sex, response before/during the COVID-19 pandemic, and generations (comparison with fathers), and examined associations with later psychiatric diagnoses. Outcomes: One third (33·4%) of adolescents reported lifetime psychotic experiences. We confirmed a three-factor structure (paranoia, bizarre thoughts, and hallucinations) of CAPE-16, and observed good scale reliability of the distress and frequency subscales (ω = ·86 and ·90). CAPE-16 measured psychotic experiences were invariant to biological sex and pandemic status. CAPE-9 was non-invariant across generations, with items related to understanding of the digital world ( electrical influences ) prone to bias. CAPE-16 sum scores were associated with a subsequent psychiatric diagnosis, particularly psychotic disorders ( frequency: OR = 2·06; 97·5% CI = 1·70–2·46; distress: OR = 1·93; 97·5% CI = 1·63–2·26). Interpretation: CAPE-16 showed robust psychometric properties across sex and pandemic status, and sum scores were associated with subsequent psychiatric diagnoses, particularly psychotic disorders. These findings suggest that with certain adjustments, CAPE-16 could have value as a screening tool for adolescents in the modern, digital world. Funding: European Union’s Horizon 2020 Programme, Research Council of Norway, South-Eastern Norway Regional Health Authority, NIMH, and the KG Jebsen Stiftelsen.


  • Open Access

    The phenotypic and genetic relationship between adolescent mental health and time spent on social media, gaming, and TV

    by Evgeniia Frei, Oleksandr Frei, Piotr P. Jaholkowski, Nadine Parker, Pravesh Parekh, Alexey A. Shadrin, Espen Hagen, Nora Refsum Bakken, Viktoria Birkenæs, Helga Ask, et al.

    Objective: Technological advances made screen devices readily available, with adolescents spending an increasing amount of time on screens. Despite rising concerns, evidence on the relationship between screen-based activities and adolescent mental health remains inconsistent. Method: Using self-reports from the Norwegian Mother, Father, and Child Cohort Study and ICD-10 codes for mental illness from health registries, we estimated the odds of psychiatric diagnoses among adolescents across different levels of screen usage (n = 22,096). Additionally, we computed polygenic risk scores (PRSs) for schizophrenia (SCZ), bipolar disorder (BP), major depression, attention-deficit hyperactivity disorder (ADHD), autism spectrum disorder (ASD), and alcohol use disorder (AUD) for adolescents without a history of mental illness and evaluated their association with time spent on screen-based activities. Results: We observed higher odds of psychiatric diagnoses among individuals with most time spent across all screen-based behaviors. However, the lowest social media usage was also associated with mental illness (OR 2.18, 95% CI: 1.80, 2.65), particularly with increased ASD occurrence. Minimum time spent on gaming was associated with lower odds of psychiatric diagnoses (OR 0.74, 95% CI: 0.65, 0.84). PRSADHD was positively associated with time spent on social media and gaming, while PRSASD was positively associated with gaming, but negatively - with social media usage. Additionally, social media usage was significantly associated with lower PRSSCZ and PRSBP, but higher PRSAUD. Conclusion: We show that screen time behaviors are associated with mental illness among adolescents, which may partly reflect a shared genetic basis.


  • Open Access

    Using iPSC Models to Understand the Role of Estrogen in Neuron–Glia Interactions in Schizophrenia and Bipolar Disorder

    by Denis Reis de Assis, Attila Szabo, Jordi Requena Osete, Francesca Puppo, Kevin S. O’connell, Ibrahim A. Akkouh, Timothy Hughes, Evgeniia Frei, Ole A. Andreassen, Srdjan Djurovic

    Schizophrenia (SCZ) and bipolar disorder (BIP) are severe mental disorders with a considerable disease burden worldwide due to early age of onset, chronicity, and lack of efficient treatments or prevention strategies. Whilst our current knowledge is that SCZ and BIP are highly heritable and share common pathophysiological mechanisms associated with cellular signaling, neurotransmission, energy metabolism, and neuroinflammation, the development of novel therapies has been hampered by the unavailability of appropriate models to identify novel targetable pathomechanisms. Recent data suggest that neuron–glia interactions are disturbed in SCZ and BIP, and are modulated by estrogen (E2). However, most of the knowledge we have so far on the neuromodulatory effects of E2 came from studies on animal models and human cell lines, and may not accurately reflect many processes occurring exclusively in the human brain. Thus, here we highlight the advantages of using induced pluripotent stem cell (iPSC) models to revisit studies of mechanisms underlying beneficial effects of E2 in human brain cells. A better understanding of these mechanisms opens the opportunity to identify putative targets of novel therapeutic agents for SCZ and BIP. In this review, we first summarize the literature on the molecular mechanisms involved in SCZ and BIP pathology and the beneficial effects of E2 on neuron–glia interactions. Then, we briefly present the most recent developments in the iPSC field, emphasizing the potential of using patient-derived iPSCs as more relevant models to study the effects of E2 on neuron–glia interactions.