Frauke Nees’s research while affiliated with Kiel University and other places

What is this page?


This page lists works of an author who doesn't have a ResearchGate profile or hasn't added the works to their profile yet. It is automatically generated from public (personal) data to further our legitimate goal of comprehensive and accurate scientific recordkeeping. If you are this author and want this page removed, please let us know.

Publications (467)


Impaired Working Memory and Risk-Taking Predict Detrimental Symptom Dynamics in Adolescence – A Moderated Cross-Lagged Panel Network Approach
  • Preprint

February 2025

·

4 Reads

·

Ilya Milos Veer

·

Peter de Jong

·

[...]

·

Background: Recent models of the development of adolescent psychopathology emphasize the dynamic interplay between substance use and internalizing and externalizing symptoms. Importantly, this interplay may be moderated by known risk factors of substance use and internalizing/externalizing symptoms, particularly impaired working memory capacity and high-risk taking behavior. However, thus far studies into the temporal symptom/substance use dynamics were not designed to test moderation of this dynamic interplay. To fill this critical gap, the present study introduces a novel adaptation of the cross-lagged panel network approach (CLPN) for identifying interaction effects between outside factors and cross-lagged estimates. We use this moderated CLPN approach (mCLPN) to examine how working memory and risk-taking moderate the temporal associations between substance use and internalizing and externalizing symptoms.Methods: Using data from the IMAGEN study (N = 1,364), we tested how working memory and risk-taking at age 14 moderated temporal associations between internalizing/externalizing symptoms and substance use over two years (ages 14 → 16). Results: Alcohol use showed reciprocal associations with externalizing symptoms and predicted heavier tobacco and cannabis use at age 16. Externalizing symptoms at age 14 predicted more externalizing symptoms and substance use at age 16. Poor working memory and high risk-taking moderated the temporal associations between both symptom domains and substance use. When risk-taking was high, the link between internalizing and externalizing symptoms at age 14 and cannabis use at age 16 was stronger.Conclusions: These findings highlight cognitive risk factors in the substance use/ symptom dynamics and illustrate the value of the moderated CLPN approach in clinical-developmental science.Keywords: executive functioning; cross-lagged panel network; substance use; internalizing symptoms; externalizing symptoms



Schematic workflow of data sets and analytic methodologies
First, using the ABCD cohort as the discovery sample, we conducted trajectory modeling for mental health symptoms to estimate their annual growth rates from preadolescence to adolescence. Next, we tested the associations between these symptom growth rates and the preadolescent cognitive functions. Third, we elucidated the brain’s mediators of these associations by using the brain growths as measured by volumetric changes between ages 10 and 12 years. Forth, we explored the neurobiological processes underlying these associations by both transcriptomic analysis for biological pathways and genomic analysis for individual genes. Fifth, we demonstrated the predictability of mental health trajectories during adolescence by using preadolescent cognitive functions, and validated these findings using the IMAGEN cohort.
Trajectories of externalizing and internalizing symptoms during the 2-year follow-up period in different groups
A Development trajectories of externalizing symptom in the high-externalizing-symptom group. B Development trajectories of externalizing symptoms in the control group. The embedded graph displays distributions of annual growth rates in externalizing symptoms for the high-externalizing-symptom (green) and control (grey) groups. C Development trajectories of internalizing symptoms in the high-internalizing-symptom group. D Development trajectories of internalizing symptoms in the control group. The embedded graph displays distributions of annual growth rates in internalizing symptoms for the high-internalizing-symptom (green) and control (grey) groups. E The inverted U-shaped relationship between the FI score and annual growth rates in externalizing symptoms in the high-externalizing-symptom group, and the linear association between the FI score and annual growth rates in externalizing symptoms in the control group. F The inverted U-shaped relationship between the FI score and annual growth rates in internalizing symptoms in the high-internalizing-symptom group. K The U-shaped relationship between the FI score and growth (i.e., differences between 2-year follow-up and baseline) in the brain total surface area in the high-externalizing-symptom group. L The U-shaped relationship between the FI score and growth in the brain total surface area in the high-internalizing-symptom group. We categorized high-symptom individuals into three groups based on their FI scores (lower_FI, <1 SD below the mean, moderate_FI, within 1 SD of the mean, and higher_FI, >1 SD above the mean). We also divided the high-symptom individuals into two groups according to the low and high symptoms at the baseline. Therefore, developmental trajectories for both symptoms G–J and growth for brain total cortical surface area M–P were reported in each of these groups. External_group: high-externalizing-symptom group; Internal_group: high-internalizing-symptom group.
Neurodevelopmental mediators for symptom-cognitive relationships
Brain regions, where brain volumetric growths were associated with annual growth rates in symptoms for the high-externalizing (A) /internalizing (C) -symptom individuals. Brain volumetric growths in the left lateral occipital cortex (LOC) (B), the right LOC (D), and the right pars triangularis (E) mediated nonlinear relationships between the FI score and annual growth rates in externalizing/internalizing symptoms in the high-symptom groups.
Transcriptomic and genomic findings
A Significant regions (pFDR < 0.05) associated with the fluid composite score squared (FI²) in the high-externalizing-symptom group. B Significant enrichments of the top 10% genes associated with neural correlations of FI² across different brain regions in the high-externalizing-symptom group. C The word cloud visualization for enrichment terms. D Significant enrichments of the top 10% genes in synapse-related biological entities. Sunburst plots depict synaptic locations starting with the synapse (centre), pre- and postsynaptic locations in the first ring and child terms (that is, terms that are subsets of the adjacent inner ring) in subsequent rings. E The Manhattan plot for the GWAS of FI2 (i.e., the square of [FI – center], where the center is the axis of symmetry in the inverted U-shaped symptom-cognitive relationships) using high-symptom individuals.
Validation in IMAGEN cohort for the high-externalizing-symptom individuals
A Developmental trajectories of externalizing symptoms. B The inverted U-shaped relationship between the pIQ and growth rates in externalizing symptoms. C The U-shape relationship between the pIQ and the volumetric growth rate in the left lateral occipital cortex (LOC). We categorized high-symptom individuals into three groups according to their pIQ (lower_pIQ, <1 SD below the mean, moderate_pIQ, within 1 SD of the mean, and higher_pIQ, >1 SD above the mean). D Developmental trajectories of cortical volumes of the left LOC for different groups. E Genetic association signals at the gene ADCY1. In both the ABCD and IMAGEN cohorts, for both the FI2 (i.e., the square of [FI – center], where the center is the axis of symmetry in the inverted U-shaped symptom-cognitive relationships) and volumetric growths in the left LOC (lLOC) during corresponding follow-up periods. SNPs are colored by their linkage disequilibrium (LD) with the index SNP which defaults to the SNP with the lowest p-value.
Cognitive predictors of mental health trajectories are mediated by inferior frontal and occipital development during adolescence
  • Article
  • Publisher preview available

February 2025

·

69 Reads

Molecular Psychiatry

Laboratory studies show brain maturation involves synaptic pruning and cognitive development. Human studies suggest links between early cognitive performance and later mental health, but inconsistencies remain. It is unclear if specific brain regions mediate this relationship, and the molecular underpinnings are not well understood. Here, our longitudinal analyses in both the Adolescent Brain Cognitive Development and IMAGEN cohorts establish inverted U-shaped relationships between baseline executive function and subsequent symptom trajectories in the high-symptom individuals, whose externalizing (n = 963) or internalizing (n = 1762) symptoms exceed a clinical threshold at any point during the follow-up period, but not in the control group (n = 4291). Volumetric changes in the left lateral occipital cortex (LOC) mediated the relationship with externalizing symptoms (outwardly directed behaviors such as aggression), while changes in the right LOC and pars triangularis mediated the relationship with internalizing symptoms (inwardly directed emotional problems such as anxiety). Transcriptomic and genomic findings highlighted synaptic biology and particularly the gene ADCY1, which is implicated in synaptic pruning, as underlying both moderate executive function and its associated brain mediators. Notably, preadolescent cognitive performance predicts late-onset externalizing symptoms and remitting internalizing symptoms with high accuracies (area under the curve: 0.87 and 0.79). Our findings highlight the predictive value of cognitive performance for adolescent mental health trajectories, and indicate how this is mediated by specific brain regions, and underpinned by particular molecular pathways.

View access options

Semantic search helper: A tool based on the use of embeddings in multi-item questionnaires as a harmonization opportunity for merging large datasets – A feasibility study

January 2025

·

29 Reads

European Psychiatry

Background Recent advances in natural language processing (NLP), particularly in language processing methods, have opened new avenues in semantic data analysis. A promising application of NLP is data harmonization in questionnaire-based cohort studies, where it can be used as an additional method, specifically when only different instruments are available for one construct as well as for the evaluation of potentially new construct-constellations. The present article therefore explores embedding models’ potential to detect opportunities for semantic harmonization. Methods Using models like SBERT and OpenAI’s ADA, we developed a prototype application (“Semantic Search Helper”) to facilitate the harmonization process of detecting semantically similar items within extensive health-related datasets. The approach’s feasibility and applicability were evaluated through a use case analysis involving data from four large cohort studies with heterogeneous data obtained with a different set of instruments for common constructs. Results With the prototype, we effectively identified potential harmonization pairs, which significantly reduced manual evaluation efforts. Expert ratings of semantic similarity candidates showed high agreement with model-generated pairs, confirming the validity of our approach. Conclusions This study demonstrates the potential of embeddings in matching semantic similarity as a promising add-on tool to assist harmonization processes of multiplex data sets and instruments but with similar content, within and across studies.


Impulsivity behaviors and white matter mediate the relationship between genetic risk for cannabis use disorder and early cannabis use in adolescents

January 2025

·

22 Reads

Addiction

Background and Aim Cannabis use disorder (CUD) is strongly influenced by genetic factors; however the mechanisms underpinning this association are not well understood. This study investigated whether a polygenic risk score (PRS) based on a genome‐wide association study for CUD in adults predicts cannabis use in adolescents and whether the association can be explained by inter‐individual variation in structural properties of brain white matter or risk‐taking behaviors. Design and setting Longitudinal and cross‐sectional analyses using data from the IMAGEN cohort, a European longitudinal study integrating genetic, neuroimaging and behavioral measures. We measured associations between PRS for CUD, novelty and sensation seeking traits and fractional anisotropy (FA) of white matter tracts. Mediation modeling explored whether novelty seeking and FA mediated the association between the PRS and cannabis use. Participants Participants were assessed at 14 ( n = 1762), 19 ( n = 1175) and 23 ( n = 1139) years old. Measurements European School Survey Project on Alcohol and Other Drugs, substance use risk profile scale, Fagerstrom Test for Nicotine Dependence, temperament and character inventory, Kirby Monetary Questionnaire, diffusor tensor imaging and CUD‐PRS. Findings CUD‐PRS was associated with adolescent total cannabis exposure [ P < 0.001, beta = 0.098 (95% confidence interval = 0.059, 0.137)] as well as with other substance use measures [alcohol P = 0.002, beta = 0.058 (0.020, 0.096); cigarettes smoked P < 0.001, beta = 0.086 (0.044, 0.128); fargestrom score P < 0.001, beta = 0.062 (0.028, 0.096); drug score P < 0.001, beta = 0.106 (0.065, 0.147)]. CUD‐PRS was also associated with impulsivity, risk‐taking behaviors [impulsivity P < 0.001, beta = 0.106 (0.060, 0.142); sensation seeking P < 0.001, beta = 0.094 (0.0523, 0.1357); novelty seeking P < 0.001, beta = 0.105 (0.064, 0.146); discounting task P < 0.001, beta = 0.051 (0.013, 0.089)] and average FA [ P < 0.001, beta = −0.010 (−0.015, −0.005)]. Longitudinal mediation models showed that these behaviors and brain measures could mediate the association of PRS with cannabis use [overall indirect effect for novelty seeking P < 0.001, beta = 0.048 (0.028, 0.068); impulsivity P = 0.016, beta = 0.019 (0.004, 0.035); sensation seeking P < 0.001, beta = 0.034 (0.017, 0.05)]. Conclusions The genetic risk of adult cannabis use disorder appears to be associated with substance use behavior and white matter structure as early as age 14. The observed mediation effect is consistent with the notion that genetic risk increases novelty seeking in a way that leads to more cannabis use in adolescents.


Relationships of eating behaviors with psychopathology, brain maturation and genetic risk for obesity in an adolescent cohort study

January 2025

·

60 Reads

Nature Mental Health

Unhealthy eating, a risk factor for eating disorders (EDs) and obesity, often coexists with emotional and behavioral problems; however, the underlying neurobiological mechanisms are poorly understood. Analyzing data from the longitudinal IMAGEN adolescent cohort, we investigated associations between eating behaviors, genetic predispositions for high body mass index (BMI) using polygenic scores (PGSs), and trajectories (ages 14–23 years) of ED-related psychopathology and brain maturation. Clustering analyses at age 23 years ( N = 996) identified 3 eating groups: restrictive, emotional/uncontrolled and healthy eaters. BMI PGS, trajectories of ED symptoms, internalizing and externalizing problems, and brain maturation distinguished these groups. Decreasing volumes and thickness in several brain regions were less pronounced in restrictive and emotional/uncontrolled eaters. Smaller cerebellar volume reductions uniquely mediated the effects of BMI PGS on restrictive eating, whereas smaller volumetric reductions across multiple brain regions mediated the relationship between elevated externalizing problems and emotional/uncontrolled eating, independently of BMI. These findings shed light on distinct contributions of genetic risk, protracted brain maturation and behaviors in ED symptomatology.


Structural Brain Correlates of Externalizing Traits and Symptoms in the IMAGEN Sample

January 2025

·

16 Reads

·

1 Citation

Personality Disorders: Theory, Research, and Treatment

The evidence supporting the presence of individual brain structure correlates of the externalizing spectrum (EXT) is sparse and mixed. To date, large-sample studies of brain–EXT relations have mainly found null to very small effects by focusing exclusively on either EXT-related personality traits (e.g., Hyatt et al., 2022) or EXT-related disorders/symptoms (e.g., Mewton et al., 2022). In this preregistered study using IMAGEN data (N = 1,370), we investigated the structural brain correlates of EXT factors that comprise both personality (e.g., antagonism) and psychopathology constructs (e.g., conduct disorder) across levels of morphometric specificity. Brain morphometry was operationalized in terms of omnibus measures (e.g., total brain volume), subcortical volume, and Desikan atlas regions (N = 161 structural magnetic resonance imaging metrics). We operationalized our integrated personality–psychopathology EXT through exploratory factor analyses of EXT-related measures, which identified two dimensions—nonsubstance use and substance use—and one overarching EXT domain. The results were consistent with previous large-sample neuroscientific investigations of EXT: The vast majority of relations were null, and all effect sizes were very small (largest marginal R² < .02). Preregistered supplementary analyses indicated that all significant relations found were driven by total intracranial volume and sex of the participant and became nonsignificant following the inclusion of these covariates. We conclude with suggestions regarding the importance of relevant covariates and large samples in clinical neuroscientific investigations.


Figure 2. Association between white matter FA values and pain severity (New Haven data set). Higher fractional anisotropy (FA) values in the right superior longitudinal fasciculus (SLF) are associated with greater pain reduction (from baseline to follow-up) in the New Haven data set.
Figure 8. Schematic representation of the sequential steps performed in structural connectivity. From left to right: data preparation, correction for confounders, and machine learning model building and testing. The dashed rectangle indicates that the combination of the data during model training was bootstrapped 50 times and validation and testing were repeated accordingly.
New Haven sample characteristics.
Mannheim sample characteristics.
A multisite validation of brain white matter pathways of resilience to chronic back pain

December 2024

·

16 Reads

eLife

Chronic back pain (CBP) is a global health concern with significant societal and economic burden. While various predictors of back pain chronicity have been proposed, including demographic and psychosocial factors, neuroimaging studies have pointed to brain characteristics as predictors of CBP. However, large-scale, multisite validation of these predictors is currently lacking. In two independent longitudinal studies, we examined white matter diffusion imaging data and pain characteristics in patients with subacute back pain (SBP) over 6- and 12-month periods. Diffusion data from individuals with CBP and healthy controls (HC) were analyzed for comparison. Whole-brain tract-based spatial statistics analyses revealed that a cluster in the right superior longitudinal fasciculus (SLF) tract had larger fractional anisotropy (FA) values in patients who recovered (SBPr) compared to those with persistent pain (SBPp), and predicted changes in pain severity. The SLF FA values accurately classified patients at baseline and follow-up in a third publicly available dataset (Area under the Receiver Operating Curve ~0.70). Notably, patients who recovered had FA values larger than those of HC suggesting a potential role of SLF integrity in resilience to CBP. Structural connectivity-based models also classified SBPp and SBPr patients from the three data sets (validation accuracy 67%). Our results validate the right SLF as a robust predictor of CBP development, with potential for clinical translation . Cognitive and behavioral processes dependent on the right SLF, such as proprioception and visuospatial attention, should be analyzed in subacute stages as they could prove important for back pain chronicity.


Reliability and validity of the self-report version of the Strengths and Difficulties Questionnaire (SDQ) in primary school children

December 2024

·

29 Reads

International Journal of Behavioral Development

This study examined the new self-report version of the Strengths and Difficulties Questionnaire (SDQ-S), SDQ-Kids, in primary school children regarding internal consistency, teacher-child agreement, and validity. Data from 2,655 children in Grades 1 to 3 and their teachers were analyzed. Children completed SDQ-Kids, previously piloted ( n = 896), while teachers completed SDQ-T. Reliability was measured using Cronbach’s alpha, and logistic regression analyzed the association between rating source (teachers vs. children) and SDQ status (“abnormal” vs. “normal”). Validity was assessed using Pearson’s correlation coefficient. SDQ-Kids showed acceptable internal consistency for total difficulties (α = .77) but lower for subscales (α = .40–.68). SDQ-T reliability was good for total difficulties (α = .90) and acceptable to good for subscales (α = .78–.89). Differences emerged, particularly in internalizing and externalizing problems. Correlations of SDQ-Kids with other instruments were acceptable to low. Differences between teacher and child reports highlight the need for a multi-informant approach. While SDQ-Kids’ total difficulties showed acceptable reliability, scale-level reliability and validity were unsatisfactory.


The Utility of Biomarkers for Assessment and Intervention in Neurodevelopmental Disorders

December 2024

·

23 Reads

The inherent heterogeneity of neurodevelopmental disorders (NDDs) such as attention deficit hyperactivity disorders (ADHD) and autism spectrum disorder (ASD) underscores a challenge, where heterogeneous clinical, neurocognitive and neurobiological traits defy consistent characterization across diagnoses. Therefore, the field of neurodevelopmental research has undergone strong advances with the aim to provide evidence on the underlying mechanisms of neurodevelopmental disorders. In the present chapter, we will provide an overview on the current state of knowledge on identified underlying biomarkers, elaborating on different types of biomarkers, the neural and neurophysiological, genetic, epigenetic and metabolic correlates of ADHD and ASD and discussing their potential for diagnostic and therapeutic strategies. Despite advancements in the field, the search for a single biomarker meeting the optimal standards of sensitivity and specificity remains unfulfilled. The chapter therefore also discusses potential needs for developmentally informed research strategies to address the dynamic nature of NDDs. This includes innovations in machine learning, which may offer promising biomarker discovery approaches in the spirit of precision medicine. Addressing these complexities and the ethical considerations, particularly the risk of discrimination and stigmatization, is imperative for future research.


Citations (40)


... Fourth, Phillips et al. (2025), also in a preregistered study, tested brain structure correlates of externalizing psychopathology, using a personality-psychopathology structural model. Data were from N = 1,370 young adults who took part in the second follow-up visit from the large-scale IMAGEN study (Mascarell Mari"ićet al., 2020). ...

Reference:

Who, How, and When? New Perspectives on Longstanding Issues in the Study of Externalizing Psychopathology
Structural Brain Correlates of Externalizing Traits and Symptoms in the IMAGEN Sample

Personality Disorders: Theory, Research, and Treatment

... While this pilot study has a limited sample size, the topic it covers is becoming increasingly important. Most recently, research talks about linking neuroimaging and mental health to satellite imagery of measurements of macro-environmental factors [100]. ...

Linking neuroimaging and mental health data from the ABCD Study to UrbanSat measurements of macro environmental factors
  • Citing Article
  • October 2024

Nature Mental Health

... The PRECOGNITION project aims to address this problem and enable multiple datasets to be combined via a novel imputation-based approach to harmonise studies, our proposed high-level framework. 32 Whilst this approach shares is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint ...

A protocol for data harmonization in large cohorts
  • Citing Article
  • October 2024

Nature Mental Health

... Finally, a rights-based approach to brain health also has important consequences for research culture. Fundamentally, it means that neuroscience researchers should think deeply about the ethical implications of their research, beyond a narrow view of ethics-as-respecting established protocols (Stahl et al., 2024). This includes a move towards diversity in brain health research in ways that promote underserved populations and bolster their rights (Murray and Slopen, 2024). ...

Rethinking ethics in interdisciplinary and big data-driven neuroscience projects

Nature Mental Health

... Currently, there are no objective tools or indices for measuring consciousness or pain. Techniques that measure brain activity, such as EEG and fMRI, can show patterns of brain activation, but it is unclear how these patterns are related to specific experiences of consciousness or pain [30,31]. For example, increased brain activity in a particular region does not necessarily prove that the activity directly corresponds to a conscious experience. ...

Long-term impact of maternal prenatal smoking on EEG brain activity and internalizing/externalizing problem symptoms in young adults
  • Citing Article
  • September 2024

Addictive Behaviors

... To address the limited insight into the underlying pathophysiological mechanisms, the applied topology methods can now be combined with publicly available gene expression and neurotransmitter atlases as well as with other neurophysiological modalities. 4 Integrating this multimodal information would greatly facilitate the interpretation of the pathophysiological mechanisms underlying the respective findings and allow for more guided hypothesis generation regarding potential interventions (Fig. 1). ...

Regional patterns of human cortex development correlate with underlying neurobiology

... At the individual level (host factors), attempts to identify the specific sub-populations of adolescents at risk for problematic alcohol use, both in general and during the pandemic, provide critical insight for targeted prevention efforts. Adolescents with behavioral issues including inattentive-hyperactivity symptoms [18], self-identified poor mental health [19], aggressive behaviors (particularly in sexual minority youth) [20], and risk-taking personality traits [21] are at increased risk of developing problematic drinking behaviors. Early alcohol use, risky drinking behaviors, and heavy episodic drinking are all predictors of long-term alcoholrelated outcomes, including progression to regular binge drinking [22] and an eventual diagnosis of alcohol use disorder [23]. ...

Personality, Social Factors, Brain Functioning, Familial Risk, and Trajectories of Alcohol Misuse in Adolescence
  • Citing Article
  • August 2024

JAMA Network Open

... ; https://doi.org/10.1101/2024.11.15.24317356 doi: medRxiv preprint studies have reported high accuracy in pain prediction, often utilizing advanced algorithms; others have shown moderate predictive performance, underscoring the complexity of pain as a subjective and multifaceted phenomenon. The variability in predictive performance across studies can be attributed to several factors, including differences in data types (e.g., EEG vs. fMRI), the underlying neural mechanism being measured, the specific algorithm employed, and the pain phenotypes being modeled [85][86][87]. Moreover, despite the promising results in pain prediction, there remains a need for robust external validation of these models to ensure their applicability in diverse clinical settings. ...

ENIGMA-Chronic Pain: a worldwide initiative to identify brain correlates of chronic pain

Pain

... The literature highlights significant discrepancies in understanding the role of anhedonia in the development of IA. Anhedonia, a complex psychopathological construct, has been extensively studied, particularly regarding its neurobiological mechanisms [64,65]. Neuroimaging studies reveal that anhedonia may stem from deficits in the brain's reward system, impairing the ability to derive pleasure from typically rewarding stimuli [66,67]. ...

Unraveling robust brain-behavior links of depressive complaints through granular network models for understanding heterogeneity

Journal of Affective Disorders

... In addition to the independent effects of FEP and CAs on cortical thickness in certain brain regions, we confirmed the interactive effects of CAs and FEP on cortical thickness, with differential effects of some CAs on specific brain regions, which is in line with previous research (LoPilato et al., 2019;Rapado-Castro et al., 2020). This finding also supports converging evidence on common underlying mechanisms and unique pathways of different types of CA that emerge from the immediate surroundings of an individual (Vaidya, Marquand, Nees, Siehl, & Schumann, 2024). It is crucial to further investigate the cumulative/interactive effects resulting from exposure to multiple adverse events, both simultaneously and successively, i.e., by mirroring the complex and interconnected nature of real-life situations. ...

The impact of psychosocial adversity on brain and behaviour: an overview of existing knowledge and directions for future research

Molecular Psychiatry