Stress-related disorders, neurodevelopmental disorders, Attention Deficit Hyperactivity Disorder (ADHD), Autism Spectrum Disorder (ASD)
Conditions
Interventions
None listed
Sponsors
Radboud Universitair Medisch Centrum
Eligibility
Age
18 Years to 64 Years
Inclusion criteria
Inclusion criteria: Gave consent for utilization of their clinical data for research purposes (MIND-Set 2, Protocol ID: 2022-1367)Adult (= 18 years old) outpatients at the Department of Psychiatry, Radboudumc, referred for diagnostic or treatment questions related to:Stress-related disorders (e.g., mood, anxiety)Neurodevelopmental disorders (e.g., ASD, ADHD, personality disorders)Impulse-regulation disorders (substance abuse disorder, eating disorders)
Exclusion criteria
Exclusion criteria: Inadequate command of the Dutch language Mentally incompetent to provide informed consentSevere somatic comorbidities (i.e., terminal illnesses)Comorbidity with any psychotic disorder
Design outcomes
Primary
| Measure | Time frame |
|---|---|
| Microbial compositionThe gut microbiota contain many different types of bacteria with different abundance. Here, our main focus is on the abundance of each microbial genus (taxonomic composition). As the gut microbiome data is compositional and zero-dispersed, we will perform the analysis on genus level, transform the raw sequencing count data to central log ratio (clr) and apply 10% prevalence filtering. We will also perform the analysis on species level as we will also obtain higher resolution data from shotgun metagenomics sequencing. We will also look into the metabolic functions associated with the taxonomic profile. As measures of relevant global microbial dynamics, we will calculate alpha diversity (difference within group) and beta diversity (difference between group) from the raw sequence read count. Beta diversity will be computed using Aitchison distance.Clinical outcomesWe derive the primary study parameters, clinical outcomes, from the MIND-Set 2 study. These are disorder-specific outcomes as measured with relevant questionnaires (for example IDS for depression) and mental health outcomes as assessed with general outcome questionnaires (OQ, WHODAS). The primary parameters will be symptom profile and functional domains. We will apply machine learning technique such as factor or hierarchical clustering analysis, supervised learning and exploratory factor analysis to establish and confirm the functional domains and symptom profile from the questionnaire data. Association analysisThe clinical outcomes and microbial measures will be statistically associated using logistic regression and generalized linear mixed model. We will associate the clr-transform taxonomic abundance to the identified symptom and functional domains, as previously done in (11). In addition, we will also associate the metabolic functions of the bacterial taxa with the symptom and functional domains using generalized linear model. All the associations are subject to multiple hypothes | — |
Secondary
| Measure | Time frame |
|---|---|
| Fluctuations of gut microbiotaTo answer the secondary research questions, we will measure the changes of the gut microbiota over time as the secondary parameter. We will use volatility measure: the Euclidean distance of gut microbiota between each measured time point. We will also perform the same data transformation protocol on the follow up measurement and then compare them to the baseline measurement. In addition, we will also use generalized mixed linear model to identify significant changes in the gut microbiota composition.We will also investigate the changes in metabolic functions over time. These will be linked to fluctuation in the composition. Considering the nature of gut microbiota data, generalized linear mixed-model is the optimal statistical test that we can use to assess these associations.Changes in symptom profile and functional domainThe questionnaire data from the follow up measure will be the basis for determining the symptom profile and functional domain. We will contrast these measures between the two time points by performing t-tests and/or linear regression.Association analysisThe gut microbiota composition and metabolic functions from both time points will be associated with symptom profiles and functional domains from both time points. To take into account the dependency within the data due to repeated measures, we will employ generalized linear mixed-models. All the association tests will be corrected for multiple comparisons with the Benjamini-Hochberg procedure. | — |
Countries
Netherlands
Contacts
Public ContactA Arias Vasquez
Radboud Universitair Medisch Centrum
Outcome results
None listed