Charlotte Fraza

This paper updates the normative modelling protocol for computational psychiatry, including new methods, software tools, and models for neuroimaging data analysis.

This paper introduces a framework using extreme value statistics to model extreme deviations from normative neuroimaging models, demonstrating its application to the UK Biobank dataset for risk estimation and anomaly detection.

This paper addresses structured missingness in biomedical data by introducing a hierarchical imputation framework and advocating for alternative evaluation metrics.

This paper introduces a framework using normative models to assess intra-individual longitudinal changes in neuroimaging data, revealing significant grey matter normalization in schizophrenia patients over time.

This paper extends hierarchical Bayesian regression to model non-Gaussian data with skewness and kurtosis, showing improved performance on neuroimaging datasets.

This study maps individual brain deviation scores in individuals with pathogenic CNVs using normative modeling and shows correlations between negative deviations and cognitive function.

This work develops normative models of brain growth and aging using neuroimaging data to quantify variability across disorders and improve case-control inferences.

This paper introduces a warped Bayesian Linear Regression framework for normative modelling, improving scalability and non-Gaussian distribution handling in neuroimaging data.

This paper introduces a standardized protocol for normative modeling in computational psychiatry using the PCNtoolkit, enabling analysis of biological heterogeneity without relying on case-control comparisons.

Generalized additive models for location, scale and shape (GAMLSS) using SHASH distributions beat Gaussian brain normative models on log score and calibration in most of 150 UK Biobank cortical-thickness regions.

The paper learns interpretable latent representations of task-fMRI data from large-scale datasets, capturing demographic and behavioral variability through a three-dimensional autoencoder and normative modeling.