Mandy Melissa Jane Wittens

Mandy Melissa Jane Wittens

Auto-generated profile, based on publicly available information

Mandy Melissa Jane Wittens

imec-SMIT, Vrije Universiteit Brussel, Pleinlaan 9, Brussels, Belgium

Research Focus & Metrics

Academic Career Length

7Years (2020-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
1 January 2020 in NeuroImage: Clinical
Verified
Journal Article
Open access
HSHanne Struyfs
Hanne Struyfs
DSDiana Maria Sima
Diana Maria Sima
MWMelissa Wittens
Melissa Wittens
ARAnnemie Ribbens
Annemie Ribbens
NBNuno Pedrosa de Barros
Nuno Pedrosa de Barros
DSDirk Smeets
Dirk Smeets
Brain volumes computed from magnetic resonance images have potential for assisting with the diagnosis of individual dementia patients, provided that they have low measurement error and high reliability. In this paper we describe and validate icobrain dm, an automatic tool that segments brain structures that are relevant for differential diagnosis of dementia, such as the hippocampi and cerebral lobes. Experiments were conducted in comparison to the widely used FreeSurfer software. The hippocampus segmentations were compared against manual segmentations, with significantly higher Dice coefficients obtained with icobrain dm (25-75th quantiles: 0.86-0.88) than with FreeSurfer (25-75th quantiles: 0.80-0.83). Other brain structures were also compared against manual delineations, with icobrain dm showing lower volumetric errors overall. Test-retest experiments show that the precision of all measurements is higher for icobrain dm than for FreeSurfer except for the parietal cortex volume. Finally, when comparing volumes obtained from Alzheimer's disease patients against age-matched healthy controls, all measures achieved high diagnostic performance levels when discriminating patients from cognitively healthy controls, with the temporal cortex volume measured by icobrain dm reaching the highest diagnostic performance level (area under the receiver operating characteristic curve = 0.99) in this dataset.

1 March 2022 in Toxicology Letters
Verified
Journal Article
Open access
MVM.C.T Verheijen
M.C.T Verheijen
JKJ. Krauskopf
J. Krauskopf
FCF. Caiment
F. Caiment
MNM. Nazaruk
M. Nazaruk
QWQ.F. Wen
Q.F. Wen
JBJacco Briede
Jacco Briede
Alzheimer's disease (AD) is the most common cause of dementia, characterized by the progressive impairment of cognition and memory loss. Sporadic AD (sAD) represents approximately 95 % of the AD cases and is induced by a complex interplay between genetic and environmental factors called "Alzheimerogens". Heavy metals (e.g. copper) and pesticides (e.g. fipronil) can affect many AD-related processes, including neuroinflammation (considered as AD-inducing factor). Research would benefit from in vitro models to investigate effects of Alzheimerogens. We compared transcriptomics changes in sAD induced pluripotent stem cell (iPSC) derived cortical neurons to differentially expressed genes (DEGs) identified in post-mortem AD brain tissue. These analyses showed that many AD-related processes could be identified in the sAD iPSC-derived neurons, and furthermore, could even identify more DEGs functioning in these processes than post-mortem AD-brain tissue. Thereafter, we exposed the iPSCs to AD-inducing factors (copper(II)chloride, fipronil sulfone and an inflammatory cytokine cocktail). Cytokine exposure induced expression of immune related genes while copper-exposure affected genes involved in lipid and cholesterol metabolism, which are known AD-related processes. Fipronil-exposure did not result in significant transcriptomic changes, although prolonged exposures or higher doses may be necessary. Overall, we show that iPSC-derived cortical neurons can be beneficial in vitro models to identify Alzheimerogens and AD-related molecular mechanisms.

19 January 2024 in Neuroradiology
Verified
Journal Article
Open access
PDF file
MWMandy M. J. Wittens
Mandy M. J. Wittens
GAGert-Jan Allemeersch
Gert-Jan Allemeersch
DSDiana Maria Sima
Diana Maria Sima
TVTim Vanderhasselt
Tim Vanderhasselt
SRSteven Raeymaeckers
Steven Raeymaeckers
SESebastiaan Engelborghs
Sebastiaan Engelborghs
Purpose: To assess the performance of the inferior lateral ventricle (ILV) to hippocampal (Hip) volume ratio on brain MRI, for Alzheimer’s disease (AD) diagnostics, comparing it to individual automated ILV and hippocampal volumes, and visual medial temporal lobe atrophy (MTA) consensus ratings. Methods: One-hundred-twelve subjects (mean age ± SD, 66.85 ± 13.64 years) with varying degrees of cognitive decline underwent MRI using a Philips Ingenia 3T. The MTA scale by Scheltens, rated on coronal 3D T1-weighted images, was determined by three experienced radiologists, blinded to diagnosis and sex. Automated volumetry was computed by icobrain dm (v. 5.10) for total, left, right hippocampal, and ILV volumes. The ILV/Hip ratio, defined as the percentage ratio between ILV and hippocampal volumes, was calculated and compared against a normative reference population (n = 1903). Inter-rater agreement, association, classification accuracy, and clinical interpretability on patient level were reported. Results: Visual MTA scores showed excellent inter-rater agreement. Ordinal logistic regression and correlation analyses demonstrated robust associations between automated brain segmentations and visual MTA ratings, with the ILV/Hip ratio consistently outperforming individual hippocampal and ILV volumes. Pairwise classification accuracy showed good performance without statistically significant differences between the ILV/Hip ratio and visual MTA across disease stages, indicating potential interchangeability. Comparison to the normative population and clinical interpretability assessments showed commensurability in classifying MTA “severity” between visual MTA and ILV/Hip ratio measurements. Conclusion: The ILV/Hip ratio shows the highest correlation to visual MTA, in comparison to automated individual ILV and hippocampal volumes, offering standardized measures for diagnostic support in different stages of cognitive decline.

12 March 2026 in Applied Ergonomics
Verified
Journal Article
Closed access
SEShirley Elprama
Shirley Elprama
JVJorre Vannieuwenhuyze
Jorre Vannieuwenhuyze
MWMandy Melissa Jane Wittens
Mandy Melissa Jane Wittens
Bram VanderborghtBram Vanderborght
Bram Vanderborght
ARAlberto Ranavolo
Alberto Ranavolo
AJAn Jacobs
An Jacobs

3 April 2025
Verified
Thesis
Closed access
MWMandy Melissa Jane Wittens
Mandy Melissa Jane Wittens
Alzheimer’s disease (AD) is a global health challenge influenced by genetic, lifestyle, and environmental factors. Though its precise onset remains unclear, AD is characterized by amyloid-β plaques, tau tangles, and neuronal loss, leading to cognitive impairments. The adoption of a clinical-biological biomarker framework enables earlier intervention and personalized treatment, driving research toward refining biomarkers for improved diagnosis and prognosis. Magnetic resonance imaging (MRI) biomarkers, including medial temporal lobe atrophy (MTA), hippocampal volume reduction, and ventricular enlargement, are valuable for detecting brain changes. Traditionally, manual segmentation has been the gold standard, but automated volumetric tools using machine learning are now enhancing accuracy and efficiency. This study evaluates icobrain dm, an automated brain volumetry tool for clinical MRI scans, assessing its feasibility for routine clinical integration. By addressing validation, standardization, and biomarker integration challenges, this research aims to enhance AD detection, staging, and management, supporting personalized treatment strategies. Our findings show that icobrain dm demonstrates robust intra- and inter-scanner reliability (mean DSC > 0.88) and high reproducibility (mean CV < 3%) across global, cortical, and subcortical structures. Icobrain’s automated MTA assessment, expressed as the ratio between inferior lateral ventricle and hippocampal volumes, correlates strongly with visual MTA ratings across cognitive decline stages. However, automated hippocampal volumes alone exhibit moderate correlation, highlighting the inferior lateral ventricle's crucial role in MTA assessment. Comparing automated MTA scores with age-matched healthy controls could aid in clinical interpretation of medial temporal atrophy severity. Additional findings suggest icobrain dm enhances diagnostic certainty by assessing hippocampal and temporal cortex volumes in AD. Brain age and Brain-Predicted Age Difference (BPAD) correlate with accelerated brain aging beyond chronological age and classify cognitive states effectively. Notably, heavy alcohol consumption accelerates brain aging, as reflected in significant BPAD differences between heavy drinkers and non-drinkers. The development of event-based models (EBMs) for AD staging has shown promising generalizability and reliability across datasets. Notably, amnestic Mild Cognitive Impairment (MCI) subjects exhibit higher biomarker event scores than non-amnestic MCI, improving precision in clinical trial recruitment. Additionally, analysis of white matter hyperintensity (WMH), neuropsychological scores, and hippocampal volumes underscores the complex relationship between WMH and cognitive function. Despite progress, challenges remain in validating automated measurements across diverse populations. Tailored EBMs for AD subtypes are crucial for accurate staging, and further research is needed to interpret the interplay between WMH load and cognitive outcomes. Addressing these challenges will enhance the adoption of automated MRI analysis, improving AD diagnosis and patient care.

2 July 2026
Verified
Book Chapter
Open access
MWMandy Melissa Jane Wittens
Mandy Melissa Jane Wittens
KKKristýna Sirka Kacafírková
Kristýna Sirka Kacafírková
SEShirley Elprama
Shirley Elprama
An JacobsAn Jacobs
An Jacobs
Designing meaningful human oversight (HO) for AI systems is increasingly challenging due to rising system complexity, automation, and temporal dynamics. Regulatory frameworks and international standards provide high-level guidance, but there are limited operational instructions for implementing context-sensitive HO. Moreover, human decision-making can be shaped by automation, intervention opportunities may be unclear or difficult to execute, and operator intentions or abilities may not align with system objectives. We propose the Human Oversight Anchor Framework, which structures oversight across three interrelated micro-level layers: task, monitoring, and decision. These are embedded within a broader process level, comprising of human and technical factors, while situating these within meso-level organizational conditions and macro-level legal, societal, cultural, and environmental contexts. It further links known and emergent risks to system characteristics, development stage, and stakeholder considerations. As a conceptual basis for future operationalization through design patterns, evaluation metrics, and HO-focused documentation, the framework supports structured reasoning about meaningful oversight in complex and evolving AI systems.