Author name cluster

Marion Smits

Possible papers associated with this exact author name in Arrow. This page groups case-insensitive exact name matches and is not a full identity disambiguation profile.

10 papers

1 author row

YNICL Journal 2021 Journal Article

Cross-cohort generalizability of deep and conventional machine learning for MRI-based diagnosis and prediction of Alzheimer’s disease

Esther E. Bron
Stefan Klein
Janne M. Papma
Lize C. Jiskoot
Vikram Venkatraghavan
Jara Linders
Pauline Aalten
Peter Paul De Deyn

This work validates the generalizability of MRI-based classification of Alzheimer's disease (AD) patients and controls (CN) to an external data set and to the task of prediction of conversion to AD in individuals with mild cognitive impairment (MCI). We used a conventional support vector machine (SVM) and a deep convolutional neural network (CNN) approach based on structural MRI scans that underwent either minimal pre-processing or more extensive pre-processing into modulated gray matter (GM) maps. Classifiers were optimized and evaluated using cross-validation in the Alzheimer's Disease Neuroimaging Initiative (ADNI; 334 AD, 520 CN). Trained classifiers were subsequently applied to predict conversion to AD in ADNI MCI patients (231 converters, 628 non-converters) and in the independent Health-RI Parelsnoer Neurodegenerative Diseases Biobank data set. From this multi-center study representing a tertiary memory clinic population, we included 199 AD patients, 139 participants with subjective cognitive decline, 48 MCI patients converting to dementia, and 91 MCI patients who did not convert to dementia. AD-CN classification based on modulated GM maps resulted in a similar area-under-the-curve (AUC) for SVM (0.940; 95%CI: 0.924–0.955) and CNN (0.933; 95%CI: 0.918–0.948). Application to conversion prediction in MCI yielded significantly higher performance for SVM (AUC = 0.756; 95%CI: 0.720-0.788) than for CNN (AUC = 0.742; 95%CI: 0.709-0.776) (p<0.01 for McNemar's test). In external validation, performance was slightly decreased. For AD-CN, it again gave similar AUCs for SVM (0.896; 95%CI: 0.855–0.932) and CNN (0.876; 95%CI: 0.836–0.913). For prediction in MCI, performances decreased for both SVM (AUC = 0.665; 95%CI: 0.576-0.760) and CNN (AUC = 0.702; 95%CI: 0.624-0.786). Both with SVM and CNN, classification based on modulated GM maps significantly outperformed classification based on minimally processed images (p=0.01). Deep and conventional classifiers performed equally well for AD classification and their performance decreased only slightly when applied to the external cohort. We expect that this work on external validation contributes towards translation of machine learning to clinical practice.

Details DOI

YNIMG Journal 2020 Journal Article

Neuro4Neuro: A neural network approach for neural tract segmentation using large-scale population-based diffusion imaging

Bo Li
Marius de Groot
Rebecca M.E. Steketee
Rozanna Meijboom
Marion Smits
Meike W. Vernooij
M. Arfan Ikram
Jiren Liu

Subtle changes in white matter (WM) microstructure have been associated with normal aging and neurodegeneration. To study these associations in more detail, it is highly important that the WM tracts can be accurately and reproducibly characterized from brain diffusion MRI. In addition, to enable analysis of WM tracts in large datasets and in clinical practice it is essential to have methodology that is fast and easy to apply. This work therefore presents a new approach for WM tract segmentation: Neuro4Neuro, that is capable of direct extraction of WM tracts from diffusion tensor images using convolutional neural network (CNN). This 3D end-to-end method is trained to segment 25 WM tracts in aging individuals from a large population-based study (N = 9752, 1. 5T MRI). The proposed method showed good segmentation performance and high reproducibility, i. e. , a high spatial agreement (Cohen’s kappa, κ = 0. 72 − 0. 83 ) and a low scan-rescan error in tract-specific diffusion measures (e. g. , fractional anisotropy: ε = 1 % − 5 % ). The reproducibility of the proposed method was higher than that of a tractography-based segmentation algorithm, while being orders of magnitude faster (0. 5s to segment one tract). In addition, we showed that the method successfully generalizes to diffusion scans from an external dementia dataset (N = 58, 3T MRI). In two proof-of-principle experiments, we associated WM microstructure obtained using the proposed method with age in a normal elderly population, and with disease subtypes in a dementia cohort. In concordance with the literature, results showed a widespread reduction of microstructural organization with aging and substantial group-wise microstructure differences between dementia subtypes. In conclusion, we presented a highly reproducible and fast method for WM tract segmentation that has the potential of being used in large-scale studies and clinical practice.

Details DOI

YNICL Journal 2018 Journal Article

Automatic normative quantification of brain tissue volume to support the diagnosis of dementia: A clinical evaluation of diagnostic accuracy

Meike W. Vernooij
Bas Jasperse
Rebecca Steketee
Marcel Koek
Henri Vrooman
M. Arfan Ikram
Janne Papma
Aad van der Lugt

Objectives: To assesses whether automated brain image analysis with quantification of structural brain changes improves diagnostic accuracy in a memory clinic setting. Methods: = 4915, aged >45 years), and were expressed as age- and sex-specific percentiles. Experienced radiologists determined the most likely imaging-based diagnosis based on structural brain MRI using three strategies (visual assessment of MRI only, quantitative normative information only, or a combination of both). Diagnostic accuracy of each strategy was calculated with the clinical diagnosis as the reference standard. Results: Providing radiologists with only quantitative data decreased diagnostic accuracy both for AD and FTD compared to conventional visual rating. The combination of quantitative with visual information, however, led to better diagnostic accuracy compared to only visual ratings for AD. This was not the case for FTD. Conclusion: Quantitative assessment of structural brain MRI combined with a reference standard in addition to standard visual assessment may improve diagnostic accuracy in a memory clinic setting.

Details DOI

YNICL Journal 2016 Journal Article

Structural and functional brain abnormalities place phenocopy frontotemporal dementia (FTD) in the FTD spectrum

Rebecca M.E. Steketee
Rozanna Meijboom
Esther E. Bron
Robert Jan Osse
Inge de Koning
Lize C. Jiskoot
Stefan Klein
Frank Jan de Jong

PURPOSE: 'Phenocopy' frontotemporal dementia (phFTD) patients may clinically mimic the behavioral variant of FTD (bvFTD), but do not show functional decline or abnormalities upon visual inspection of routine neuroimaging. We aimed to identify abnormalities in gray matter (GM) volume and perfusion in phFTD and to assess whether phFTD belongs to the FTD spectrum. We compared phFTD patients with both healthy controls and bvFTD patients. MATERIALS & METHODS: Seven phFTD and 11 bvFTD patients, and 20 age-matched controls underwent structural T1-weighted magnetic resonance imaging (MRI) and 3D pseudo-continuous arterial spin labeling (pCASL) at 3T. Normalized GM (nGM) volumes and perfusion, corrected for partial volume effects, were quantified regionally as well as in the entire supratentorial cortex, and compared between groups taking into account potential confounding effects of gender and scanner. RESULTS: PhFTD patients showed cortical atrophy, most prominently in the right temporal lobe. Apart from this regional atrophy, GM volume was generally not different from either controls or from bvFTD. BvFTD however showed extensive frontotemporal atrophy. Perfusion was increased in the left prefrontal cortex compared to bvFTD and to a lesser extent to controls. CONCLUSION: PhFTD and bvFTD show overlapping cortical structural abnormalities indicating a continuum of changes especially in the frontotemporal regions. Together with functional changes suggestive of a compensatory response to incipient pathology in the left prefrontal regions, these findings are the first to support a possible neuropathological etiology of phFTD and suggest that phFTD may be a neurodegenerative disease on the FTD spectrum.

Details DOI

JBHI Journal 2015 Journal Article

Feature Selection Based on the SVM Weight Vector for Classification of Dementia

Esther E. Bron
Marion Smits
Wiro J. Niessen
Stefan Klein

Computer-aided diagnosis of dementia using a support vector machine (SVM) can be improved with feature selection. The relevance of individual features can be quantified from the SVM weights as a significance map (p-map). Although these p-maps previously showed clusters of relevant voxels in dementia-related brain regions, they have not yet been used for feature selection. Therefore, we introduce two novel feature selection methods based on p-maps using a direct approach (filter) and an iterative approach (wrapper). To evaluate these p-map feature selection methods, we compared them with methods based on the SVM weight vector directly, t-statistics, and expert knowledge. We used MRI data from the Alzheimer's disease neuroimaging initiative classifying Alzheimer's disease (AD) patients, mild cognitive impairment (MCI) patients who converted to AD (MCIc), MCI patients who did not convert to AD (MCInc), and cognitively normal controls (CN). Features for each voxel were derived from gray matter morphometry. Feature selection based on the SVM weights gave better results than t-statistics and expert knowledge. The p-map methods performed slightly better than those using the weight vector. The wrapper method scored better than the filter method. Recursive feature elimination based on the p-map improved most for AD-CN: the area under the receiver-operating-characteristic curve (AUC) significantly increased from 90. 3% without feature selection to 92. 0% when selecting 1. 5%–3% of the features. This feature selection method also improved the other classifications: AD-MCI 0. 1% improvement in AUC (not significant), MCI-CN 0. 7%, and MCIc-MCInc 0. 1% (not significant). Although the performance improvement due to feature selection was limited, the methods based on the p-map generally had the best performance, and were therefore better in estimating the relevance of individual features.

Details DOI

YNIMG Journal 2015 Journal Article

Standardized evaluation of algorithms for computer-aided diagnosis of dementia based on structural MRI: The CADDementia challenge

Esther E. Bron
Marion Smits
Wiesje M. van der Flier
Hugo Vrenken
Frederik Barkhof
Philip Scheltens
Janne M. Papma
Rebecca M.E. Steketee

Algorithms for computer-aided diagnosis of dementia based on structural MRI have demonstrated high performance in the literature, but are difficult to compare as different data sets and methodology were used for evaluation. In addition, it is unclear how the algorithms would perform on previously unseen data, and thus, how they would perform in clinical practice when there is no real opportunity to adapt the algorithm to the data at hand. To address these comparability, generalizability and clinical applicability issues, we organized a grand challenge that aimed to objectively compare algorithms based on a clinically representative multi-center data set. Using clinical practice as the starting point, the goal was to reproduce the clinical diagnosis. Therefore, we evaluated algorithms for multi-class classification of three diagnostic groups: patients with probable Alzheimer's disease, patients with mild cognitive impairment and healthy controls. The diagnosis based on clinical criteria was used as reference standard, as it was the best available reference despite its known limitations. For evaluation, a previously unseen test set was used consisting of 354 T1-weighted MRI scans with the diagnoses blinded. Fifteen research teams participated with a total of 29 algorithms. The algorithms were trained on a small training set (n=30) and optionally on data from other sources (e. g. , the Alzheimer's Disease Neuroimaging Initiative, the Australian Imaging Biomarkers and Lifestyle flagship study of aging). The best performing algorithm yielded an accuracy of 63. 0% and an area under the receiver-operating-characteristic curve (AUC) of 78. 8%. In general, the best performances were achieved using feature extraction based on voxel-based morphometry or a combination of features that included volume, cortical thickness, shape and intensity. The challenge is open for new submissions via the web-based framework: http: //caddementia. grand-challenge. org.

Details DOI

YNICL Journal 2013 Journal Article

The influence of cerebral small vessel disease on default mode network deactivation in mild cognitive impairment

Janne M. Papma
Tom den Heijer
Inge de Koning
Francesco U. Mattace-Raso
Aad van der Lugt
Fedde van der Lijn
John C. van Swieten
Peter J. Koudstaal

Introduction Cerebral small vessel disease (CSVD) is thought to contribute to cognitive dysfunction in patients with mild cognitive impairment (MCI). The underlying mechanisms, and more specifically, the effects of CSVD on brain functioning in MCI are incompletely understood. The objective of the present study was to examine the effects of CSVD on brain functioning, activation and deactivation, in patients with MCI using task-related functional MRI (fMRI). Methods We included 16 MCI patients with CSVD, 26 MCI patients without CSVD and 25 controls. All participants underwent a physical and neurological examination, neuropsychological testing, structural MRI, and fMRI during a graded working memory paradigm. Results MCI patients with and without CSVD had a similar neuropsychological profile and task performance during fMRI, but differed with respect to underlying (de)activation patterns. MCI patients with CSVD showed impaired deactivation in the precuneus/posterior cingulate cortex, a region known to be involved in the default mode network. In MCI patients without CSVD, brain activation depended on working memory load, as they showed relative ‘hyperactivation’ during vigilance, and ‘hypoactivation’ at a high working memory load condition in working memory related brain regions. Conclusions We present evidence that the potential underlying mechanism of CSVD affecting cognition in MCI is through network interference. The observed differences in brain activation and deactivation between MCI patients with and without CSVD, who had a similar ‘clinical phenotype’, support the view that, in patients with MCI, different types of pathology can contribute to cognitive impairment through different pathways.

Details DOI

YNIMG Journal 2011 Journal Article

Neurobiological substrate of smoking-related attentional bias

Maartje Luijten
Dick J. Veltman
Wim van den Brink
Rob Hester
Matt Field
Marion Smits
Ingmar. H.A. Franken

Substance-dependent patients automatically and involuntarily allocate their attention to drug cues in the environment, a process referred to as attentional bias. Attentional bias is increased during periods of subjective craving and predictive of treatment outcome and relapse in substance dependence. Despite recent theoretical and clinical advances with regard to attentional bias, the underlying neurobiological mechanisms are largely unknown. The objective of the current study was to investigate the neural substrate of attentional bias and associated subjective craving in smokers. A group of smokers (n =20) and a group of age- and gender-matched nonsmoking controls (n =22) were recruited from the general population and participated in a single session of fMRI scanning while attentional processes were manipulated. Main outcome measures were blood oxygen level-dependent (BOLD) fMRI activation during an attentional bias paradigm and self-reported cigarette craving. Results of the current study show that the dorsal anterior cingulate cortex, the superior parietal gyrus, and the superior temporal gyrus were more strongly activated in smokers, as compared to controls, when they had to pay attention to task-relevant information (line counting) while smoking cues were present as distracters (attentional bias). Subjective craving measures during attentional bias correlated with brain activation in the insula and putamen. To our knowledge, this is the first controlled study that shows the brain regions involved in attentional bias in smokers. The current study demonstrates that brain regions contributing to top-down attentional processing are implicated in attentional bias in smokers, suggesting that smokers have to employ more attentional resources to focus on a standard cognitive task when smoking cues are present.

Details DOI

YNIMG Journal 2011 Journal Article

Remedial action and feedback processing in a time-estimation task: Evidence for a role of the rostral cingulate zone in behavioral adjustments without learning

Frederik M. van der Veen
Christian H. Röder
Gabry W. Mies
Aad van der Lugt
Marion Smits

The present study examined the role of the rostral cingulate zone (RCZ) in feedback processing, and especially focused on effects of modality of the feedback stimulus and remedial action. Participants performed a time-estimation task in which they had to estimate a 1-second interval. After the estimation participants received verbal (correct/false) or facial (fearful face/happy face) feedback. Percentage of positive and negative feedback was kept at 50% by dynamically adjusting the interval in which estimations were labeled correct. Contrary to predictions of the reinforcement learning theory, which predicts more RCZ activation when the outcome of behavior is worse than expected, we found that the RCZ was more active after positive feedback than after negative feedback, independent of the modality of the feedback stimulus. More in line with the suggested role of the RCZ in reinforcement learning was the finding that the RCZ was more active after negative feedback that was followed by a correct adjustment as compared to negative feedback followed by an incorrect adjustment. Both findings can be explained in terms of the RCZ being involved in facilitating remedial action as opposed to the suggested signaling function (outcome is worse than expected) proposed by the reinforcement learning theory.

Details DOI

YNIMG Journal 2011 Journal Article

The anterior cingulate cortex responds differently to the validity and valence of feedback in a time-estimation task

Gabry W. Mies
Maurits W. van der Molen
Marion Smits
Michiel W. Hengeveld
Frederik M. van der Veen

This study examined the role of the medial frontal cortex in the processing of valence and validity of performance feedback using a time-estimation paradigm. Participants had to produce 1s intervals followed by positive and negative feedback that could be valid or invalid (i. e. , related or unrelated to task performance). Performance results showed that participants used the validity information to adjust their time estimations to negative feedback. The rostral cingulate zone (RCZ) was more active after valid feedback than after invalid feedback, but was insensitive to the valence of the feedback. The rostral anterior cingulate cortex (rACC), posterior cingulate and right superior frontal gyrus, however, appeared to be primarily sensitive to the valence of the feedback; being more active after positive feedback. The results are discussed along the lines of the ACC's cognitive and affective subdivisions and their structural and functional connections.

Details DOI