Monica Roascio’s research while affiliated with University of Genoa 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 (14)


Preprocessing pipeline: overview of the main preliminary image processing steps performed on (A) 3D T1-weighted, whose key step is skull-stripping and (B) HARDI scans, whose core is represented by denoising as well as distortion correction, for an example subject.
An intuitive visualization of Canonical Correlation Analysis: Let N be the number of observations. n datasets—variable depending on each diffusion model—Xk ∈ RNxVk are transformed by projections Wk ∈ R VkxD such that each paired embedding (Ai, Aj) is maximally correlated with unit length in the projected space.
Microstructural models: parametric scalar maps derived from all the HARDI models employed for this study: (A) Diffusion Kurtosis Imaging (DKI), (B) Neurite Orientation Dispersion and Density Imaging (NODDI), (C) Fiber Orientation Estimated using Continuous Axially Symmetric Tensors (FORECAST), (D) Multi-Shell Multi-Tissue Constrained Spherical Deconvolution (MSMT CSD).
TBSS pipeline: overview of the main steps of the TBSS framework, from spatial normalization of DTI volumes to bootstrapping the within-population template to skeletonization of the template DTI-FA map and projection of each subject's DTI-FA onto the skeleton.
Experimental design for SVM classification: in a first phase, an SVM classification estimator is chosen to best perform on DTI-FA skeletonized data; in a second phase the, selected model is extended to other non-FA measures.

+6

Multi-view fusion of diffusion MRI microstructural models: a preterm birth study
  • Article
  • Full-text available

December 2024

·

19 Reads

·

Monica Roascio

·

·

[...]

·

Objective High Angular Resolution Diffusion Imaging (HARDI) models have emerged as a valuable tool for investigating microstructure with a higher degree of detail than standard diffusion Magnetic Resonance Imaging (dMRI). In this study, we explored the potential of multiple advanced microstructural diffusion models for investigating preterm birth in order to identify non-invasive markers of altered white matter development. Approach Rather than focusing on a single MRI modality, we studied on a compound of HARDI techniques in 46 preterm babies studied on a 3T scanner at term-equivalent age and in 23 control neonates born at term. Furthermore, we investigated discriminative patterns of preterm birth using multiple analysis methods, drawn from two only seemingly divergent modeling goals, namely inference and prediction. We thus resorted to (i) a traditional univariate voxel-wise inferential method, as the Tract-Based Spatial Statistics (TBSS) approach; (ii) a univariate predictive approach, as the Support Vector Machine (SVM) classification; and (iii) a multivariate predictive Canonical Correlation Analysis (CCA). Main results The TBSS analysis revealed significant differences between preterm and term cohorts in several white matter areas for multiple HARDI features. SVM classification on skeletonized HARDI measures yielded satisfactory accuracy, particularly for highly informative parameters about fiber directionality. Assessment of the degree of overlap between the two methods in voting for the most discriminating features exhibited a good, though parameter-dependent, rate of agreement. Finally, CCA identified joint changes precisely for those measures exhibiting less correspondence between TBSS and SVM. Significance Our results suggest that a data-driven intramodal imaging approach is crucial for gathering deep and complementary information. The main contribution of this methodological outline is to thoroughly investigate prematurity-related white matter changes through different inquiry focuses, with a view to addressing this issue, both aiming toward mechanistic insight and optimizing predictive accuracy.

Download

Figure 4. Average fraction of sleep events out of the total duration of sleep separated for patients without (green) and w lesions (red) visible in the MRI. Boxes present median (black line) and inter-quartile (black whiskers). Statist significance is shown above individual tests with the relative pvalues.
Sleep architecture correlates with neurological and neurobehavioral short- and mid-term outcome in a sample of very preterm infants

December 2024

·

53 Reads

Newborns spend most of their time sleeping. This activity fosters neurodevelopment. Prematurity, defined by birth occurring prior the 37th week of gestation, disrupts normal brain in-utero programming, with long-lasting consequences that carry a high social burden. Sleep alterations may contribute to these sequelae. In this pilot study we aimed to describe the 24-hours distribution of sleep states among very preterm infants (VPI), and to correlate it with neurobehavioral assessment up to 6 months of corrected age (CA). Secondly, we aimed to assess if the presence of a brain lesion detected at MRI could affect sleep duration, architecture, and quality. Ten VPI were assessed at 32 weeks PMA with a 24-hours video-polysomnography and received a neurobehavioral examination at the time of the recording, at term equivalent age (TEA), and at 6 months CA. Analysis of sleep stages distribution and transitions, and power spectra were conducted. Total sleep time and amount of quiet sleep positively correlated with neurological, and neurobehavioral assessment at 32 weeks PMA, at TEA, and at 6 months CA, while Sleep Onset Active Sleep (SOAS) had a negative association. Infants carrying brain lesions showed lower Total Sleep time accompanied by a higher prevalence of AS+SOAS and showed a gradient for higher power of posterior slow activity (slow δ and δ) during SOAS in the left hemisphere posterior regions. Understanding sleep dynamics among preterm infants might provide future therapeutic/management strategies, which need to encompass sleep care.


EEG-based Machine Learning Models for the Prediction of Phenoconversion Time and Subtype in iRBD

February 2024

·

47 Reads

·

7 Citations

Sleep

Study Objectives Isolated rapid eye movement sleep behavior disorder (iRBD) is a prodromal stage of α-synucleinopathies and eventually phenoconverts to overt neurodegenerative diseases including Parkinson’s disease (PD), dementia with Lewy bodies (DLB), and multiple system atrophy (MSA). Associations of baseline resting-state electroencephalography (EEG) with phenoconversion have been reported. In this study, we aimed to develop machine learning models to predict phenoconversion time and subtype using baseline EEG features in patients with iRBD. Methods At baseline, resting-state EEG and neurological assessments were performed on patients with iRBD. Calculated EEG features included spectral power, weighted phase lag index, and Shannon entropy. Three models were used for survival prediction, and four models were used for α-synucleinopathy subtype prediction. The models were externally validated using data from a different institution. Results A total of 236 iRBD patients were followed up for up to 8 years (mean 3.5 years), and 31 patients converted to α-synucleinopathies (16 PD, 9 DLB, 6 MSA). The best model for survival prediction was the random survival forest model with an integrated Brier score of 0.114 and a concordance index of 0.775. The K-nearest neighbor model was the best model for subtype prediction with an area under the receiver operating characteristic curve of 0.901. Slowing of the EEG was an important feature for both models. Conclusions Machine learning models using baseline EEG features can be used to predict phenoconversion time and its subtype in patients with iRBD. Further research including large sample data from many countries is needed to make a more robust model.


EEG-based Machine Learning Models for the Prediction of Phenoconversion Time and Subtype in iRBD

September 2023

·

52 Reads

·

1 Citation

Background: Idiopathic/Isolated rapid eye movement sleep behavior disorder (iRBD) is a prodromal stage of alpha-synucleinopathies and eventually phenoconverts to overt neurodegenerative diseases including Parkinson's disease (PD), dementia with Lewy bodies (DLB) and multiple system atrophy (MSA). Associations of baseline resting-state electroencephalography (EEG) with phenoconversion have been reported. Objectives: In this study, we aimed to develop machine learning models to predict phenoconversion time and subtype using baseline EEG features in patients with iRBD. Methods: At baseline, resting-state EEG and neurological assessments were performed on patients with iRBD. Calculated EEG features included spectral power, weighted phase lag index and Shannon entropy. Three models were used for survival prediction, and four models were used for alpha-synucleinopathy subtype prediction. The models were externally validated using data from a different institution. Results: A total of 236 iRBD patients were followed-up for up to eight years (mean 3.5 years), and 31 patients converted to alpha-synucleinopathies (16 PD, 9 DLB, 6 MSA). The best model for survival prediction was the random survival forest model with an integrated Brier score of 0.114 and a concordance index of 0.775. The K-nearest neighbor model was the best model for subtype prediction with an area under the receiver operating characteristic curve of 0.901. EEG slowing was an important feature for both models. Conclusions: Machine learning models using baseline EEG features can be used to predict phenoconversion time and its subtype in patients with iRBD. Further research including large sample data from many countries is needed to make a more robust model.


Altered Brain Dynamics in idiopathic REM sleep behavior disorder: Implications for a continuum from prodromal to overt alpha-synucleinopathies

May 2023

·

91 Reads

Idiopathic/isolated REM sleep behavior disorder (iRBD) is considered a prodromal stage of alpha-synucleinopathies. Cortical and sub-cortical brain modifications begin years before the emergence of overt neurodegenerative symptoms. To better understand the pathophysiological process impacting the brain from the prodromal to the overt stage of alpha-synucleinopathy, it is essential to assess iRBD patients over time. Recent evidence suggests that the human brain operates at an operating point near a critical phase transition between subcritical and supercritical phases in the system′s state space to maintain cognitive and physiological performance. In contrast, a deviation from the critical regime leading to altered oscillatory dynamics has been observed in several pathologies. Here, we investigated if the alpha-synucleinopathy produces a deviation of the operating point already evident in the prodromal phase and if this shift correlates with biological and clinical disease severity. We analyzed a dataset of 59 patients with iRBD (age 69.61 +/- 6.98, 50 male) undergoing resting-state high-density EEG, presynaptic dopaminergic imaging, and clinical evaluations. Thirty-one patients (age 72.41 +/- 7.05, 31 male) also underwent clinical and instrumental follow-up (mean follow-up period 25.85 +/- 10.20 months). To localize the individual operating points along the excitation-inhibition (EI) continuum, we assessed both measures of neuronal EI balance and measures of critical brain dynamics such as long-range temporal correlation (LRTCs) and neuronal bistability in spontaneous narrow-band oscillations. Finally, we correlated critical brain dynamics and EI balance metrics with phase synchronization, nigro-striatal dopaminergic functioning, and clinical performances. Compared to 48 healthy subjects (age 70.25 +/- 10.15, 23 male), iRBD patients showed higher values of LRTCs and bistability in the 2-7 Hz band at diagnosis. Patients who eventually phenoconverted to overt alpha-synucleinopathy exhibited a more excitation-dominated (fEI > 1) condition than stable iRBD patients in 5-7 Hz. This higher excitation also directly correlated with phase synchronization in 2-7 Hz, further suggesting a shift of the operating point toward a supercritical state with the disease progression. Moreover, excitation-dominated state and low bistability were associated with deterioration of the nigro-striatal dopaminergic function and tended to correlate with stronger clinical symptoms. In conclusion, this study shows for the first time a deviation of the working point from inhibition- to excitation-dominated states along the continuum from prodromal to overt phases of the disease. These cortical brain dynamics modifications are associated with nigro-striatal dopaminergic impairment. These results increase our knowledge of the physiopathological process underlying alpha-synucleinopathies since prodromal stages, possibly providing new clues on disease-modifying strategies.


Influence of adaptive denoising on Diffusion Kurtosis Imaging at 3T and 7T

March 2023

·

24 Reads

·

1 Citation

Computer Methods and Programs in Biomedicine

Background and objective: Choosing the most appropriate denoising method to improve the quality of diagnostic images maximally is key in pre-processing of diffusion MRI images. Recent advancements in acquisition and reconstruction techniques have questioned traditional noise estimation methods favoring adaptive denoising frameworks, circumventing the need to know a priori information that is hardly available in a clinical setting. In this observational study, we compared two innovative adaptive techniques sharing some features, Patch2Self and Nlsam, through application on reference adult data at 3T and 7T. The primary aim was identifying the most effective method in case of Diffusion Kurtosis Imaging (DKI) data - particularly susceptible to noise and signal fluctuations - at 3T and 7T fields. A side goal consisted of investigating the dependence of kurtosis metrics' variability with respect to the magnetic field on the adopted denoising methodology. Methods: For comparison purposes, we focused on qualitative and quantitative analysis of DKI data and related microstructural maps before and after applying the two denoising approaches. Specifically, we assessed computational efficiency, preservation of anatomical details via perceptual metrics, consistency of microstructure model fitting, alleviation of degeneracies in model estimation, and joint variability with varying field strength and denoising method. Results: Accounting for all these factors, Patch2Self framework has turned out to be specifically suitable for DKI data, with improving performance at 7T. Nlsam method is more robust in alleviating degenerate black voxels while introducing some blurring, which in turn is reflected in an overall loss of image sharpness. Regarding the impact of denoising on field-dependent variability, both methods have been shown to make variations from standard to Ultra-High Field more concordant with theoretical evidence, claiming that kurtosis metrics are sensitive to susceptibility-induced background gradients, directly proportional to the magnetic field strength and sensitive to the microscopic distribution of iron and myelin. Conclusions: This study serves as a proof-of-concept stressing the need for an accurate choice of a denoising methodology, specifically tailored for the data under analysis and allowing higher spatial resolution acquisition within clinically compatible timings, with all the potential benefits that improving suboptimal quality of diagnostic images entails.


Data-driven characterization of Preterm Birth through intramodal Diffusion MRI

January 2023

·

33 Reads

Preterm birth still represents a concrete emergency to be managed and addressed globally. Since cerebral white matter injury is the major form of brain impairment in survivors of premature birth, the identification of reliable, non-invasive markers of altered white matter development is of utmost importance in diagnostics. Diffusion MRI has recently emerged as a valuable tool to investigate these kinds of alterations. In this work, rather than focusing on a single MRI modality, we worked on a compound of beyond-DTI High Angular Resolution Diffusion Imaging (HARDI) techniques in a group of 46 preterm babies studied on a 3T scanner at term equivalent age and in 23 control neonates born at term. After extracting relevant derived parameters, we examined discriminative patterns of preterm birth through (i) a traditional voxel-wise statistical method such as the Tract-Based Spatial Statistics approach (TBSS); (ii) an advanced Machine Learning approach such as the Support Vector Machine (SVM) classification; (iii) establishing the degree of association between the two methods in voting white matter most discriminating areas. Finally, we applied a multi-set Canonical Correlation Analysis (CCA) in search for sources of linked alterations across modalities. TBSS analysis showed significant differences between preterm and term cohorts in several white matter areas for multiple HARDI features. SVM classification performed on skeletonized HARDI measures produced satisfactory accuracy rates, especially as for highly informative parameters about fibers directionality. Assessment of the degree of overlap between the relevant measures identified by the two methods exhibited a good, though parameter-dependent rate of agreement. Finally, CCA analysis identified joint changes precisely for those features exhibiting less correspondence between TBSS and SVM. Our results suggest that a data-driven intramodal imaging approach is crucial to extract deep and complementary information that cannot be extracted from a single modality.



Figure 2: EEG features change in iRBD patients. Group-level averaged (continuous lines) and single-subject (dotted lines) of (a) PSD, (b) wPLI, and (c) oCC for 2-classes: RBD (green) and controls (black). Shaded areas represent confidence intervals at 5% around population mean (bootstrap, N = 1000). Dashed lines in (b) and (c) represent surrogate averages. Empty black circles show a significant difference between people with RBD and healthy subjects (Kruskal Wallis test). Filled black circles show a significant difference between the two population after the BH correction. Legend. Power spectral density -PSD, weighted Phase Lag IndexwPLI, orthogonalized Correlation Coefficient -oCC, idiopathic Rapid eye-movement sleep Behavior Disorder -iRBD, Healthy control subjects -Controls.
Figure 3: Large-scale EEG features improve the robustness and performance of the classifier. (a) Violin plots of F1-score of classification model with 7 different inputs of variables as reported on the x-axis. For each model, we used three different shuffling (yellow, orange, and light red, respectively) to shuffle data before the split in learning and testing sets. (b-e) Averaged and standard deviation across shuffling of (b) f1-score (red), (c) accuracy (blue), (d) precision (purple), and (e) recall (pink). Legend. Power spectral density -PSD, weighted Phase Lag Index -wPLI, orthogonalized Correlation Coefficient -oCC, Electroencephalographic -EEG.
Figure 4: Performance increased by adopting less than 50% of the features as input variables. Averaged and standard deviation of (a) f1-score (red), (b) accuracy (blue), (c) precision (purple), and (d) recall (pink) across shuffling and split for different input variables. The first bubble for each group of features shows the performance of the classifier without variable selection. Other bubbles show the classifier performance with features selection (the number of selected variables is shown on the x-axis). The bubble size shows the percentage of selected variables, while the bubble color shows the performance of the model. On the y-axis, it was shown the group of features that are used as input variables to carry out the RBD/healthy classification. Legend. Power spectral density -PSD, weighted Phase Lag Index -wPLI, orthogonalized Correlation Coefficient -oCC, idiopathic Rapid eye-movement sleep Behavior Disorder -iRBD.
Main clinical and demographic data of idiopathic RBD patients and HC subjects. Legend. Idiopathic Rapid eye-movement
Large-scale network metrics improve the classification performance of rapid-eye-movement sleep behavior disorder patients

August 2022

·

74 Reads

·

1 Citation

Clinical decision support systems based on machine-learning algorithms are largely applied in the context of the diagnosis of neurodegenerative diseases (NDDs). While recent models yield robust classifications in supervised two classes-problems accurately separating Parkinson’s disease (PD) from healthy control (HC) subjects, few works looked at prodromal stages of NDDs. Idiopathic Rapid-eye Movement (REM) sleep behavior disorder (iRBD) is considered a prodromal stage of PD with a high chance of phenoconversion but with heterogeneous symptoms that hinder accurate disease prediction. Machine learning (ML) based methods can be used to develop personalized trajectory models, but these require large amounts of observational points with homogenous features significantly reducing the possible imaging modalities to non-invasive and cost-effective techniques such as high-density electrophysiology (hdEEG). In this work, we aimed at quantifying the increase in accuracy and robustness of the classification model with the inclusion of network-based metrics compared to the classical Fourier-based power spectral density (PSD). We performed a series of analyses to quantify significance in cohort-wise metrics, the performance of classification tasks, and the effect of feature selection on model accuracy. We report that amplitude correlation spectral profiles show the largest difference between iRBD and HC subjects mainly in delta and theta bands. Moreover, the inclusion of amplitude correlation and phase synchronization improves the classification performance by up to 11% compared to using PSD alone. Our results show that hdEEG features alone can be used as potential biomarkers in classification problems using iRBD data and that large-scale network metrics improve the performance of the model. This evidence suggests that large-scale brain network metrics should be considered important tools for investigating prodromal stages of NDD as they yield more information without harming the patient, allowing for constant and frequent longitudinal evaluation of patients at high risk of phenoconversion. Highlights Network-based features are important tools to investigate prodromal stages of PD Amplitude correlation shows the largest difference between two groups in 9/30 bands Amplitude correlation improved up to 11% the performance compared to PSD alone Classification robustness increases when we use both network-based EEG features Classifier performance worsens when PSD is added to network-based EEG features


Diffusion Kurtosis Imaging of Neonatal Spinal Cord in Clinical Routine

May 2022

·

88 Reads

·

2 Citations

Frontiers in Radiology

Diffusion kurtosis imaging (DKI) has undisputed advantages over the more classical diffusion magnetic resonance imaging (dMRI) as witnessed by the fast-increasing number of clinical applications and software packages widely adopted in brain imaging. However, in the neonatal setting, DKI is still largely underutilized, in particular in spinal cord (SC) imaging, because of its inherently demanding technological requirements. Due to its extreme sensitivity to non-Gaussian diffusion, DKI proves particularly suitable for detecting complex, subtle, fast microstructural changes occurring in this area at this early and critical stage of development, which are not identifiable with only DTI. Given the multiplicity of congenital anomalies of the spinal canal, their crucial effect on later developmental outcome, and the close interconnection between the SC region and the brain above, managing to apply such a method to the neonatal cohort becomes of utmost importance. This study will (i) mention current methodological challenges associated with the application of advanced dMRI methods, like DKI, in early infancy, (ii) illustrate the first semi-automated pipeline built on Spinal Cord Toolbox for handling the DKI data of neonatal SC, from acquisition setting to estimation of diffusion measures, through accurate adjustment of processing algorithms customized for adult SC, and (iii) present results of its application in a pilot clinical case study. With the proposed pipeline, we preliminarily show that DKI is more sensitive than DTI-related measures to alterations caused by brain white matter injuries in the underlying cervical SC.


Citations (4)


... Using electroencephalogram (EEG) data, ML models have detected the transition from iRBD to PD with 90.1% accuracy. 92 Arnaldi et al applied ML to analyze clinical data and presynaptic dopaminergic imaging from the international RBD study group, achieving 77% sensitivity and 85% specificity in distinguishing synucleinopathy-related RBD from non-phenotypically transformed RBD. Moreover, their approach demonstrated 85% sensitivity and 86% specificity in differentiating PD converters from dementia with Lewy bodies (DLB). ...

Reference:

From Night to Light: A Bibliometric Analysis of the Global Research Trajectory of Sleep Disorders in Parkinson’s Disease
EEG-based Machine Learning Models for the Prediction of Phenoconversion Time and Subtype in iRBD
  • Citing Article
  • February 2024

Sleep

... A number of markers have been proposed to distinguish patients with iRBD who will convert towards a synucleinopathic disease from those who will not convert [4-11, 29, 35, 58-60]. However, these markers still lack the sensitivity and specificity to predict the disease outcome at the single-subject level [60]. Even when several markers are considered, machine learning models tend to be overfitted, leading to a reproducibility problem, as these models perform poorly when applied to different cohorts [58][59][60]. ...

EEG-based Machine Learning Models for the Prediction of Phenoconversion Time and Subtype in iRBD
  • Citing Preprint
  • September 2023

... For each EEG epoch, the fast Fourier transforms (FFT) using the Hanning window was applied with a frequency of interest range of 1-50 Hz in 0.5 Hz steps. In our study, four frequency bands were used: delta (2-3.5 Hz), theta (4-7.5 Hz), alpha (8-12.5 Hz), beta (13)(14)(15)(16)(17)(18)(19)(20)(21)(22)(23)(24)(25)(26)(27)(28)(29)(30) Hz). Absolute power was averaged across all electrodes and converted decibels. ...

Large-scale network metrics improve the classification performance of rapid-eye-movement sleep behavior disorder patients

... Preliminary EEG analysis demonstrated the slowing of the power spectrum and decreased FC toward lower frequencies as a preclinical index (42,43). It is speculated as an active compensatory mechanism of cognitive impairment in the prodromal stage of synucleinopathies (44). Nevertheless, we did not observe significantly aberrant PSD regardless of the relationship between delta PSD and cognitive performance as well as olfactory function in iRBD. ...

Phase and amplitude EEG correlations change with disease progression in people with idiopathic rapid eye-movement sleep behavior disorder

Sleep