Dataset Information

Prediction of the information processing speed performance in multiple sclerosis using a machine learning approach in a large multicenter magnetic resonance imaging data set.

ABSTRACT: Many patients with multiple sclerosis (MS) experience information processing speed (IPS) deficits, and the Symbol Digit Modalities Test (SDMT) has been recommended as a valid screening test. Magnetic resonance imaging (MRI) has markedly improved the understanding of the mechanisms associated with cognitive deficits in MS. However, which structural MRI markers are the most closely related to cognitive performance is still unclear. We used the multicenter 3T-MRI data set of the Italian Neuroimaging Network Initiative to extract multimodal data (i.e., demographic, clinical, neuropsychological, and structural MRIs) of 540 MS patients. We aimed to assess, through machine learning techniques, the contribution of brain MRI structural volumes in the prediction of IPS deficits when combined with demographic and clinical features. We trained and tested the eXtreme Gradient Boosting (XGBoost) model following a rigorous validation scheme to obtain reliable generalization performance. We carried out a classification and a regression task based on SDMT scores feeding each model with different combinations of features. For the classification task, the model trained with thalamus, cortical gray matter, hippocampus, and lesions volumes achieved an area under the receiver operating characteristic curve of 0.74. For the regression task, the model trained with cortical gray matter and thalamus volumes, EDSS, nucleus accumbens, lesions, and putamen volumes, and age reached a mean absolute error of 0.95. In conclusion, our results confirmed that damage to cortical gray matter and relevant deep and archaic gray matter structures, such as the thalamus and hippocampus, is among the most relevant predictors of cognitive performance in MS.

SUBMITTER: Marzi C

PROVIDER: S-EPMC9783441 | biostudies-literature | 2023 Jan

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Prediction of the information processing speed performance in multiple sclerosis using a machine learning approach in a large multicenter magnetic resonance imaging data set.

Marzi Chiara C d'Ambrosio Alessandro A Diciotti Stefano S Bisecco Alvino A Altieri Manuela M Filippi Massimo M Rocca Maria Assunta MA Storelli Loredana L Pantano Patrizia P Tommasin Silvia S Cortese Rosa R De Stefano Nicola N Tedeschi Gioacchino G Gallo Antonio A

Human brain mapping 20221018 1

Many patients with multiple sclerosis (MS) experience information processing speed (IPS) deficits, and the Symbol Digit Modalities Test (SDMT) has been recommended as a valid screening test. Magnetic resonance imaging (MRI) has markedly improved the understanding of the mechanisms associated with cognitive deficits in MS. However, which structural MRI markers are the most closely related to cognitive performance is still unclear. We used the multicenter 3T-MRI data set of the Italian Neuroimagin ...[more]

PMID: 36255155

Similar Datasets

Project description:Cognitive impairment affects about 50% of multiple sclerosis (MS) patients, but the mechanisms underlying this remain unclear. The default mode network (DMN) has been linked with cognition, but in MS its role is still poorly understood. Moreover, within an extended DMN network including the cerebellum (CBL-DMN), the contribution of cortico-cerebellar connectivity to MS cognitive performance remains unexplored. The present study investigated associations of DMN and CBL-DMN structural connectivity with cognitive processing speed in MS, in both cognitively impaired (CIMS) and cognitively preserved (CPMS) MS patients. 68 MS patients and 22 healthy controls (HCs) completed a symbol digit modalities test (SDMT) and had 3T brain magnetic resonance imaging (MRI) scans that included a diffusion weighted imaging protocol. DMN and CBL-DMN tracts were reconstructed with probabilistic tractography. These networks (DMN and CBL-DMN) and the cortico-cerebellar tracts alone were modeled using a graph theoretical approach with fractional anisotropy (FA) as the weighting factor. Brain parenchymal fraction (BPF) was also calculated. In CIMS SDMT scores strongly correlated with the FA-weighted global efficiency (GE) of the network [GE(CBL-DMN): ρ = 0.87, R 2 = 0.76, p < 0.001; GE(DMN): ρ = 0.82, R 2 = 0.67, p < 0.001; GE(CBL): ρ = 0.80, R 2 = 0.64, p < 0.001]. In CPMS the correlation between these measures was significantly lower [GE(CBL-DMN): ρ = 0.51, R 2 = 0.26, p < 0.001; GE(DMN): ρ = 0.48, R 2 = 0.23, p = 0.001; GE(CBL): ρ = 0.52, R 2 = 0.27, p < 0.001] and SDMT scores correlated most with BPF (ρ = 0.57, R 2 = 0.33, p < 0.001). In a multivariable regression model where SDMT was the independent variable, FA-weighted GE was the only significant explanatory variable in CIMS, while in CPMS BPF and expanded disability status scale were significant. No significant correlation was found in HC between SDMT scores, MRI or network measures. DMN structural GE is related to cognitive performance in MS, and results of CBL-DMN suggest that the cerebellum structural connectivity to the DMN plays an important role in information processing speed decline.

Project description:IntroductionMultiple sclerosis (MS) is a chronic autoimmune-mediated demyelinating disease of the central nervous system (CNS). A clinical presentation of the disease is highly differentiated even from the earliest stages of the disease. The application of stratifying tests in clinical practice would allow for improving clinical decision-making including a proper assessment of treatment benefit/risk balance.MethodsThis prospective study included patients with MS diagnosed up to 1 year before recruitment. We analyzed serum biomarkers such as CXCL13, CHI3L1, OPN, IL-6, and GFAP and neurofilament light chains (NfLs); brain MRI parameters of linear atrophy such as bicaudate ratio (BCR), third ventricle width (TVW); and information processing speed were measured using the Symbol Digit Modalities Test (SDMT) during the 2 years follow-up.ResultsThe study included a total of 50 patients recruited shortly after the diagnosis of MS diagnosis (median 0 months; range 0-11 months), and the mean time of observation was 28 months (SD = 4.75). We observed a statistically significant increase in the EDSS score (Wilcoxon test: Z = 3.06, p = 0.002), BCR (Wilcoxon test: Z = 4.66, p < 0.001), and TVW (Wilcoxon test: Z = 2.84, p = 0.005) after 2 years of disease. Patients who had a significantly higher baseline level of NfL suffered from a more severe disease course as per the EDSS score (Mann-Whitney U-test: U = 107, Z = -2,74, p = 0.006) and presence of relapse (Mann-Whitney U-test: U = 188, Z = -2.01, p = 0.044). In the logistic regression model, none of the parameters was a significant predictor for the achieving of no evidence of disease activity status (NEDA). In the model considering all assessed parameters, only the level of NfL had a significant impact on disease progression, measured as the increase in EDSS (logistic regression: β = 0.002, p = 0.017).ConclusionWe confirmed that NfL levels in serum are associated with more active disease. Moreover, we found that TVW at the time of diagnosis was associated with an impairment in cognitive function measured by information processing speed at the end of the 2-year observation. The inclusion of serum NfL and TVW assessment early in the disease may be a good predictor of disease progression independent of NEDA.

Project description:BackgroundIn participants with major depressive disorder who are trained to upregulate their amygdalar hemodynamic responses during positive autobiographical memory recall with real-time functional magnetic resonance imaging neurofeedback (rtfMRI-nf) training, depressive symptoms diminish. This study tested whether amygdalar rtfMRI-nf also changes emotional processing of positive and negative stimuli in a variety of behavioral and imaging tasks.MethodsPatients with major depressive disorder completed two rtfMRI-nf sessions (18 received amygdalar rtfMRI-nf, 16 received control parietal rtfMRI-nf). One week before and following rtfMRI-nf training, participants performed tasks measuring responses to emotionally valenced stimuli including a backward-masking task, which measures the amygdalar hemodynamic response to emotional faces presented for traditionally subliminal duration and followed by a mask, and the Emotional Test Battery in which reaction times and performance accuracy are measured during tasks involving emotional faces and words.ResultsDuring the backward-masking task, amygdalar responses increased while viewing masked happy faces but decreased to masked sad faces in the experimental versus control group following rtfMRI-nf. During the Emotional Test Battery, reaction times decreased to identification of positive faces and during self-identification with positive words and vigilance scores increased to positive faces and decreased to negative faces during the faces dot-probe task in the experimental versus control group following rtfMRI-nf.ConclusionsrtfMRI-nf training to increase the amygdalar hemodynamic response to positive memories was associated with changes in amygdalar responses to happy and sad faces and improved processing of positive stimuli during performance of the Emotional Test Battery. These results may suggest that amygdalar rtfMRI-nf training alters responses to emotional stimuli in a manner similar to antidepressant pharmacotherapy.

Dataset Information

Prediction of the information processing speed performance in multiple sclerosis using a machine learning approach in a large multicenter magnetic resonance imaging data set.

Publications

Prediction of the information processing speed performance in multiple sclerosis using a machine learning approach in a large multicenter magnetic resonance imaging data set.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets