FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W2897982562> ?p ?o ?g. }

• W2897982562 abstract "Alzheimer's disease (AD) has a long preclinical phase, which provides opportunities for secondary prevention of dementia. To effectively design trials in non-demented subjects, there is a need for identifying subjects at risk of cognitive decline. This study aimed to predict cognitive decline in non-demented subjects based on MRI, clinical, and baseline cognitive data using machine-learning techniques. Individuals were selected from the EMIF-AD Multimodal Biomarker Discovery study, which pooled data and scans from 11 European prospective cohort studies. We included 389 subjects with normal cognition (CN, n = 92) or mild cognitive impairment (MCI, n = 297), a MRI scan and at least 1 follow-up. Cognition was assessed by the MMSE. Potential predictors of cognitive decline included demographics, cognitive performance, and APOE ε4 genotype. Imaging predictors included the results of visual ratings, cortical thickness, subcortical volumes and surface area from FreeSurfer. Outcomes were at least 1 point decline on the MMSE or progression from control to MCI or from MCI to dementia. We first examined univariate predictors of decline using linear mixed effects models, after correction for age, gender, education and site. A support vector regression (SVR) model with leave-one-out cross-validation was applied to select the best features predicting MMSE scores at four follow-up stages. Nested random forests were used for feature selection. After on average 2.2 +/- 1.1 years follow-up, 78 (20%) participants showed clinical progression and 310 (80%) remained stable [Table 1]. In univariate analysis, atrophy measures showed the best association with baseline and longitudinal MMSE measures [Table 2]. Our algorithm selected 38 out of 199 features to predict MMSE values at follow up. Correlation between real and predicted MMSE scores ranged between 0.92-0.96 across time points. The prediction accuracy was high and similar in the stable (mean absolute error: 0.93-1.24) and progressive groups (mean absolute error: 0.43-1.71) [Figure 1]. SVR prediction of MMSE at follow-up Fig 1 shows the SVR performances to predict cognitive decline. Panel A represents the subject composition for the different time points considered (i.e.: M12, M24. M36, K148- M84); Panel B reports the evaluation metrics of the Support Vector Machine (SVM.svr); Panel C shows the rank of the most important biomarkers after feature selection step. Acronyms R2-score: coefficient of determination; MAE: Mean absolute error. MApE: Mean absolute percentage error. Our SVR model could predict, with high confidence, future MMSE scores based on baseline MRI and cognitive measures. The proposed model could be aid the selection of at-risk subjects for secondary prevention trials of AD dementia. Validation of these findings on larger cohorts is necessary." @default.
• W2897982562 created "2018-10-26" @default.
• W2897982562 creator A5001521779 @default.
• W2897982562 creator A5002493210 @default.
• W2897982562 creator A5004136948 @default.
• W2897982562 creator A5015869436 @default.
• W2897982562 creator A5016439888 @default.
• W2897982562 creator A5018434240 @default.
• W2897982562 creator A5020936422 @default.
• W2897982562 creator A5022295403 @default.
• W2897982562 creator A5025499867 @default.
• W2897982562 creator A5029875703 @default.
• W2897982562 creator A5033328565 @default.
• W2897982562 creator A5036068671 @default.
• W2897982562 creator A5036259423 @default.
• W2897982562 creator A5039782136 @default.
• W2897982562 creator A5041038663 @default.
• W2897982562 creator A5041715128 @default.
• W2897982562 creator A5045907715 @default.
• W2897982562 creator A5046609202 @default.
• W2897982562 creator A5050002274 @default.
• W2897982562 creator A5051868596 @default.
• W2897982562 creator A5054596281 @default.
• W2897982562 creator A5062777125 @default.
• W2897982562 creator A5068903009 @default.
• W2897982562 creator A5071341444 @default.
• W2897982562 creator A5073183618 @default.
• W2897982562 creator A5076585111 @default.
• W2897982562 creator A5078311665 @default.
• W2897982562 creator A5079246880 @default.
• W2897982562 creator A5079766290 @default.
• W2897982562 creator A5079993632 @default.
• W2897982562 creator A5084571580 @default.
• W2897982562 creator A5086914602 @default.
• W2897982562 date "2018-07-01" @default.
• W2897982562 modified "2023-10-18" @default.
• W2897982562 title "P2‐458: PREDICTING COGNITIVE DECLINE THROUGH STRUCTURAL MRI BIOMARKERS: RESULTS FROM THE EMIF‐AD BIOMARKER DISCOVERY STUDY" @default.
• W2897982562 doi "https://doi.org/10.1016/j.jalz.2018.06.1151" @default.
• W2897982562 hasPublicationYear "2018" @default.
• W2897982562 type Work @default.
• W2897982562 sameAs 2897982562 @default.
• W2897982562 citedByCount "0" @default.
• W2897982562 crossrefType "journal-article" @default.
• W2897982562 hasAuthorship W2897982562A5001521779 @default.
• W2897982562 hasAuthorship W2897982562A5002493210 @default.
• W2897982562 hasAuthorship W2897982562A5004136948 @default.
• W2897982562 hasAuthorship W2897982562A5015869436 @default.
• W2897982562 hasAuthorship W2897982562A5016439888 @default.
• W2897982562 hasAuthorship W2897982562A5018434240 @default.
• W2897982562 hasAuthorship W2897982562A5020936422 @default.
• W2897982562 hasAuthorship W2897982562A5022295403 @default.
• W2897982562 hasAuthorship W2897982562A5025499867 @default.
• W2897982562 hasAuthorship W2897982562A5029875703 @default.
• W2897982562 hasAuthorship W2897982562A5033328565 @default.
• W2897982562 hasAuthorship W2897982562A5036068671 @default.
• W2897982562 hasAuthorship W2897982562A5036259423 @default.
• W2897982562 hasAuthorship W2897982562A5039782136 @default.
• W2897982562 hasAuthorship W2897982562A5041038663 @default.
• W2897982562 hasAuthorship W2897982562A5041715128 @default.
• W2897982562 hasAuthorship W2897982562A5045907715 @default.
• W2897982562 hasAuthorship W2897982562A5046609202 @default.
• W2897982562 hasAuthorship W2897982562A5050002274 @default.
• W2897982562 hasAuthorship W2897982562A5051868596 @default.
• W2897982562 hasAuthorship W2897982562A5054596281 @default.
• W2897982562 hasAuthorship W2897982562A5062777125 @default.
• W2897982562 hasAuthorship W2897982562A5068903009 @default.
• W2897982562 hasAuthorship W2897982562A5071341444 @default.
• W2897982562 hasAuthorship W2897982562A5073183618 @default.
• W2897982562 hasAuthorship W2897982562A5076585111 @default.
• W2897982562 hasAuthorship W2897982562A5078311665 @default.
• W2897982562 hasAuthorship W2897982562A5079246880 @default.
• W2897982562 hasAuthorship W2897982562A5079766290 @default.
• W2897982562 hasAuthorship W2897982562A5079993632 @default.
• W2897982562 hasAuthorship W2897982562A5084571580 @default.
• W2897982562 hasAuthorship W2897982562A5086914602 @default.
• W2897982562 hasConcept C118552586 @default.
• W2897982562 hasConcept C119857082 @default.
• W2897982562 hasConcept C126322002 @default.
• W2897982562 hasConcept C144301174 @default.
• W2897982562 hasConcept C15744967 @default.
• W2897982562 hasConcept C161584116 @default.
• W2897982562 hasConcept C169900460 @default.
• W2897982562 hasConcept C185592680 @default.
• W2897982562 hasConcept C199163554 @default.
• W2897982562 hasConcept C2779134260 @default.
• W2897982562 hasConcept C2779483572 @default.
• W2897982562 hasConcept C2781197716 @default.
• W2897982562 hasConcept C2984863031 @default.
• W2897982562 hasConcept C38180746 @default.
• W2897982562 hasConcept C41008148 @default.
• W2897982562 hasConcept C55493867 @default.
• W2897982562 hasConcept C71924100 @default.
• W2897982562 hasConcept C72563966 @default.
• W2897982562 hasConcept C99508421 @default.
• W2897982562 hasConceptScore W2897982562C118552586 @default.
• W2897982562 hasConceptScore W2897982562C119857082 @default.
• W2897982562 hasConceptScore W2897982562C126322002 @default.
• W2897982562 hasConceptScore W2897982562C144301174 @default.
• W2897982562 hasConceptScore W2897982562C15744967 @default.
• W2897982562 hasConceptScore W2897982562C161584116 @default.

• next