Dataset Information

Machine learning of neuroimaging for assisted diagnosis of cognitive impairment and dementia: A systematic review.

ABSTRACT:

Introduction

Advanced machine learning methods might help to identify dementia risk from neuroimaging, but their accuracy to date is unclear.

Methods

We systematically reviewed the literature, 2006 to late 2016, for machine learning studies differentiating healthy aging from dementia of various types, assessing study quality, and comparing accuracy at different disease boundaries.

Results

Of 111 relevant studies, most assessed Alzheimer's disease versus healthy controls, using AD Neuroimaging Initiative data, support vector machines, and only T1-weighted sequences. Accuracy was highest for differentiating Alzheimer's disease from healthy controls and poor for differentiating healthy controls versus mild cognitive impairment versus Alzheimer's disease or mild cognitive impairment converters versus nonconverters. Accuracy increased using combined data types, but not by data source, sample size, or machine learning method.

Discussion

Machine learning does not differentiate clinically relevant disease categories yet. More diverse data sets, combinations of different types of data, and close clinical integration of machine learning would help to advance the field.

SUBMITTER: Pellegrini E

PROVIDER: S-EPMC6197752 | biostudies-literature | 2018

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Machine learning of neuroimaging for assisted diagnosis of cognitive impairment and dementia: A systematic review.

Pellegrini Enrico E Ballerini Lucia L Hernandez Maria Del C Valdes MDCV Chappell Francesca M FM González-Castro Victor V Anblagan Devasuda D Danso Samuel S Muñoz-Maniega Susana S Job Dominic D Pernet Cyril C Mair Grant G MacGillivray Tom J TJ Trucco Emanuele E Wardlaw Joanna M JM

Alzheimer's & dementia (Amsterdam, Netherlands) 20180811

<h4>Introduction</h4>Advanced machine learning methods might help to identify dementia risk from neuroimaging, but their accuracy to date is unclear.<h4>Methods</h4>We systematically reviewed the literature, 2006 to late 2016, for machine learning studies differentiating healthy aging from dementia of various types, assessing study quality, and comparing accuracy at different disease boundaries.<h4>Results</h4>Of 111 relevant studies, most assessed Alzheimer's disease versus healthy controls, us ...[more]

PMID: 30364671

Similar Datasets

Project description:BackgroundMild cognitive impairment and cognitive impairment, no dementia, are emerging terms that encompass the clinical state between normal cognition and dementia in elderly people. Controversy surrounds their characterization, definition and application in clinical practice. In this article, we provide physicians with practical guidance on the definition, diagnosis and treatment of mild cognitive impairment and cognitive impairment, no dementia, based on recommendations from the Third Canadian Consensus Conference on the Diagnosis and Treatment of Dementia, held in March 2006.MethodsWe developed evidence-based guidelines using systematic literature searches, with specific criteria for study selection and quality assessment, and a clear and transparent decision-making process. We selected studies published from January 1996 to December 2005 that had mild cognitive impairment or cognitive impairment, no dementia, as the outcome. Subsequent to the conference, we searched for additional articles published between January 2006 and January 2008. We graded the strength of evidence using the criteria of the Canadian Task Force on Preventive Health Care.ResultsWe identified 2483 articles, of which 314 were considered to be relevant and of good or fair quality. From a synthesis of the evidence in these studies, we made 16 recommendations. In brief, family physicians should be aware that most types of dementia are preceded by a recognizable phase of mild cognitive decline. They should be familiar with the concepts of mild cognitive impairment and of cognitive impairment, no dementia. Patients with these conditions should be closely monitored because of their increased risk for dementia. Leisure activities, cognitive stimulation and physical activity could be promoted as part of a healthy lifestyle in elderly people and those with mild cognitive impairment. Vascular risk factors should be treated optimally. No other specific therapies can yet be recommended.InterpretationPhysicians will increasingly see elderly patients with mild memory loss, and learning an approach to diagnosing states such as mild cognitive impairment is now warranted. Close monitoring for progression to dementia, promotion of a healthy lifestyle and treatment of vascular risk factors are recommended for the management of patients with mild cognitive impairment.

Project description:INTRODUCTION:Persisting neurotropic viruses are proposed to increase the risk of dementia, but evidence of association from robust, adequately powered population studies is lacking. This is essential to inform clinical trials of targeted preventive interventions. METHODS AND ANALYSIS:We will carry out a comprehensive systematic review of published and grey literature of the association between infection with, reactivation of, vaccination against or treatment of any of the eight human herpesviruses and dementia or mild cognitive impairment. We will search the Cochrane Library, Embase, Global Health, Medline, PsycINFO, Scopus, Web of Science, clinical trials registers, the New York Academy of Medicine Grey Literature Report, Electronic Theses Online Service through the British Library and the ISI Conference Proceedings Citation Index for randomised controlled trials, cohort, caseâ€"control, case crossover or self-controlled case series studies reported in any language up to January 2017. Titles, abstracts and full-text screening will be conducted by two researchers independently. Data will be extracted systematically from eligible studies using a piloted template. We will assess risk of bias of individual studies in line with the Cochrane Collaboration tool. We will conduct a narrative synthesis, grouping studies by exposure and outcome definitions, and will describe any differences by population subgroups and dementia subtypes. We will consider performing meta-analyses if there are adequate numbers of sufficiently homogeneous studies. The overall quality of cumulative evidence will be assessed using selected Grading of Recommendations, Assessment, Development and Evaluations criteria. ETHICS AND DISSEMINATION:As this is a review of existing studies, no ethical approval is required. Results will be disseminated through a peer-reviewed publication and at national and international conferences. We anticipate the review will clarify the current extent and quality of evidence for a link between herpesviruses and dementia, identify gaps and inform the direction of future research. PROSPERO REGISTRATION NUMBER:CRD42017054684.

Project description:Accumulating evidence suggests an association between coronary heart disease and risk for cognitive impairment or dementia, but no study has systematically reviewed this association. Therefore, we summarized the available evidence on the association between coronary heart disease and risk for cognitive impairment or dementia.Medline, Embase, PsycINFO, and CINAHL were searched for all publications until 8th January 2016. Articles were included if they fulfilled the inclusion criteria: (1) myocardial infarction, angina pectoris or coronary heart disease (combination of both) as predictor variable; (2) cognition, cognitive impairment or dementia as outcome; (3) population-based study; (4) prospective (?1 year follow-up), cross-sectional or case-control study design; (5) ?100 participants; and (6) aged ?45 years. Reference lists of publications and secondary literature were hand-searched for possible missing articles. Two reviewers independently screened all abstracts and extracted information from potential relevant full-text articles using a standardized data collection form. Study quality was assessed with the Newcastle-Ottawa Scale. We pooled estimates from the most fully adjusted model using random-effects meta-analysis.We identified 6,132 abstracts, of which 24 studies were included. A meta-analysis of 10 prospective cohort studies showed that coronary heart disease was associated with increased risk of cognitive impairment or dementia (OR = 1.45, 95%CI = 1.21-1.74, p<0.001). Between-study heterogeneity was low (I2 = 25.7%, 95%CI = 0-64, p = 0.207). Similar significant associations were found in separate meta-analyses of prospective cohort studies for the individual predictors (myocardial infarction, angina pectoris). In contrast, meta-analyses of cross-sectional and case-control studies were inconclusive.This meta-analysis suggests that coronary heart disease is prospectively associated with increased odds of developing cognitive impairment or dementia. Given the projected worldwide increase in the number of people affected by coronary heart disease and dementia, insight into causal mechanisms or common pathways underlying the heart-brain connection is needed.

Dataset Information

Machine learning of neuroimaging for assisted diagnosis of cognitive impairment and dementia: A systematic review.

Introduction

Methods

Results

Discussion

Publications

Machine learning of neuroimaging for assisted diagnosis of cognitive impairment and dementia: A systematic review.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure