Classification of Alzheimer’s Disease with and without Imagery Using Gradient Boosted Machines and ResNet-50

Background. Alzheimer’s is a disease for which there is no cure. Diagnosing Alzheimer’s disease (AD) early facilitates family planning and cost control. The purpose of this study is to predict the presence of AD using socio-demographic, clinical, and magnetic resonance imaging (MRI) data. Early detection of AD enables family planning and may reduce costs by delaying long-term care. Accurate, non-imagery methods also reduce patient costs. The Open Access Series of Imaging Studies (OASIS-1) cross-sectional MRI data were analyzed. A gradient boosted machine (GBM) predicted the presence of AD as a function of gender, age, education, socioeconomic status (SES), and a mini-mental state exam (MMSE). A residual network with 50 layers (ResNet-50) predicted the clinical dementia rating (CDR) presence and severity from MRI’s (multi-class classification). The GBM achieved a mean 91.3% prediction accuracy (10-fold stratified cross validation) for dichotomous CDR using socio-demographic and MMSE variables. MMSE was the most important feature. ResNet-50 using image generation techniques based on an 80% training set resulted in 98.99% three class prediction accuracy on 4139 images (20% validation set) at Epoch 133 and nearly perfect multi-class predication accuracy on the training set (99.34%). Machine learning methods classify AD with high accuracy. GBM models may help provide initial detection based on non-imagery analysis, while ResNet-50 network models might help identify AD patients automatically prior to provider review.

Download Full-text

Classification of Alzheimer's Disease with and without Imagery Using Gradient Boosted Machines and ResNet-50

10.20944/preprints201907.0345.v1 ◽

2019 ◽

Author(s):

Lawrence V. Fulton ◽

Diane Dolezel ◽

Jordan Harrop ◽

Yan Yan ◽

Christopher P. Fulton

Keyword(s):

Alzheimer’S Disease ◽

Alzheimer's Disease ◽

Family Planning ◽

Prediction Accuracy ◽

Network Models ◽

Training Set ◽

Cross Sectional ◽

Term Care ◽

Class Prediction ◽

Multi Class Classification

Alzheimer's is a disease for which there is no cure. Diagnosing Alzheimer's Disease (AD) early facilitates family planning and cost control. The purpose of this study is to predict the presence of AD using socio-demographic, clinical, and Magnetic Resonance Imaging (MRI) data. Early detection of AD enables family planning and may reduce costs by delaying long-term care. Accurate, non-imagery methods also reduce patient costs. The Open Access Series of Imaging Studies (OASIS-1) cross-sectional MRI data were analyzed. A gradient boosted machine (GBM) predicted the presence of AD as a function of gender, age, education, socioeconomic status (SES), and Mini-Mental State Exam (MMSE). A Residual Network with 50 layers (ResNet-50) predicted CDR presence and severity from MRI's (multi-class classification). The GBM achieved a mean 91.3% prediction accuracy (10-fold stratified cross validation) for dichotomous CDR using socio-demographic and MMSE variables. MMSE was the most important feature. ResNet-50 using image generation techniques based on an 80% training set resulted in 98.99% three class prediction accuracy on 4,139 images (20% validation set) at Epoch 133 and nearly perfect multi-class predication accuracy on the training set (99.34%). Machine Learning methods classify AD with high accuracy. GBM models may help provide initial detection based on non-imagery analysis, while ResNet-50 network models might help identify AD patients automatically prior to provider review.

Download Full-text

Neuropsychiatric symptoms and severity of dementia

Dementia & Neuropsychologia ◽

10.1590/s1980-57642013dn70200006 ◽

2013 ◽

Vol 7 (2) ◽

pp. 171-175

Author(s):

Gustavo Henrique de Oliveira Caldas ◽

Sueli Luciano Pires ◽

Milton Luiz Gorzoni

Keyword(s):

Alzheimer’S Disease ◽

Alzheimer's Disease ◽

Motor Behavior ◽

Neuropsychiatric Symptoms ◽

Prevalence Ratio ◽

Cross Sectional Study ◽

Clinical Dementia Rating ◽

Cross Sectional ◽

Specific Stage ◽

Severity Of Dementia

ABSTRACT Neuropsychiatric symptoms (NPS) cause distress, disabilitiy and reduced quality of life for both the patient and their families Objective: To evaluate the prevalence of NPS as a specific stage of dementia status. Methods: A cross-sectional study in patients attending an outpatient clinic for dementia was performed. We applied the Neuropsychiatric Inventory and Clinical Dementia Rating (CDR) scale. Statistical analysis was carried out with SPSS 17 software. Results: The 124 subjects (mean age of 80.4±7.0 years), 88 women (70.9%) had average duration of dementia of 7.1±3.2 years, most common dementias of Alzheimer's disease (35.5%) and mixed (31.5%) and most prevalent NPS of apathy (75%) and irritability (66.9%). Correlation between apathy and a CDR 1 had a PR (prevalence ratio) = 0.289 and p<0.001 while between apathy and CDR 4-5 (PR=8.333, p<0.005). A similar result was found between aberrant motor behavior (AMB) and CDR 1 (PR=0.352, p<0.003) and between AMB and CDR4-5 (PR=2.929, p<0.006). Conclusion: Alzheimer's disease and mixed dementia were predominant, while apathy and AMB were detected in association with the progressive stages of dementia.

Download Full-text

Enablers And Barriers To Utilisation Of Available Formal Support By Institutionalised Individuals With Alzheimer’s Disease And Related Dementias In Rural South Western Uganda: A Qualitative Descriptive Study

10.21203/rs.2.15396/v1 ◽

2019 ◽

Author(s):

VALLENCE NIYONZIMA ◽

Samuel Maling ◽

Zeina Chemali

Keyword(s):

Alzheimer’S Disease ◽

Alzheimer's Disease ◽

Rural Communities ◽

Support Services ◽

The Elderly ◽

Cross Sectional Study ◽

Cross Sectional ◽

Term Care ◽

Formal Support ◽

Western Uganda

Abstract Background: The world’s population is ageing resulting in a larger number of people living with dementia. Prevalence of dementia for persons aged above 60 years is expected to increase . There is paucity of information regarding formal support available for institutionalised individuals with AD/ADRD, enablers and barriers to its utilisation particularly South Western Uganda. Therefore this study, filling this knowledge gap, sought to assess formal support available for individuals with AD/ADRD, enablers and barriers to its utilisation in rural communities in South Western Uganda Methods: This study was conducted in two nongovernmental community based facilities for the elderly in South Western Uganda employing a qualitative cross sectional study design. Purposive sampling was used to recruit caretakers of individuals with Alzheimer’s disease and related dementias. A total of 28 in depth interviews and 6 key informant interviews were conducted. Results: Three arching themes emerged: 1) Formal support services available, 2) Enablers and 3) Barriers to the utilisation of the formal support services in institutionalised patients with AD/ADRD in rural communities in South Western Uganda. Conclusions: These findings highlight the importance of formal care in the long-term care of individuals with AD/ADRD. There is a dire need to equip caretakers involved in the care of patients with AD/ADRD with the required skill set and include formal support services in the mainstream primary health care to make the services accessible.

Download Full-text

Inter-Cohort Validation of SuStaIn Model for Alzheimer’s Disease

Frontiers in Big Data ◽

10.3389/fdata.2021.661110 ◽

2021 ◽

Vol 4 ◽

Author(s):

Damiano Archetti ◽

Alexandra L. Young ◽

Neil P. Oxtoby ◽

Daniel Ferreira ◽

Gustav Mårtensson ◽

...

Keyword(s):

Alzheimer’S Disease ◽

Alzheimer's Disease ◽

Clinical Setting ◽

Neurodegenerative Disorder ◽

Amyloid Β ◽

Research Data ◽

Training Set ◽

Cross Sectional ◽

Data Set ◽

Test Sets

Alzheimer’s disease (AD) is a neurodegenerative disorder which spans several years from preclinical manifestations to dementia. In recent years, interest in the application of machine learning (ML) algorithms to personalized medicine has grown considerably, and a major challenge that such models face is the transferability from the research settings to clinical practice. The objective of this work was to demonstrate the transferability of the Subtype and Stage Inference (SuStaIn) model from well-characterized research data set, employed as training set, to independent less-structured and heterogeneous test sets representative of the clinical setting. The training set was composed of MRI data of 1043 subjects from the Alzheimer’s disease Neuroimaging Initiative (ADNI), and the test set was composed of data from 767 subjects from OASIS, Pharma-Cog, and ViTA clinical datasets. Both sets included subjects covering the entire spectrum of AD, and for both sets volumes of relevant brain regions were derived from T1-3D MRI scans processed with Freesurfer v5.3 cross-sectional stream. In order to assess the predictive value of the model, subpopulations of subjects with stable mild cognitive impairment (MCI) and MCIs that progressed to AD dementia (pMCI) were identified in both sets. SuStaIn identified three disease subtypes, of which the most prevalent corresponded to the typical atrophy pattern of AD. The other SuStaIn subtypes exhibited similarities with the previously defined hippocampal sparing and limbic predominant atrophy patterns of AD. Subject subtyping proved to be consistent in time for all cohorts and the staging provided by the model was correlated with cognitive performance. Classification of subjects on the basis of a combination of SuStaIn subtype and stage, mini mental state examination and amyloid-β1-42 cerebrospinal fluid concentration was proven to predict conversion from MCI to AD dementia on par with other novel statistical algorithms, with ROC curves that were not statistically different for the training and test sets and with area under curve respectively equal to 0.77 and 0.76. This study proves the transferability of a SuStaIn model for AD from research data to less-structured clinical cohorts, and indicates transferability to the clinical setting.

Download Full-text

SERUM 25(OH)D SIGNIFICANTLY IMPACTS ALZHEIMER’S DISEASE IN OLDER ADULTS LIVING IN LONG-TERM CARE COMMUNITIES.

Innovation in Aging ◽

10.1093/geroni/igz038.3170 ◽

2019 ◽

Vol 3 (Supplement_1) ◽

pp. S863-S863

Author(s):

Robert A Churchill ◽

Ronna Robbins ◽

Nalini Ranjit ◽

Sara Sweitzer ◽

Margaret Briley

Keyword(s):

Alzheimer’S Disease ◽

Older Adults ◽

Alzheimer's Disease ◽

Serum Levels ◽

Cross Sectional Study ◽

Vitamin D Supplementation ◽

Cross Sectional ◽

Term Care ◽

Logistic Regression Models ◽

Hydroxyvitamin D

Abstract 6.08 million Americans suffer from Alzheimer’s Disease (AD), with some estimating diagnoses will reach 15.0 million by 2060. Age is the strongest risk factor for AD, and the prevalence of AD among older adults necessitates investigation into preventable risk factors. 25-hydroxyvitamin D [25(OH)D] deficiency is more prevalent in older adults than other age demographics. Research shows a strong correlation between deficient 25(OH)D serum levels (≤20ng/ml) and AD diagnosis. The association with insufficient (≤30 ng/ml) levels remains unclear. Older adults (age > 65 yo) of five LTC communities in Texas participated in the cross-sectional study. One-year medical history and demographics were abstracted from medical records using double-blinded data abstraction and entry. Blood draws measured 25(OH)D serum levels. Adjusted logistic regression models examined if insufficient 25(OH)D serum levels (≤30 ng/ml) are associated with AD diagnosis. Confounders were total daily vitamin D supplementation, BMI, race, gender, age, years in community, and diagnosis of liver and renal disease. Participants (n=174, mean age: 83 yo) consisted of 63% female and 89% Caucasian. Fifty five percent had insufficient serum 25(OH)D levels (mean level: 32.6 ng/ml; mean supplementation rate: 1,138 IU per/d), and 20% had diagnosis of AD. 25% had both insufficient serum levels and AD, while 12.6% had adequate serum levels and AD. Those with insufficient 25(OH)D serum levels had elevated odds (OR=2.8; CL: 1.14, 7.02; p=0.024) of having AD after adjusting for confounders. Insufficient serum 25(OH)D levels (≤30 ng/ml) are associated with increased diagnoses of AD, indicating the importance of adequate levels among LTC residents.

Download Full-text

Relationship of frontal lobe dysfunction and aberrant motor behaviors in patients with Alzheimer's disease

International Psychogeriatrics ◽

10.1017/s1041610209991323 ◽

2009 ◽

Vol 22 (3) ◽

pp. 463-469 ◽

Cited By ~ 16

Author(s):

Tomoyuki Nagata ◽

Shunichiro Shinagawa ◽

Yusuke Ochiai ◽

Hirohide Kada ◽

Hiroo Kasahara ◽

...

Keyword(s):

Alzheimer’S Disease ◽

Alzheimer's Disease ◽

Frontal Lobe ◽

Motor Behavior ◽

Cross Sectional Study ◽

Stepwise Multiple Regression ◽

Clinical Dementia Rating ◽

Cross Sectional ◽

Motor Behaviors ◽

Relationship Of

ABSTRACTBackground: In order to address the neuropsychological pathogenesis of aberrant motor behaviors in Alzheimer's disease (AD), we used a cross-sectional study design to investigate the association between frontal lobe function, including executive function, and activity disturbances (wandering, purposeless activities and inappropriate activities).Methods: Among 75 consecutive outpatients with AD, 50 subjects with a Clinical Dementia Rating (CDR) score of 1 or 2 were selected and divided into two groups based on data obtained from interviews with their caregivers: an aberrant motor behaviors (AMB) group (n = 22), and a non-aberrant motor behaviors (NAMB) group (n = 28). Aberrant motor behavior was defined according to whether the “activity disturbance” score (ranging from 0 to 9) of the Behavioral Pathology in Alzheimer Disease (Behave-AD) scale was 0 or ≥1. The total and subtest scores of the Frontal Assessment Battery (FAB) were then compared between the two groups.Results: Significant differences were found between the FAB total (P < 0.05) and the subtest scores (lexical fluency, conflicting instructions; P < 0.05) in the two groups. The FAB score was significantly associated with the activity disturbance score (r = −0.49; P<0.001). A stepwise multiple regression analysis showed that only the FAB score significantly influenced the activity disturbance score (P < 0.001).Conclusions: This finding suggested that in addition to episodic memory disturbance, frontal lobe dysfunctions might lead patients with AD to develop aberrant motor behavior.

Download Full-text

Diagnostic Value of Subjective Memory Complaints Assessed with a Single Item in Dominantly Inherited Alzheimer’s Disease: Results of the DIAN Study

BioMed Research International ◽

10.1155/2015/828120 ◽

2015 ◽

Vol 2015 ◽

pp. 1-7 ◽

Cited By ~ 3

Author(s):

Christoph Laske ◽

Hamid R. Sohrabi ◽

Mateusz S. Jasielec ◽

Stephan Müller ◽

Niklas K. Koehler ◽

...

Keyword(s):

Alzheimer’S Disease ◽

Alzheimer's Disease ◽

Depression Scale ◽

Diagnostic Value ◽

Subjective Memory Complaints ◽

Subjective Memory ◽

Clinical Dementia Rating ◽

Cross Sectional ◽

Single Item ◽

Memory Complaints

Objective. We examined the diagnostic value of subjective memory complaints (SMCs) assessed with a single item in a large cross-sectional cohort consisting of families with autosomal dominant Alzheimer’s disease (ADAD) participating in the Dominantly Inherited Alzheimer Network (DIAN).Methods. The baseline sample of 183 mutation carriers (MCs) and 117 noncarriers (NCs) was divided according to Clinical Dementia Rating (CDR) scale into preclinical (CDR 0; MCs:n=107; NCs:n=109), early symptomatic (CDR 0.5; MCs:n=48; NCs:n=8), and dementia stage (CDR ≥ 1; MCs:n=28; NCs:n=0). These groups were subdivided by the presence or absence of SMCs.Results. At CDR 0, SMCs were present in 12.1% of MCs and 9.2% of NCs(P=0.6). At CDR 0.5, SMCs were present in 66.7% of MCs and 62.5% of NCs(P=1.0). At CDR ≥ 1, SMCs were present in 96.4% of MCs. SMCs in MCs were significantly associated with CDR, logical memory scores, Geriatric Depression Scale, education, and estimated years to onset.Conclusions. The present study shows that SMCs assessed by a single-item scale have no diagnostic value to identify preclinical ADAD in asymptomatic individuals. These results demonstrate the need of further improvement of SMC measures that should be examined in large clinical trials.

Download Full-text

Factors associated with caregiver burden: comparative study between Brazilian and Spanish caregivers of patients with Alzheimer's disease (AD)

International Psychogeriatrics ◽

10.1017/s1041610216000508 ◽

2016 ◽

Vol 28 (8) ◽

pp. 1363-1374 ◽

Cited By ~ 19

Author(s):

Maria Fernanda B. Sousa ◽

Raquel L. Santos ◽

Oriol Turró-Garriga ◽

Rachel Dias ◽

Marcia C. N. Dourado ◽

...

Keyword(s):

Alzheimer’S Disease ◽

Alzheimer's Disease ◽

Day Care ◽

Neuropsychiatric Symptoms ◽

Care Center ◽

Clinical Dementia Rating ◽

Cross Sectional ◽

Day Care Center ◽

High Burden ◽

Factors Associated

ABSTRACTBackground:Transcultural studies regarding the comparison of levels of burden in caregivers of patients with Alzheimer's disease (AD) from Europe and Latin America are rare. We designed this study to investigate the differentiating factors associated with burden in Brazilian and Spanish caregivers of patients with AD.Methods:This is a cross-sectional study composed by samples of outpatients with AD and their caregivers from Brazil (n = 128) and Spain (n = 146). Caregivers answered the Zarit Burden Interview (ZBI) and a Sociodemographic Questionnaire. Patients were assessed with the Mini-Mental State Examination (MMSE), Functional Activities Questionnaire (FAQ), Disability Assessment for Dementia (DAD), Neuropsychiatric Inventory (NPI), and Clinical Dementia Rating (CDR) Scale.Results:In the multivariate regression analysis, high burden levels were reported in Brazil, when caregivers were female (p = 0.025) and when patients did not attend Day Care Center (p = 0.025). In Spain, high burden levels were associated with living with the patient (p = 0.014), younger caregivers (p = 0.003), and participation of patients at Day Care Center (p = 0.046). Also, different neuropsychiatric symptoms explained high burden levels: in Brazil, depression (p < 0.001) and anxiety (p = 0.024) and, in Spain, apathy/indifference (p < 0.001), agitation/aggression (p = 0.019) and irritability/lability (p = 0.027).Conclusions:Caregivers’ gender, patients who attended Day Care Center and neuropsychiatric symptoms were differentiating factors in the burden of Brazilian and Spanish caregivers.

Download Full-text

Clinical progression of moderate-to-severe Alzheimer's disease and caregiver burden: a 12-month multicenter prospective observational study

International Psychogeriatrics ◽

10.1017/s104161021000150x ◽

2010 ◽

Vol 22 (8) ◽

pp. 1265-1279 ◽

Cited By ~ 27

Author(s):

Luis Agüera-Ortiz ◽

Ana Frank-García ◽

Pedro Gil ◽

Alfonso Moreno

Keyword(s):

Alzheimer’S Disease ◽

Alzheimer's Disease ◽

Observational Study ◽

Caregiver Burden ◽

Major Influence ◽

Functional Impairments ◽

Clinical Progression ◽

Clinical Dementia Rating ◽

Cross Sectional ◽

Use Of Resources

ABSTRACTBackground: Prospective studies on the clinical progression of Alzheimer's disease (AD) and its relationship to caregiver burden are needed to improve illness management and use of resources.Methods: This national, multicenter, observational study evaluated 1235 moderate to severe AD patients under routine care in Spain. Baseline cross-sectional sociodemographic and clinical data, and changes from baseline to month 12 of various neuropsychological tests and clinical ratings, including Blessed Dementia Scale, Mini-mental State Examination (MMSE), Hughes Clinical Dementia Rating sum-of-boxes (CDR-SB), Clinical Global Impression of Change (CGIC) and Zarit Caregiver Burden scales, were recorded and comprehensively analyzed.Results: Baseline data were in accordance with characteristics consistently reported to influence AD risk regarding anthropometrics, sociocultural features and comorbidities. Significant progressive functional impairments (i.e. in routine activities and essential daily tasks) and cognitive (i.e. MMSE and CDR-SB) impairments were found at month 12. However, patients' behavior and caregivers' burden improved slightly, but significantly, corroborating the major influence of behavioral symptoms on caregivers' distress. Caregivers showed significantly lower burden with patients with higher levels of education and, to a lesser extent, when patients received AD-specific medication. Physicians accurately detected AD clinical evolution as their CGIC ratings significantly correlated with all tests.Conclusions: These findings reinforce previous AD knowledge and add data on the clinical course of advanced stages of AD. Caregiver burden depended more on patients' behavioral alterations than on their functional or cognitive declines; and it was diminished by their patients having higher levels of education and being treated with AD-specific medications. Research into unexplored factors that might reduce caregiver burden, ultimately benefiting both patients and caregivers, is encouraged.

Download Full-text

Assessing Dependency in a Multiethnic Community Cohort of Individuals With Alzheimer’s Disease

Innovation in Aging ◽

10.1093/geroni/igy011 ◽

2018 ◽

Vol 2 (1) ◽

Cited By ~ 2

Author(s):

Maria Meinerding ◽

Brittany DeFeis ◽

Preeti Sunderaraman ◽

Martina Azar ◽

Siobhan Lawless ◽

...

Keyword(s):

Alzheimer’S Disease ◽

Alzheimer's Disease ◽

Disease Severity ◽

Living Arrangement ◽

Institutional Care ◽

Clinical Dementia Rating ◽

Cross Sectional ◽

Scale Scores ◽

Dependence Scale ◽

Design And Methods

Abstract Background and Objectives Clinic-based studies of patients with Alzheimer’s disease (AD) have demonstrated the value of assessing dependence when characterizing patients’ functional status. The Dependence Scale, a validated tool to assess level of caregiving needs, is associated with markers of disease severity, cost, and progression, while offering independent functional information about patients. This study examines whether such associations between the Dependence Scale and markers of disease severity demonstrated in clinical cohorts are similarly exhibited in a multiethnic community population of individuals with AD. Research Design and Methods One hundred fifty four elders with AD enrolled in the Predictors 3 cohort were assessed with the Dependence Scale, modified Mini-Mental State Examination (mMMS), instrumental (IADL) and basic (BADL) activities of daily living, and Clinical Dementia Rating (CDR) Scale, and were assigned an Equivalent Institutional Care (EIC) rating. Cross-sectional associations were examined using bivariate correlations and one-way analysis of variance analyses. Fisher-z tests examined differences in strengths of associations across previous clinic and current community cohorts. Results Dependence Scale scores were associated with CDR (r = .20, p = .013), mMMS (r = −.23, p = .005), IADL (r = .39, p < .001), BADL (r = .65, p < .001), and EIC (r = .51, p < .001). Dependence was unassociated with ethnicity (F[3,144] = 1.027, p = .3822), age (r = .120, p = .145), and education (r = −.053, p = .519). The strength of the correlations was comparable across cohorts except that BADLs were more strongly associated with dependence (z = −4.60, p < .001) in the community cohort, and living arrangement was not associated with dependence (r = .13, p = .130). Discussion and Implications Associations between the Dependence Scale and markers of disease severity in a clinic-based cohort of AD patients are similar to associations in a multiethnic community cohort of individuals diagnosed with AD. The Dependence Scale relates to markers of disease severity rather than demographic factors, and may offer an unbiased assessment of care required in multiethnic and community populations.

Download Full-text