RESUMO
BACKGROUND: Cerebral amyloid angiopathy (CAA) is associated with white matter damage and neurodegeneration. Gait is impaired in CAA; however, the neural basis of this impairment is unclear. RESEARCH QUESTION: Are gait impairments in patients with CAA associated with altered cerebral white matter diffusivity and/or atrophy of cortical and subcortical grey matter. METHODS: Participants with CAA (n=29), Alzheimer's disease (AD; n=16), and normal controls (n=47) were included. Gait was assessed using a 6â¯m walkway with parameters categorized into rhythm, pace, postural control, and variability domains. The dual-task cost (DTC) of gait speed was calculated for counting backwards, animal fluency, and serial sevens tasks. Whole-brain white matter disruption was quantified using the peak width of skeletonized mean diffusivity (PSMD), and thickness and volume of select cortical, subcortical, and cerebellar regions were quantified using FreeSurfer. RESULTS: In CAA participants, associations were found between PSMD and pace (standardized parameter estimate (ß), 95â¯% confidence interval (CI): 0.17, 0.03-0.32), and medial orbital frontal cortical thickness and counting backwards DTC (parameter estimate (PE), 95â¯% CI: -5.7â¯%/SD, -0.24 to -11.23). Across all groups, including CAA, associations were found between PSMD and pace, variability, counting backwards DTC, and animal fluency DTC; between frontal cortical thickness and pace, counting backwards DTC, and animal fluency DTC; between cortical regions affected by AD (inferior parietal cortex, inferior and middle temporal gyrus) and counting backwards DTC; and between thalamus volume and postural control. SIGNIFICANCE: Reduced white matter structural integrity and grey matter loss is associated with poor overall gait performance in CAA, AD, and normal controls.
RESUMO
Mild cognitive impairment (MCI) and gait deficits are commonly associated with Parkinson's disease (PD). Early detection of MCI associated with Parkinson's disease (PD-MCI) and its biomarkers is critical to managing disability in PD patients, reducing caregiver burden and healthcare costs. Gait is considered a surrogate marker for cognitive decline in PD. However, gait kinematic and kinetic features in PD-MCI patients remain unknown. This study was designed to explore the difference in gait kinematics and kinetics during single-task and dual-task walking between PD patients with and without MCI. Kinematic and kinetic data of 90 PD patients were collected using 3D motion capture system. Differences in gait kinematic and kinetic gait features between groups were identified by using: first, univariate statistical analysis and then a supervised machine learning analysis. The findings of this study showed that the presence of MCI in PD patients is coupled with kinematic and kinetic deviations of gait cycle which may eventually identify two different phenotypes of the disease. Indeed, as shown by the demographical and clinical comparison between the two groups, PD-MCI patients were older and more impaired. Moreover, PD-MCI kinematic results showed that cognitive dysfunction coexists with more severe axial symptoms and an increase postural flexion. A lack of physiological distal-to-proximal shift in joint kinetics was evidenced in the PD phenotype associated with cognitive impairments.
Assuntos
Disfunção Cognitiva , Transtornos Neurológicos da Marcha , Marcha , Doença de Parkinson , Humanos , Doença de Parkinson/complicações , Doença de Parkinson/fisiopatologia , Disfunção Cognitiva/fisiopatologia , Disfunção Cognitiva/diagnóstico , Masculino , Feminino , Fenômenos Biomecânicos , Idoso , Pessoa de Meia-Idade , Transtornos Neurológicos da Marcha/fisiopatologia , Transtornos Neurológicos da Marcha/etiologia , Marcha/fisiologia , Cinética , Aprendizado de Máquina , Caminhada/fisiologia , AlgoritmosRESUMO
Background: Alzheimer's disease (AD) and Lewy body disease (LBD) are characterized by early and gradual worsening perturbations in speeded cognitive responses. Objective: Using simple and choice reaction time tasks, we compared two indicators of cognitive speed within and across the AD and LBD spectra: mean rate (average reaction time across trials) and inconsistency (within person variability). Methods: The AD spectrum cohorts included subjective cognitive impairment (SCI, nâ=â28), mild cognitive impairment (MCI, nâ=â121), and AD (nâ=â45) participants. The LBD spectrum included Parkinson's disease (PD, nâ=â32), mild cognitive impairment in PD (PD-MCI, nâ=â21), and LBD (nâ=â18) participants. A cognitively unimpaired (CU, nâ=â39) cohort served as common benchmark. We conducted multivariate analyses of variance and discrimination analyses. Results: Within the AD spectrum, the AD cohort was slower and more inconsistent than the CU, SCI, and MCI cohorts. The MCI cohort was slower than the CU cohort. Within the LBD spectrum, the LBD cohort was slower and more inconsistent than the CU, PD, and PD-MCI cohorts. The PD-MCI cohort was slower than the CU and PD cohorts. In cross-spectra (corresponding cohort) comparisons, the LBD cohort was slower and more inconsistent than the AD cohort. The PD-MCI cohort was slower than the MCI cohort. Discrimination analyses clarified the group difference patterns. Conclusions: For both speed tasks, mean rate and inconsistency demonstrated similar sensitivity to spectra-related comparisons. Both dementia cohorts were slower and more inconsistent than each of their respective non-dementia cohorts.
Assuntos
Doença de Alzheimer , Disfunção Cognitiva , Doença por Corpos de Lewy , Testes Neuropsicológicos , Tempo de Reação , Humanos , Feminino , Masculino , Idoso , Doença por Corpos de Lewy/psicologia , Doença de Alzheimer/psicologia , Disfunção Cognitiva/psicologia , Tempo de Reação/fisiologia , Testes Neuropsicológicos/estatística & dados numéricos , Estudos de Coortes , Doença de Parkinson/psicologia , Idoso de 80 Anos ou mais , Cognição/fisiologia , Pessoa de Meia-IdadeRESUMO
Cerebral amyloid angiopathy (CAA) is characterized by the accumulation of amyloid protein in the walls of cerebral blood vessels. This deposition of amyloid causes damage to the cerebral vasculature, resulting in blood-brain barrier disruption, cerebral hemorrhage, cognitive decline, and dementia. The role of the immune system in CAA is complex and not fully understood. While the immune system has a clear role in the rare inflammatory variants of CAA (CAA related inflammation and Abeta related angiitis), the more common variants of CAA also have immune system involvement. In a protective role, immune cells may facilitate the clearance of beta-amyloid from the cerebral vasculature. The immune system can also contribute to CAA pathology, promoting vascular injury, blood-brain barrier breakdown, inflammation, and progression of CAA. In this review, we summarize the role of the immune system in CAA, including the potential of immune based treatment strategies to slow vascular disease in CAA and associated cognitive impairment, white matter disease progression, and reduce the risk of cerebral hemorrhage. HIGHLIGHTS: The immune system has a role in cerebral amyloid angiopathy (CAA) which is summarized in this review. There is an inflammatory response to beta-amyloid that may contribute to brain injury and cognitive impairment. Immune cells may facilitate the clearance of beta-amyloid from the cerebral vasculature. Improved understanding of the immune system in CAA may afford novel treatment to improve outcomes in patients with CAA.
Assuntos
Peptídeos beta-Amiloides , Angiopatia Amiloide Cerebral , Angiopatia Amiloide Cerebral/patologia , Humanos , Peptídeos beta-Amiloides/metabolismo , Sistema Imunitário , Inflamação/imunologia , Barreira Hematoencefálica , Animais , Encéfalo/patologia , Encéfalo/imunologiaRESUMO
White matter damage quantified as white matter hyperintensities (WMH) may aggravate cognitive and motor impairments, but whether and how WMH burden impacts these problems in Parkinson's disease (PD) is not fully understood. This study aimed to examine the association between WMH and cognitive and motor performance in PD through a systematic review and meta-analysis. We compared the WMH burden across the cognitive spectrum (cognitively normal, mild cognitive impairment, dementia) in PD including controls. Motor signs were compared in PD with low/negative and high/positive WMH burden. We compared baseline WMH burden of PD who did and did not convert to MCI or dementia. MEDLINE and EMBASE databases were used to conduct the literature search resulting in 50 studies included for data extraction. Increased WMH burden was found in individuals with PD compared with individuals without PD (i.e. control) and across the cognitive spectrum in PD (i.e. PD, PD-MCI, PDD). Individuals with PD with high/positive WMH burden had worse global cognition, executive function, and attention. Similarly, PD with high/positive WMH presented worse motor signs compared with individuals presenting low/negative WMH burden. Only three longitudinal studies were retrieved from our search and they showed that PD who converted to MCI or dementia, did not have significantly higher WMH burden at baseline, although no data was provided on WMH burden changes during the follow up. We conclude, based on cross-sectional studies, that WMH burden appears to increase with PD worse cognitive and motor status in PD.
Assuntos
Disfunção Cognitiva , Doença de Parkinson , Substância Branca , Humanos , Doença de Parkinson/complicações , Doença de Parkinson/patologia , Doença de Parkinson/diagnóstico por imagem , Doença de Parkinson/fisiopatologia , Substância Branca/diagnóstico por imagem , Substância Branca/patologia , Disfunção Cognitiva/etiologia , Disfunção Cognitiva/fisiopatologia , Disfunção Cognitiva/patologia , Disfunção Cognitiva/diagnóstico por imagem , Demência/patologia , Demência/etiologia , Demência/fisiopatologiaRESUMO
PET imaging is increasingly recognized as an important diagnostic tool to investigate patients with cognitive disturbances of possible neurodegenerative origin. PET with 2-[18F]fluoro-2-deoxy-D-glucose ([18F]FDG), assessing glucose metabolism, provides a measure of neurodegeneration and allows a precise differential diagnosis among the most common neurodegenerative diseases, such as Alzheimer's disease, frontotemporal dementia or dementia with Lewy bodies. PET tracers specific for the pathological deposits characteristic of different neurodegenerative processes, namely amyloid and tau deposits typical of Alzheimer's Disease, allow the visualization of these aggregates in vivo. [18F]FDG and amyloid PET imaging have reached a high level of clinical validity and are since 2022 investigations that can be offered to patients in standard clinical care in most of Canada.This article will briefly review and summarize the current knowledge on these diagnostic tools, their integration into diagnostic algorithms as well as perspectives for future developments.
RESUMO
T1-weighted magnetization-prepared rapid gradient-echo (MPRAGE) is commonly included in brain studies for structural imaging using magnitude images; however, its phase images can provide an opportunity to assess microbleed burden using quantitative susceptibility mapping (QSM). This potential application for MPRAGE-based QSM was evaluated using in vivo and simulated measurements. Possible factors affecting image quality were also explored. Detection sensitivity was evaluated against standard multiecho gradient echo (MEGE) QSM using 3-T in vivo data of 15 subjects with a combined total of 108 confirmed microbleeds. The two methods were compared based on the microbleed size and susceptibility measurements. In addition, simulations explored the detection sensitivity of MPRAGE-QSM at different representative magnetic field strengths and echo times using microbleeds of different size, susceptibility, and location. Results showed that in vivo microbleeds appeared to be smaller (× 0.54) and of higher mean susceptibility (× 1.9) on MPRAGE-QSM than on MEGE-QSM, but total susceptibility estimates were in closer agreement (slope: 0.97, r2: 0.94), and detection sensitivity was comparable. In simulations, QSM at 1.5 T had a low contrast-to-noise ratio that obscured the detection of many microbleeds. Signal-to-noise ratio (SNR) levels at 3 T and above resulted in better contrast and increased detection. The detection rates for microbleeds of minimum one-voxel diameter and 0.4-ppm susceptibility were 0.55, 0.80, and 0.88 at SNR levels of 1.5, 3, and 7 T, respectively. Size and total susceptibility estimates were more consistent than mean susceptibility estimates, which showed size-dependent underestimation. MPRAGE-QSM provides an opportunity to detect and quantify the size and susceptibility of microbleeds of at least one-voxel diameter at B0 of 3 T or higher with no additional time cost, when standard T2*-weighted images are not available or have inadequate spatial resolution. The total susceptibility measure is more robust against sequence variations and might allow combining data from different protocols.
Assuntos
Hemorragia Cerebral , Imageamento por Ressonância Magnética , Humanos , Hemorragia Cerebral/diagnóstico por imagem , Masculino , Feminino , Pessoa de Meia-Idade , Imageamento por Ressonância Magnética/métodos , Idoso , Simulação por Computador , AdultoRESUMO
Distributed learning is a promising alternative to central learning for machine learning (ML) model training, overcoming data-sharing problems in healthcare. Previous studies exploring federated learning (FL) or the traveling model (TM) setup for medical image-based disease classification often relied on large databases with a limited number of centers or simulated artificial centers, raising doubts about real-world applicability. This study develops and evaluates a convolution neural network (CNN) for Parkinson's disease classification using data acquired by 83 diverse real centers around the world, mostly contributing small training samples. Our approach specifically makes use of the TM setup, which has proven effective in scenarios with limited data availability but has never been used for image-based disease classification. Our findings reveal that TM is effective for training CNN models, even in complex real-world scenarios with variable data distributions. After sufficient training cycles, the TM-trained CNN matches or slightly surpasses the performance of the centrally trained counterpart (AUROC of 83% vs. 80%). Our study highlights, for the first time, the effectiveness of TM in 3D medical image classification, especially in scenarios with limited training samples and heterogeneous distributed data. These insights are relevant for situations where ML models are supposed to be trained using data from small or remote medical centers, and rare diseases with sparse cases. The simplicity of this approach enables a broad application to many deep learning tasks, enhancing its clinical utility across various contexts and medical facilities.
RESUMO
Parkinson's disease (PD) is the second most common neurodegenerative disease. Accurate PD diagnosis is crucial for effective treatment and prognosis but can be challenging, especially at early disease stages. This study aimed to develop and evaluate an explainable deep learning model for PD classification from multimodal neuroimaging data. The model was trained using one of the largest collections of T1-weighted and diffusion-tensor magnetic resonance imaging (MRI) datasets. A total of 1264 datasets from eight different studies were collected, including 611 PD patients and 653 healthy controls (HC). These datasets were pre-processed and non-linearly registered to the MNI PD25 atlas. Six imaging maps describing the macro- and micro-structural integrity of brain tissues complemented with age and sex parameters were used to train a convolutional neural network (CNN) to classify PD/HC subjects. Explainability of the model's decision-making was achieved using SmoothGrad saliency maps, highlighting important brain regions. The CNN was trained using a 75%/10%/15% train/validation/test split stratified by diagnosis, sex, age, and study, achieving a ROC-AUC of 0.89, accuracy of 80.8%, specificity of 82.4%, and sensitivity of 79.1% on the test set. Saliency maps revealed that diffusion tensor imaging data, especially fractional anisotropy, was more important for the classification than T1-weighted data, highlighting subcortical regions such as the brainstem, thalamus, amygdala, hippocampus, and cortical areas. The proposed model, trained on a large multimodal MRI database, can classify PD patients and HC subjects with high accuracy and clinically reasonable explanations, suggesting that micro-structural brain changes play an essential role in the disease course.
RESUMO
Sharing multicenter imaging datasets can be advantageous to increase data diversity and size but may lead to spurious correlations between site-related biological and non-biological image features and target labels, which machine learning (ML) models may exploit as shortcuts. To date, studies analyzing how and if deep learning models may use such effects as a shortcut are scarce. Thus, the aim of this work was to investigate if site-related effects are encoded in the feature space of an established deep learning model designed for Parkinson's disease (PD) classification based on T1-weighted MRI datasets. Therefore, all layers of the PD classifier were frozen, except for the last layer of the network, which was replaced by a linear layer that was exclusively re-trained to predict three potential bias types (biological sex, scanner type, and originating site). Our findings based on a large database consisting of 1880 MRI scans collected across 41 centers show that the feature space of the established PD model (74% accuracy) can be used to classify sex (75% accuracy), scanner type (79% accuracy), and site location (71% accuracy) with high accuracies despite this information never being explicitly provided to the PD model during original training. Overall, the results of this study suggest that trained image-based classifiers may use unwanted shortcuts that are not meaningful for the actual clinical task at hand. This finding may explain why many image-based deep learning models do not perform well when applied to data from centers not contributing to the training set.
Assuntos
Doença de Parkinson , Humanos , Doença de Parkinson/diagnóstico por imagem , Imageamento por Ressonância Magnética/métodos , Aprendizado de Máquina , Máquina de Vetores de SuporteRESUMO
BACKGROUND: Within the spectrum of Lewy body disorders (LBD), both Parkinson's disease (PD) and dementia with Lewy bodies (DLB) are characterized by gait and balance disturbances, which become more prominent under dual-task (DT) conditions. The brain substrates underlying DT gait variations, however, remain poorly understood in LBD. OBJECTIVE: To investigate the relationship between gray matter volume loss and DT gait variations in LBD. METHODS: Seventy-nine participants including cognitively unimpaired PD, PD with mild cognitive impairment, PD with dementia (PDD), or DLB and 20 cognitively unimpaired controls were examined across a multi-site study. PDD and DLB were grouped together for analyses. Differences in gait speed between single and DT conditions were quantified by dual task cost (DTC). Cortical, subcortical, ventricle, and cerebellum brain volumes were obtained using FreeSurfer. Linear regression models were used to examine the relationship between gray matter volumes and DTC. RESULTS: Smaller amygdala and total cortical volumes, and larger ventricle volumes were associated with a higher DTC across LBD and cognitively unimpaired controls. No statistically significant interaction between group and brain volumes were found. Adding cognitive and motor covariates or white matter hyperintensity volumes separately to the models did not affect brain volume and DTC associations. CONCLUSION: Gray matter volume loss is associated with worse DT gait performance compared to single task gait, across cognitively unimpaired controls through and the LBD spectrum. Impairment in DT gait performance may be driven by age-related cortical neurodegeneration.
Assuntos
Doença de Alzheimer , Doença por Corpos de Lewy , Doença de Parkinson , Humanos , Envelhecimento , Doença de Alzheimer/complicações , Marcha , Substância Cinzenta/diagnóstico por imagem , Corpos de Lewy , Doença por Corpos de Lewy/diagnóstico por imagem , Doença por Corpos de Lewy/complicações , Doença de Parkinson/complicaçõesRESUMO
Background: People living with Parkinson's disease (PD) have a high risk for falls. Objective: To examine gaps in falls prevention targeting people with PD as part of the Task Force on Global Guidelines for Falls in Older Adults. Methods: A Delphi consensus process was used to identify specific recommendations for falls in PD. The current narrative review was conducted as educational background with a view to identifying gaps in fall prevention. Results: A recent Cochrane review recommended exercises and structured physical activities for PD; however, the types of exercises and activities to recommend and PD subgroups likely to benefit require further consideration. Freezing of gait, reduced gait speed, and a prior history of falls are risk factors for falls in PD and should be incorporated in assessments to identify fall risk and target interventions. Multimodal and multi-domain fall prevention interventions may be beneficial. With advanced or complex PD, balance and strength training should be administered under supervision. Medications, particularly cholinesterase inhibitors, show promise for falls prevention. Identifying how to engage people with PD, their families, and health professionals in falls education and implementation remains a challenge. Barriers to the prevention of falls occur at individual, environmental, policy, and health system levels. Conclusion: Effective mitigation of fall risk requires specific targeting and strategies to reduce this debilitating and common problem in PD. While exercise is recommended, the types and modalities of exercise and how to combine them as interventions for different PD subgroups (cognitive impairment, freezing, advanced disease) need further study.
RESUMO
BACKGROUND: Older adults presenting with dual-decline in cognition and walking speed face a 6-fold higher risk for dementia compared with those showing no decline. We hypothesized that the metabolomics profile of dual-decliners would be unique even before they show signs of decline in cognition and gait speed. OBJECTIVE: The objective of this study was to determine if plasma metabolomics signatures can discriminate dual-decliners from no decliners, purely cognitive decliners, and purely motor decliners prior to decline. METHODS: A retrospective cross-sectional study using baseline plasma for untargeted metabolomics analyses to investigate early signals of later dual-decline status in study participants (nâ=â76) with convenient sampling. Dual-decline was operationalized as decline in gait speed (>10âcm/s) and cognition (>2 points decline in Montreal Cognitive Assessment score) on at least two consecutive 6-monthly assessments. The participants' decliner status was evaluated 3 years after the blood sample was collected. Pair-wise comparison of detected compounds was completed using principal components and hierarchical clustering analyses. RESULTS: Analyses did not detect any cluster separation in untargeted metabolomes across baseline groups. However, follow-up analyses of specific molecules detected 4 compounds (17-Hydroxy-12-(hydroxymethyl)-10-oxo-8 oxapentacyclomethyl hexopyranoside, Fleroxacin, Oleic acid, and 5xi-11,12-Dihydroxyabieta-8(14),9(11),12-trien-20-oic acid) were at significantly higher concentration among the dual-decliners compared to non-decliners. The pure cognitive decliner group had significantly lower concentration of six compounds (1,3-nonanediol acetate, 4-(2-carboxyethyl)-2-methoxyphenyl beta-D-glucopyranosiduronic acid, oleic acid, 2E-3-[4-(sulfo-oxy)phenyl] acrylic acid, palmitelaidic acid, and myristoleic acid) compared to the non-decliner group. CONCLUSIONS: The unique metabolomics profile of dual-decliners warrants follow-up metabolomics analysis. Results may point to modifiable pathways.
RESUMO
OBJECTIVE: This work investigates if deep learning (DL) models can classify originating site locations directly from magnetic resonance imaging (MRI) scans with and without correction for intensity differences. MATERIAL AND METHODS: A large database of 1880 T1-weighted MRI scans collected across 41 sites originally for Parkinson's disease (PD) classification was used to classify sites in this study. Forty-six percent of the datasets are from PD patients, while 54% are from healthy participants. After preprocessing the T1-weighted scans, 2 additional data types were generated: intensity-harmonized T1-weighted scans and log-Jacobian deformation maps resulting from nonlinear atlas registration. Corresponding DL models were trained to classify sites for each data type. Additionally, logistic regression models were used to investigate the contribution of biological (age, sex, disease status) and non-biological (scanner type) variables to the models' decision. RESULTS: A comparison of the 3 different types of data revealed that DL models trained using T1-weighted and intensity-harmonized T1-weighted scans can classify sites with an accuracy of 85%, while the model using log-Jacobian deformation maps achieved a site classification accuracy of 54%. Disease status and scanner type were found to be significant confounders. DISCUSSION: Our results demonstrate that MRI scans encode relevant site-specific information that models could use as shortcuts that cannot be removed using simple intensity harmonization methods. CONCLUSION: The ability of DL models to exploit site-specific biases as shortcuts raises concerns about their reliability, generalization, and deployability in clinical settings.
Assuntos
Encéfalo , Aprendizado Profundo , Humanos , Encéfalo/diagnóstico por imagem , Encéfalo/patologia , Reprodutibilidade dos Testes , Imageamento por Ressonância Magnética/métodos , NeuroimagemRESUMO
Patients with Parkinson's Disease (PD) often suffer from cognitive decline. Accurate prediction of cognitive decline is essential for early treatment of at-risk patients. The aim of this study was to develop and evaluate a multimodal machine learning model for the prediction of continuous cognitive decline in patients with early PD. We included 213 PD patients from the Parkinson's Progression Markers Initiative (PPMI) database. Machine learning was used to predict change in Montreal Cognitive Assessment (MoCA) score using the difference between baseline and 4-years follow-up data as outcome. Input features were categorized into four sets: clinical test scores, cerebrospinal fluid (CSF) biomarkers, brain volumes, and genetic variants. All combinations of input feature sets were added to a basic model, which consisted of demographics and baseline cognition. An iterative scheme using RReliefF-based feature ranking and support vector regression in combination with tenfold cross validation was used to determine the optimal number of predictive features and to evaluate model performance for each combination of input feature sets. Our best performing model consisted of a combination of the basic model, clinical test scores and CSF-based biomarkers. This model had 12 features, which included baseline cognition, CSF phosphorylated tau, CSF total tau, CSF amyloid-beta1-42, geriatric depression scale (GDS) scores, and anxiety scores. Interestingly, many of the predictive features in our model have previously been associated with Alzheimer's disease, showing the importance of assessing Alzheimer's disease pathology in patients with Parkinson's disease.
Assuntos
Doença de Alzheimer , Disfunção Cognitiva , Doença de Parkinson , Humanos , Idoso , Doença de Parkinson/complicações , Doença de Parkinson/diagnóstico , Doença de Parkinson/líquido cefalorraquidiano , Doença de Alzheimer/complicações , Disfunção Cognitiva/diagnóstico , Disfunção Cognitiva/etiologia , Disfunção Cognitiva/líquido cefalorraquidiano , Cognição , Peptídeos beta-Amiloides/líquido cefalorraquidiano , Biomarcadores/líquido cefalorraquidiano , Proteínas tau/líquido cefalorraquidiano , Progressão da DoençaRESUMO
Background: Persons with Parkinson's disease (PD) differentially progress to cognitive impairment and dementia. With a 3-year longitudinal sample of initially non-demented PD patients measured on multiple dementia risk factors, we demonstrate that machine learning classifier algorithms can be combined with explainable artificial intelligence methods to identify and interpret leading predictors that discriminate those who later converted to dementia from those who did not. Method: Participants were 48 well-characterized PD patients (Mbaseline age = 71.6; SD = 4.8; 44% female). We tested 38 multi-modal predictors from 10 domains (e.g., motor, cognitive) in a computationally competitive context to identify those that best discriminated two unobserved baseline groups, PD No Dementia (PDND), and PD Incipient Dementia (PDID). We used Random Forest (RF) classifier models for the discrimination goal and Tree SHapley Additive exPlanation (Tree SHAP) values for deep interpretation. Results: An excellent RF model discriminated baseline PDID from PDND (AUC = 0.84; normalized Matthews Correlation Coefficient = 0.76). Tree SHAP showed that ten leading predictors of PDID accounted for 62.5% of the model, as well as their relative importance, direction, and magnitude (risk threshold). These predictors represented the motor (e.g., poorer gait), cognitive (e.g., slower Trail A), molecular (up-regulated metabolite panel), demographic (age), imaging (ventricular volume), and lifestyle (activities of daily living) domains. Conclusion: Our data-driven protocol integrated RF classifier models and Tree SHAP applications to selectively identify and interpret early dementia risk factors in a well-characterized sample of initially non-demented persons with PD. Results indicate that leading dementia predictors derive from multiple complementary risk domains.
RESUMO
Importance: Exercise, cognitive training, and vitamin D may enhance cognition in older adults with mild cognitive impairment (MCI). Objective: To determine whether aerobic-resistance exercises would improve cognition relative to an active control and if a multidomain intervention including exercises, computerized cognitive training, and vitamin D supplementation would show greater improvements than exercise alone. Design, Setting, and Participants: This randomized clinical trial (the SYNERGIC Study) was a multisite, double-masked, fractional factorial trial that evaluated the effects of aerobic-resistance exercise, computerized cognitive training, and vitamin D on cognition. Eligible participants were between ages 65 and 84 years with MCI enrolled from September 19, 2016, to April 7, 2020. Data were analyzed from February 2021 to December 2022. Interventions: Participants were randomized to 5 study arms and treated for 20 weeks: arm 1 (multidomain intervention with exercise, cognitive training, and vitamin D), arm 2 (exercise, cognitive training, and placebo vitamin D), arm 3 (exercise, sham cognitive training, and vitamin D), arm 4 (exercise, sham cognitive training, and placebo vitamin D), and arm 5 (control group with balance-toning exercise, sham cognitive training, and placebo vitamin D). The vitamin D regimen was a 10â¯000 IU dose 3 times weekly. Main Outcomes and Measures: Primary outcomes were changes in ADAS-Cog-13 and Plus variant at 6 months. Results: Among 175 randomized participants (mean [SD] age, 73.1 [6.6] years; 86 [49.1%] women), 144 (82%) completed the intervention and 133 (76%) completed the follow-up (month 12). At 6 months, all active arms (ie, arms 1 through 4) with aerobic-resistance exercise regardless of the addition of cognitive training or vitamin D, improved ADAS-Cog-13 when compared with control (mean difference, -1.79 points; 95% CI, -3.27 to -0.31 points; P = .02; d = 0.64). Compared with exercise alone (arms 3 and 4), exercise and cognitive training (arms 1 and 2) improved the ADAS-Cog-13 (mean difference, -1.45 points; 95% CI, -2.70 to -0.21 points; P = .02; d = 0.39). No significant improvement was found with vitamin D. Finally, the multidomain intervention (arm 1) improved the ADAS-Cog-13 score significantly compared with control (mean difference, -2.64 points; 95% CI, -4.42 to -0.80 points; P = .005; d = 0.71). Changes in ADAS-Cog-Plus were not significant. Conclusions and Relevance: In this clinical trial, older adults with MCI receiving aerobic-resistance exercises with sequential computerized cognitive training significantly improved cognition, although some results were inconsistent. Vitamin D supplementation had no effect. Our findings suggest that this multidomain intervention may improve cognition and potentially delay dementia onset in MCI. Trial Registration: ClinicalTrials.gov Identifier: NCT02808676.
Assuntos
Disfunção Cognitiva , Treino Cognitivo , Humanos , Feminino , Idoso , Masculino , Disfunção Cognitiva/terapia , Disfunção Cognitiva/psicologia , Cognição , Vitaminas/uso terapêutico , Vitaminas/farmacologia , Vitamina D/uso terapêutico , Vitamina D/farmacologia , Suplementos NutricionaisRESUMO
Background: Previous reports have suggested that patients with cerebral amyloid angiopathy (CAA) may harbor smaller white matter, basal ganglia, and cerebellar volumes compared to age-matched healthy controls (HC) or patients with Alzheimer's disease (AD). We investigated whether CAA is associated with subcortical atrophy. Methods: The study was based on the multi-site Functional Assessment of Vascular Reactivity cohort and included 78 probable CAA (diagnosed according to the Boston criteria v2.0), 33 AD, and 70 HC. Cerebral and cerebellar volumes were extracted from brain 3D T1-weighted MRI using FreeSurfer (v6.0). Subcortical volumes, including total white matter, thalamus, basal ganglia, and cerebellum were reported as proportion (%) of estimated total intracranial volume. White matter integrity was quantified by the peak width of skeletonized mean diffusivity. Results: Participants in the CAA group were older (74.0 ± 7.0, female 44%) than the AD (69.7 ± 7.5, female 42%) and HC (68.8 ± 7.8, female 69%) groups. CAA participants had the highest white matter hyperintensity volume and worse white matter integrity of the three groups. After adjusting for age, sex, and study site, CAA participants had smaller putamen volumes (mean differences, -0.024% of intracranial volume; 95% confidence intervals, -0.041% to -0.006%; p = 0.005) than the HCs but not AD participants (-0.003%; -0.024 to 0.018%; p = 0.94). Other subcortical volumes including subcortical white matter, thalamus, caudate, globus pallidus, cerebellar cortex or cerebellar white matter were comparable between all three groups. Conclusion: In contrast to prior studies, we did not find substantial atrophy of subcortical volumes in CAA compared to AD or HCs, except for the putamen. Differences between studies may reflect heterogeneity in CAA presenting syndromes or severity.
RESUMO
Introduction: Cerebral amyloid angiopathy (CAA) is a small vessel disease that causes covert and symptomatic brain hemorrhaging. We hypothesized that persons with CAA would have increased brain iron content detectable by quantitative susceptibility mapping (QSM) on magnetic resonance imaging (MRI), and that higher iron content would be associated with worse cognition. Methods: Participants with CAA (n = 21), mild Alzheimer's disease with dementia (AD-dementia; n = 14), and normal controls (NC; n = 83) underwent 3T MRI. Post-processing QSM techniques were applied to obtain susceptibility values for regions of the frontal and occipital lobe, thalamus, caudate, putamen, pallidum, and hippocampus. Linear regression was used to examine differences between groups, and associations with global cognition, controlling for multiple comparisons using the false discovery rate method. Results: No differences were found between regions of interest in CAA compared to NC. In AD, the calcarine sulcus had greater iron than NC (ß = 0.99 [95% CI: 0.44, 1.53], q < 0.01). However, calcarine sulcus iron content was not associated with global cognition, measured by the Montreal Cognitive Assessment (p > 0.05 for all participants, NC, CAA, and AD). Discussion: After correcting for multiple comparisons, brain iron content, measured via QSM, was not elevated in CAA compared to NC in this exploratory study.