Exploring cognitive progression subtypes in the Framingham Heart Study.

INTRODUCTION: Alzheimer's disease (AD) is a heterogeneous disorder characterized by complex underlying neuropathology that is not fully understood. This study aimed to identify cognitive progression subtypes and examine their correlation with clinical outcomes. METHODS: Participants of this study were recruited from the Framingham Heart Study. The Subtype and Stage Inference (SuStaIn) method was used to identify cognitive progression subtypes based on eight cognitive domains. RESULTS: Three cognitive progression subtypes were identified, including verbal learning (Subtype 1), abstract reasoning (Subtype 2), and visual memory (Subtype 3). These subtypes represent different domains of cognitive decline during the progression of AD. Significant differences in age of onset among the different subtypes were also observed. A higher SuStaIn stage was significantly associated with increased mortality risk. DISCUSSION: This study provides a characterization of AD heterogeneity in cognitive progression, emphasizing the importance of developing personalized approaches for risk stratification and intervention. Highlights: We used the Subtype and Stage Inference (SuStaIn) method to identify three cognitive progression subtypes.Different subtypes have significant variations in age of onset.Higher stages of progression are associated with increased mortality risk.

Multimodal Machine Learning for 10-Year Dementia Risk Prediction: The Framingham Heart Study.

BACKGROUND: Early prediction of dementia risk is crucial for effective interventions. Given the known etiologic heterogeneity, machine learning methods leveraging multimodal data, such as clinical manifestations, neuroimaging biomarkers, and well-documented risk factors, could predict dementia more accurately than single modal data. OBJECTIVE: This study aims to develop machine learning models that capitalize on neuropsychological (NP) tests, magnetic resonance imaging (MRI) measures, and clinical risk factors for 10-year dementia prediction. METHODS: This study included participants from the Framingham Heart Study, and various data modalities such as NP tests, MRI measures, and demographic variables were collected. CatBoost was used with Optuna hyperparameter optimization to create prediction models for 10-year dementia risk using different combinations of data modalities. The contribution of each modality and feature for the prediction task was also quantified using Shapley values. RESULTS: This study included 1,031 participants with normal cognitive status at baseline (age 75±5 years, 55.3% women), of whom 205 were diagnosed with dementia during the 10-year follow-up. The model built on three modalities demonstrated the best dementia prediction performance (AUC 0.90±0.01) compared to single modality models (AUC range: 0.82-0.84). MRI measures contributed most to dementia prediction (mean absolute Shapley value: 3.19), suggesting the necessity of multimodal inputs. CONCLUSION: This study shows that a multimodal machine learning framework had a superior performance for 10-year dementia risk prediction. The model can be used to increase vigilance for cognitive deterioration and select high-risk individuals for early intervention and risk management.

Prediction of Progression from Mild Cognitive Impairment to Alzheimer's disease with Longitudinal and Multimodal Data.

Introduction: Accurate prediction of the progression from mild cognitive impairment (MCI) to Alzheimer's disease (AD) within a certain time frame is crucial for appropriate therapeutic interventions. However, it is challenging to capture the dynamic changes in cognitive and functional abilities over time, resulting in limited predictive performance. Our study aimed to investigate whether incorporating longitudinal multimodal data with advanced analytical methods could improve the capability to predict the risk of progressing to AD. Methods: This study included participants from the Alzheimer's Disease Neuroimaging Initiative (ADNI), a large-scale multi-center longitudinal study. Three data modalities, including demographic variables, neuropsychological tests, and neuroimaging measures were considered. A Long Short-Term Memory (LSTM) model using data collected at five-time points (baseline, 6-month, 12-month, 18-month, and 24-month) was developed to predict the risk of progression from MCI to AD within two years from the index exam (the exam at 24-month). In contrast, a random forest model was developed to predict the risk of progression just based on the data collected at the index exam. Results: The study included 347 participants with MCI at 24-month (age: mean 75, SD 7 years; 39.8% women) from ADNI, of whom 77 converted to AD over a 2-year follow-up period. The longitudinal LSTM model showed superior prediction performance of MCI-to-AD progression (AUC 0.93±0.06) compared to the random forest model (AUC 0.90±0.09). A similar pattern was also observed across different age groups. Discussion: Our study suggests that the incorporation of longitudinal data can provide better predictive performance for 2-year MCI-to-AD progression risk than relying solely on cross-sectional data. Therefore, repeated or multiple times routine health surveillance of MCI patients are essential in the early detection and intervention of AD.

Association Between Acoustic Features and Brain Volumes: the Framingham Heart Study.

Introduction: Although brain magnetic resonance imaging (MRI) is a valuable tool for investigating structural changes in the brain associated with neurodegeneration, the development of non-invasive and cost-effective alternative methods for detecting early cognitive impairment is crucial. The human voice has been increasingly used as an indicator for effectively detecting cognitive disorders, but it remains unclear whether acoustic features are associated with structural neuroimaging. Methods: This study aims to investigate the association between acoustic features and brain volume and compare the predictive power of each for mild cognitive impairment (MCI) in a large community-based population. The study included participants from the Framingham Heart Study (FHS) who had at least one voice recording and an MRI scan. Sixty-five acoustic features were extracted with the OpenSMILE software (v2.1.3) from each voice recording. Nine MRI measures were derived according to the FHS MRI protocol. We examined the associations between acoustic features and MRI measures using linear regression models adjusted for age, sex, and education. Acoustic composite scores were generated by combining acoustic features significantly associated with MRI measures. The MCI prediction ability of acoustic composite scores and MRI measures were compared by building random forest models and calculating the mean area under the receiver operating characteristic curve (AUC) of 10-fold cross-validation. Results: The study included 4,293 participants (age 57 ± 13 years, 53.9% women). During 9.3±3.7 years follow-up, 106 participants were diagnosed with MCI. Seven MRI measures were significantly associated with more than 20 acoustic features after adjusting for multiple testing. The acoustic composite scores can improve the AUC for MCI prediction to 0.794, compared to 0.759 achieved by MRI measures. Discussion: We found multiple acoustic features were associated with MRI measures, suggesting the potential for using acoustic features as easily accessible digital biomarkers for the early diagnosis of MCI.

Association Between Elevated Depressive Symptoms and Cognitive Function Moderated by APOE4 Status: Framingham Offspring Study.

BACKGROUND: Depression and Apolipoprotein E4 (APOE4) are associated with decreased cognitive function and differences in brain structure. OBJECTIVE: This study investigated whether APOE4 status moderates the association between elevated depressive symptoms, cognitive function, and brain structure. METHODS: Stroke- and dementia-free participants (n = 1,968) underwent neuropsychological evaluation, brain MRI, and depression screening. Linear and logistic regression was used to examine all associations. Secondary analyses were performed using interaction terms to assess effect modification by APOE4 status. RESULTS: Elevated depressive symptoms were associated with lower cognitive performance in several domains. In stratified analyses, elevated depressive symptoms were associated with poorer visual short- and long-term memory performance for APOE4 + participants. Elevated depressive symptoms were not associated with any brain structure in this study sample. CONCLUSION: Elevated depressive symptoms impact cognitive function in non-demented individuals. Having the APOE4 allele may exacerbate the deleterious effects of elevated depressive symptoms on visual memory performance. Screening for elevated depressive symptoms in both research studies and clinical practice may be warranted to avoid false positive identification of neurodegeneration, particularly among those who are APOE4 + .

Deep ensemble learning for Alzheimer's disease classification.

Ensemble learning uses multiple algorithms to obtain better predictive performance than any single one of its constituent algorithms could. With the growing popularity of deep learning technologies, researchers have started to ensemble these technologies for various purposes. Few, if any, however, have used the deep learning approach as a means to ensemble Alzheimer's disease classification algorithms. This paper presents a deep ensemble learning framework that aims to harness deep learning algorithms to integrate multisource data and tap the 'wisdom of experts'. At the voting layer, two sparse autoencoders are trained for feature learning to reduce the correlation of attributes and diversify the base classifiers ultimately. At the stacking layer, a nonlinear feature-weighted method based on a deep belief network is proposed to rank the base classifiers, which may violate the conditional independence. The neural network is used as a meta classifier. At the optimizing layer, over-sampling and threshold-moving are used to cope with the cost-sensitive problem. Optimized predictions are obtained based on an ensemble of probabilistic predictions by similarity calculation. The proposed deep ensemble learning framework is used for Alzheimer's disease classification. Experiments with the clinical dataset from National Alzheimer's Coordinating Center demonstrate that the classification accuracy of our proposed framework is 4% better than six well-known ensemble approaches, including the standard stacking algorithm as well. Adequate coverage of more accurate diagnostic services can be provided by utilizing the wisdom of averaged physicians. This paper points out a new way to boost the primary care of Alzheimer's disease from the view of machine learning.

Identification of digital voice biomarkers for cognitive health.

AIM: Human voice contains rich information. Few longitudinal studies have been conducted to investigate the potential of voice to monitor cognitive health. The objective of this study is to identify voice biomarkers that are predictive of future dementia. METHODS: Participants were recruited from the Framingham Heart Study. The vocal responses to neuropsychological tests were recorded, which were then diarized to identify participant voice segments. Acoustic features were extracted with the OpenSMILE toolkit (v2.1). The association of each acoustic feature with incident dementia was assessed by Cox proportional hazards models. RESULTS: Our study included 6, 528 voice recordings from 4, 849 participants (mean age 63 ± 15 years old, 54.6% women). The majority of participants (71.2%) had one voice recording, 23.9% had two voice recordings, and the remaining participants (4.9%) had three or more voice recordings. Although all asymptomatic at the time of examination, participants who developed dementia tended to have shorter segments than those who were dementia free (P < 0.001). Additionally, 14 acoustic features were significantly associated with dementia after adjusting for multiple testing (P < 0.05/48 = 1 × 10-3). The most significant acoustic feature was jitterDDP_sma_de (P = 7.9 × 10-7), which represents the differential frame-to-frame Jitter. A voice based linear classifier was also built that was capable of predicting incident dementia with area under curve of 0.812. CONCLUSIONS: Multiple acoustic and linguistic features are identified that are associated with incident dementia among asymptomatic participants, which could be used to build better prediction models for passive cognitive health monitoring.

Using data science to diagnose and characterize heterogeneity of Alzheimer's disease.

INTRODUCTION: Despite the availability of age- and education-adjusted standardized scores for most neuropsychological tests, there is a lack of objective rules in how to interpret multiple concurrent neuropsychological test scores that characterize the heterogeneity of Alzheimer's disease. METHODS: Using neuropsychological test scores of 2091 participants from the Framingham Heart Study, we devised an automated algorithm that follows general diagnostic criteria and explores the heterogeneity of Alzheimer's disease. RESULTS: We developed a series of stepwise diagnosis rules that evaluate information from multiple neuropsychological tests to produce an intuitive and objective Alzheimer's disease dementia diagnosis with more than 80% accuracy. DISCUSSION: A data-driven stepwise diagnosis system is useful for diagnosis of Alzheimer's disease from neuropsychological tests. It demonstrated better performance than the traditional dichotomization of individuals' performance into satisfactory and unsatisfactory outcomes, making it more reflective of dementia as a spectrum disorder. This algorithm can be applied to both within clinic and outside-of-clinic settings.