A Summary Score for the Framingham Heart Study Neuropsychological Battery

Prediction of Alzheimer's disease progression within 6 years using speech: A novel approach leveraging language models

INTRODUCTION

Identification of individuals with mild cognitive impairment (MCI) who are at risk of developing Alzheimer's disease (AD) is crucial for early intervention and selection of clinical trials.

METHODS

We applied natural language processing techniques along with machine learning methods to develop a method for automated prediction of progression to AD within 6 years using speech. The study design was evaluated on the neuropsychological test interviews of n = 166 participants from the Framingham Heart Study, comprising 90 progressive MCI and 76 stable MCI cases.

RESULTS

Our best models, which used features generated from speech data, as well as age, sex, and education level, achieved an accuracy of 78.5% and a sensitivity of 81.1% to predict MCI-to-AD progression within 6 years.

DISCUSSION

The proposed method offers a fully automated procedure, providing an opportunity to develop an inexpensive, broadly accessible, and easy-to-administer screening tool for MCI-to-AD progression prediction, facilitating development of remote assessment.

HIGHLIGHTS

Voice recordings from neuropsychological exams coupled with basic demographics can lead to strong predictive models of progression to dementia from mild cognitive impairment. The study leveraged AI methods for speech recognition and processed the resulting text using language models. The developed AI-powered pipeline can lead to fully automated assessment that could enable remote and cost-effective screening and prognosis for Alzehimer's disease.

Establishing cognitive baseline in three generations: Framingham Heart Study

Introduction

Generational changes warrant recalibrating normative cognitive measures to detect changes indicative of dementia risk within each generation.

Methods

We performed linear regressions to compare eight neuropsychological (NP) tests among three‐generation cohorts at baseline in Framingham Heart Study (FHS, n = 4787) and conducted Cox regressions to investigate the relationships of NP tests with generation‐specific dementia risk.

Results

The FHS second and third generations performed better than the first generation for seven NP tests (0.14–0.81 standard deviation improvement, P ≤ .001) while the second and third generations performed similarly for six of eight NP tests (P > .05). One standard deviation better performance was associated with a higher reduction in incident dementia risk in the second than the first generation (35% vs. 24%, P_interaction = .02) for the similarities test.

Discussion

Our findings suggest cohort‐based norms are needed for cognitive assessment for the diagnosis of cognitive impairment and dementia.

Automated detection of mild cognitive impairment and dementia from voice recordings: A natural language processing approach

INTRODUCTION

Automated computational assessment of neuropsychological tests would enable widespread, cost-effective screening for dementia.

METHODS

A novel natural language processing approach is developed and validated to identify different stages of dementia based on automated transcription of digital voice recordings of subjects' neuropsychological tests conducted by the Framingham Heart Study (n = 1084). Transcribed sentences from the test were encoded into quantitative data and several models were trained and tested using these data and the participants' demographic characteristics.

RESULTS

Average area under the curve (AUC) on the held-out test data reached 92.6%, 88.0%, and 74.4% for differentiating Normal cognition from Dementia, Normal or Mild Cognitive Impairment (MCI) from Dementia, and Normal from MCI, respectively.

DISCUSSION

The proposed approach offers a fully automated identification of MCI and dementia based on a recorded neuropsychological test, providing an opportunity to develop a remote screening tool that could be adapted easily to any language.

Neuropsychological test validation of speech markers of cognitive impairment in the Framingham Cognitive Aging Cohort

AIM

Although clinicians primarily diagnose dementia based on a combination of metrics such as medical history and formal neuropsychological tests, recent work using linguistic analysis of narrative speech to identify dementia has shown promising results. We aim to build upon research by Thomas JA & Burkardt HA et al. (J Alzheimers Dis. 2020;76:905-22) and Alhanai et al. (arXiv:1710.07551v1. 2020) on the Framingham Heart Study (FHS) Cognitive Aging Cohort by 1) demonstrating the predictive capability of linguistic analysis in differentiating cognitively normal from cognitively impaired participants and 2) comparing the performance of the original linguistic features with the performance of expanded features.

METHODS

Data were derived from a subset of the FHS Cognitive Aging Cohort. We analyzed a sub-selection of 98 participants, which provided 127 unique audio files and clinical observations (n = 127, female = 47%, cognitively impaired = 43%). We built on previous work which extracted original linguistic features from transcribed audio files by extracting expanded features. We used both feature sets to train logistic regression classifiers to distinguish cognitively normal from cognitively impaired participants and compared the predictive power of the original and expanded linguistic feature sets, and participants' Mini-Mental State Examination (MMSE) scores.

RESULTS

Based on the area under the receiver-operator characteristic curve (AUC) of the models, both the original (AUC = 0.882) and expanded (AUC = 0.883) feature sets outperformed MMSE (AUC = 0.870) in classifying cognitively impaired and cognitively normal participants. Although the original and expanded feature sets had similar AUC, the expanded feature set showed better positive and negative predictive value [expanded: positive predictive value (PPV) = 0.738, negative predictive value (NPV) = 0.889; original: PPV = 0.701, NPV = 0.869].

CONCLUSIONS

Linguistic analysis has been shown to be a potentially powerful tool for clinical use in classifying cognitive impairment. This study expands the work of several others, but further studies into the plausibility of speech analysis in clinical use are vital to ensure the validity of speech analysis for clinical classification of cognitive impairment.

Sex, diabetes status and cognition: findings from the study of longevity in diabetes

INTRODUCTION

Women comprise two-thirds of people with dementia, making female sex a significant dementia risk factor. Both type 1 diabetes (T1D) and type 2 diabetes (T2D) are known dementia risk factors with an increasing global incidence. Understanding whether subtle sex differences persist in cognitive function prior to dementia in the context of diabetes may help elucidate the magnitude of sex effects on dementia risk.

RESEARCH DESIGN AND METHODS

We examined cross-sectional data from the Study of Longevity in Diabetes (SOLID), a prospective cohort study of members of Kaiser Permanente Northern California aged 60 years and older with T1D (n=758), T2D (n=232) and without either T1D or T2D (n=247). We used factor analysis to generate summary scores of cognitive domains and used regression analyses to examine the associations between sex and cognition adjusting for sociodemographic and cardiovascular confounders.

RESULTS

We included 1237 participants (630 women and 607 men) with mean age 68 years. By design, the distribution of men and women in T1D, T2D and no diabetes was similar. Women had better cognitive performance than men in global cognition (β=0.21, 95% CI 0.16 to 0.26), language (β=0.08, 95% CI 0.004 to 0.15), executive function (β=0.13, 95% CI 0.05 to 0.20), episodic verbal memory (β=0.68, 95% CI 0.59 to 0.77) and attention (β=0.20, 95% CI 0.11 to 0.28) but not in episodic visual memory (β=0.006, 95% CI -0.07 to 0.09) adjusting for age and education independent of diabetes status. We did not find an interaction between sex and diabetes status for any of the cognitive outcomes.

CONCLUSIONS

Women in late mid-life have better cognitive performance than men in many cognitive domains independent of the presence of T1D or T2D. Further work is required to understand whether these differences change over time or in older cohorts and to understand their relationship to subsequent dementia.

Long-term dietary flavonoid intake and change in cognitive function in the Framingham Offspring cohort

OBJECTIVE

To examine the association between long-term intake of total and the six classes of dietary flavonoids and decline in cognitive function over a follow-up period of up to 15 years.

DESIGN

In this longitudinal study, we evaluated change in eight cognitive domain scores (verbal and visual memory, verbal learning, attention and concentration, abstract reasoning, language, visuoperceptual organisation and the global function) based on three neuropsychological exams and characterised the annualised change between consecutive exams. Long-term intakes of total and six flavonoid classes were assessed up to four times by a validated FFQ. Repeated-measures regression models were used to examine the longitudinal association between total and six flavonoid classes and annualised change in the eight cognitive domains.

SETTING

The Framingham Heart Study (FHS), a prospective cohort study.

PARTICIPANTS

One thousand seven hundred and seventy-nine subjects who were free of dementia, aged ≥45 years and had attended at least two of the last three FHS Offspring cohort study exams.

RESULTS

Over a median follow-up of 11·8 years with 1779 participants, nominally significant trends towards a slower decline in cognitive function were observed among those with higher flavanol and flavon-3-ol intakes for global function, verbal and visual memory; higher total flavonoids and flavonoid polymers for visual memory; and higher flavanols for verbal learning.

CONCLUSIONS

In spite of modest nominal trends, overall, our findings do not support a clear association between higher long-term flavonoid intake and slowing age-related cognitive decline.

Cerebral tract integrity relates to white matter hyperintensities, cortex volume, and cognition

We examined the relationship among white matter (WM) tract integrity, WM hyperintensities (WMH), lobar gray matter (GM) volumes, and cognition in the cross-sectional Framingham Offspring Study. Six hundred eighty participants (71.7 ± 7.7 years) completed cognitive testing and magnetic resonance imaging. Diffusion tensor imaging probabilistic tractography was used to reconstruct major WM tracts. We computed tract-specific mean fractional anisotropy (FA) and tract-specific WMH ratio. Linear regressions identified relations between tracts and lobar GM volumes. Partial least squares regression examined associations between integrity of combined tracts, lobar GM volumes and cognition, including scores of memory and processing speed. Five tracts were particularly vulnerable to WMH, and tract-specific WMH volumes were inversely associated with tract-specific FA (p values < 0.05). Tract-specific FA related to lobar GM volumes. Memory was associated with lobar GM, while processing speed related to both tract integrity and lobar GM volumes. We conclude that subtle microstructural WM tract degeneration relates to specific lobar GM atrophy. The integrity of associated WM tracts and GM lobes differentially impacts memory and processing speed.

Risk for cognitive impairment across 22 measures of cognitive ability in early-stage chronic kidney disease

Background

Chronic kidney disease (CKD) is a significant risk factor for cognitive impairment. Previous studies have examined differences in cognitive impairment between persons with and without CKD using multiple cognitive outcomes, but few have done this for an extensive battery of cognitive tests. We relate early-stage CKD to two indices of impairment for 22 measures of cognitive ability.

Methods

The study was community-based and cross-sectional with 898 individuals free from dementia and end-stage renal disease. Estimated glomerular filtration rate (eGFR) was calculated using the chronic kidney disease epidemiology collaboration equation and classified as <60 or ≥60 mL/min/1.73 m2, based on consensus definitions of Stage 3 or greater CKD. The eGFR classifications were related to modest [≥1 standard deviation (SD) below the mean] and severe (≥1.5 SD below the mean) impairment on each measure using logistic regression analyses adjusting for potential risk factors.

Results

A total of 146 individuals (16.3%) had eGFR <60 mL/min/1.73 m2 (mean 51.6 ± 10.1 mL/min/1.73 m2). These participants had significantly greater risk for modestly impaired abilities in the scanning and tracking and visual-spatial organization/memory (VSOM) domains after accounting for comorbidity-related risk factors [odds ratios (ORs) between 1.68 and 2.16], as well as greater risk for severely impaired functioning in the language domain (OR = 2.65).

Conclusions

Participants with eGFR <60 mL/min/1.73 m2 were at higher risk for cognitive impairment than those with eGFR ≥60 mL/min/1.73 m2 on the majority of cognitive abilities, specifically those within the VSOM, Language, and scanning and tracking domains. Targeted screening for cognitive deficits in kidney disease patients early in their disease course may be warranted.

Racial and ethnic differences in cognitive function among older adults in the USA

OBJECTIVE

Examine differences in cognition between Hispanic, non-Hispanic black (NHB), and non-Hispanic white (NHW) older adults in the United States.

DATA/METHODS

The final sample includes 18 982 participants aged 51 or older who received a modified version of the Telephone Interview for Cognitive Status during the 2010 Health and Retirement Study follow-up. Ordinary least squares will be used to examine differences in overall cognition according to race/ethnicity.

RESULTS

Hispanics and NHB had lower cognition than NHW for all age groups (51-59, 60-69, 70-79, 80+). Hispanics had higher cognition than NHB for all age groups but these differences were all within one point. The lower cognition among NHB compared to NHW remained significant after controlling for age, gender, and education, whereas the differences in cognition between Hispanics and NHW were no longer significant after controlling for these covariates. Cognitive scores increased with greater educational attainment for all race/ethnic groups, but Hispanics exhibited the least benefit.

DISCUSSION

Our results highlight the role of education in race/ethnic differences in cognitive function during old age. Education seems beneficial for cognition in old age for all race/ethnic groups, but Hispanics appear to receive a lower benefit compared to other race/ethnic groups. Further research is needed on the racial and ethnic differences in the pathways of the benefits of educational attainment for late-life cognitive function. Copyright © 2016 John Wiley & Sons, Ltd.