Caselli, Richard J.

Background: Disease progression prediction based on neuroimaging biomarkers is vital in Alzheimer’s disease (AD) research. Convolutional neural networks (CNN) have been proved to be powerful for various computer vision research by refining reliable and high-level feature maps from image patches. Objective: A key challenge in applying CNN to neuroimaging research is the limited labeled samples with high dimensional features. Another challenge is how to improve the prediction accuracy by joint analysis of multiple data sources (i.e., multiple time points or multiple biomarkers). To address these two challenges, we propose a novel multi-task learning framework based on CNN. Methods: First, we pre-trained CNN on the ImageNet dataset and transferred the knowledge from the pre-trained model to neuroimaging representation. We used this deep model as feature extractor to generate high-level feature maps of different tasks. Then a novel unsupervised learning method, termed Multi-task Stochastic Coordinate Coding (MSCC), was proposed for learning sparse features of multi-task feature maps by using shared and individual dictionaries. Finally, Lasso regression was performed on these multi-task sparse features to predict AD progression measured by the Mini-Mental State Examination (MMSE) and the Alzheimer’s Disease Assessment Scale cognitive subscale (ADAS-Cog). Results: We applied this novel CNN-MSCC system on the Alzheimer’s Disease Neuroimaging Initiative dataset to predict future MMSE/ADAS-Cog scales. We found our method achieved superior performances compared with seven other methods. Conclusion: Our work may add new insights into data augmentation and multi-task deep model research and facilitate the adoption of deep models in neuroimaging research.

AbstractIntroductionWe previously characterized associations between brain imaging measurements of amyloid‐β (Aβ) plaque burden and apolipoprotein E (APOE) ε4 gene dose in a small number of cognitively unimpaired late‐middle‐aged APOE ε4 homozygotes (HMs), heterozygotes (HTs), and noncarriers (NCs). We now characterize cross‐sectional Aβ plaque, tau tangle, and cortical atrophy (neurodegeneration) measurements, classifications, and associations with age in a larger number of unimpaired HMs, HTs, and NCs over a wider age range.MethodsWe analyzed 11C Pittsburgh compound B (Aβ) positron emission tomography (PET), flortaucipir (tau) PET, and volumetric magnetic resonance imaging data from 164 study participants of age 47–86 years, including 26 APOE ε4 HMs, 48 HTs, and 90 NCs matched for age and sex.ResultsAβ PET measurements rose, plateaued at the respective ages of 68 and 76, and then declined with age in unimpaired HM and HT groups. Compared with NCs, these two groups began to have significantly higher Aβ PET measurements at ages 62 and 70, respectively, and no longer had significantly higher measurements by ages 71 and 78, respectively. They began to have significantly higher entorhinal cortex tau PET measurements at ages 66 and 70, respectively, and no longer had significantly higher measurements by ages 74 and 78, respectively. Brain atrophy measurements tended to decline slowly with age in all three genetic groups. Their elevated tau PET measurements were attributable to those with positive Aβ PET scans. 41.0%, 18.0%, and 5.0% of the 47‐ to 70‐year‐old HMs, HTs, and NCs and 25.0%, 79.0%, and 38.0% of the 71‐ to 86‐year‐old HMs, HTs, and NCs had positive Aβ PET scans, and the long‐term recall memory scores are significantly higher in the older HMs than in HT and NC groups, suggesting resistance to Aβ deposition in those HMs who remained unimpaired at older ages.ConclusionsThis study provides information about Aβ plaque burden, tau tangle burden, and neurodegeneration in cognitively unimpaired persons at three levels of genetic risk for AD. Unimpaired APOE ε4 HMs can be studied before their 70s to evaluate the understanding of factors, processes, and interventions involved in the predisposition to and prevention of AD, and after their 70s, to discover factors, processes, and interventions involved in the resilience or resistance to and prevention of AD.

Abstract Introduction Some Alzheimer's disease biomarker studies found amyloid changes 20 years or more in advance of expected symptoms, while cognitive changes lagged for more than a decade, but this apparent lag might reflect the sensitivities of the biomarker and cognitive assays used. How far in advance of incident amnestic mild cognitive impairment (MCI) does cognition begin to decline? Methods Longitudinal neuropsychological study of an apolipoprotein E e4 enriched cohort of cognitively normal individuals at entry. Linear mixed models for MCI converters (n = 65) and nonconverters (n = 719) fitted for each neuropsychological measure; annual changes compared between groups before and after linear model intersections (inflection points). Results 34 of 35 cognitive measures and 9 of 18 behavioral measures declined faster post‐inflection in the MCI converters; the earliest cognitive inflection point was nearly 20 years in advance of MCI diagnosis. Discussion The preclinical duration of cognitive and behavioral changes approaches the earliest reported biomarker changes.

Objective: To investigate the association of a 6-month Zumba intervention with cognition and quality of life among older cognitively unimpaired apolipoprotein ∊4 (APOE4) carrier and noncarrier women. Methods: Fifty-three women were randomly assigned to either twice-weekly Zumba group classes or maintenance of habitual exercise (control group) for 6 months. At baseline, 3, and 6 months, all participants underwent neuropsychological, physical activity, and quality-of-life assessments. Results: Overall, neuropsychological test scores and level of physical activity did not differ between intervention and control groups at any time. However, compared to the control group, quality of life was higher at 3 months, and visuospatial working memory and response inhibition improved more in the intervention group by 6 months. Apolipoprotein ∊4 status did not affect the results. Discussion: Zumba may strengthen performance on visuospatial working memory among cognitively unimpaired older women but this needs to be tested in a larger clinical trial.

AbstractAlzheimer's disease (AD) is the most common type of dementia in the elderly with no effective treatment currently. Recent studies of noninvasive neuroimaging, resting‐state functional magnetic resonance imaging (rs‐fMRI) with graph theoretical analysis have shown that patients with AD and mild cognitive impairment (MCI) exhibit disrupted topological organization in large‐scale brain networks. In previous work, it is a common practice to threshold such networks. However, it is not only difficult to make a principled choice of threshold values, but also worse is the discard of potential important information. To address this issue, we propose a threshold‐free feature by integrating a prior persistent homology‐based topological feature (the zeroth Betti number) and a newly defined connected component aggregation cost feature to model brain networks over all possible scales. We show that the induced topological feature (Integrated Persistent Feature) follows a monotonically decreasing convergence function and further propose to use its slope as a concise and persistent brain network topological measure. We apply this measure to study rs‐fMRI data from the Alzheimer's Disease Neuroimaging Initiative and compare our approach with five other widely used graph measures across five parcellation schemes ranging from 90 to 1,024 region‐of‐interests. The experimental results demonstrate that the proposed network measure shows more statistical power and stronger robustness in group difference studies in that the absolute values of the proposed measure of AD are lower than MCI and much lower than normal controls, providing empirical evidence for decreased functional integration in AD dementia and MCI.