Archives

  • 2018-07
  • 2019-04
  • 2019-05
  • 2019-06
  • 2019-07
  • 2019-08
  • 2019-09
  • 2019-10
  • 2019-11
  • 2019-12
  • 2020-01
  • 2020-02
  • 2020-03
  • 2020-04
  • 2020-05
  • 2020-06
  • 2020-07
  • 2020-08
  • 2020-09
  • 2020-10
  • 2020-11
  • 2020-12
  • 2021-01
  • 2021-02
  • 2021-03
  • 2021-04
  • 2021-05
  • 2021-06
  • 2021-07
  • 2021-08
  • 2021-09
  • 2021-10
  • 2021-11
  • 2021-12
  • 2022-01
  • 2022-02
  • 2022-03
  • 2022-04
  • 2022-05
  • 2022-06
  • 2022-07
  • 2022-08
  • 2022-09
  • 2022-10
  • 2022-11
  • 2022-12
  • 2023-01
  • 2023-02
  • 2023-03
  • 2023-04
  • 2023-05
  • 2023-06
  • 2023-08
  • 2023-09
  • 2023-10
  • 2023-11
  • 2023-12
  • 2024-01
  • 2024-02
  • 2024-03
  • 2024-04
  • 2024-05
  • br Special report Financial and

    2024-05-14


    Special report – Financial and personal benefits of early diagnosis
    Acknowledgments The Alzheimer’s Association acknowledges the contributions of Joseph Gaugler, Ph.D., Bryan James, Ph.D., Tricia Johnson, Ph.D., Allison Marin, Ph.D., and Jennifer Weuve, M.P.H., Sc.D., in the preparation of 2018 Alzheimer’s Disease Facts and Figures.
    Main Text
    Introduction Alzheimer's disease (AD) is a severe neurodegenerative disorder that results in a loss of mental function due to the deterioration of 11021 tissue, leading directly to death (Khachaturian, 1985). It accounts for 60–70% of age related dementia, affecting an estimated 30 million individuals in 2011 and the number is projected to be over 114 million by 2050 (Wimo et al., 2003). The cause of AD is poorly understood and currently there is no cure for AD. AD has a long preclinical phase, lasting a decade or more. There is increasing research emphasis on detecting AD in the pre-clinical phase, before the onset of the irreversible neuron loss that characterizes the dementia phase of the disease, since therapies/treatment are most likely to be effective in this early phase. The Alzheimer's Disease Neuroimaging Initiative (ADNI, http://adni.loni.usc.edu/) has been facilitating the scientific evaluation of neuroimaging data including magnetic resonance imaging (MRI), positron emission tomography (PET), along with other biomarkers, clinical and neuropsychological assessments for predicting the onset and progression of MCI (mild cognitive impairment) and AD. Early diagnosis of AD is key to the development, assessment, and monitoring of new treatments for AD. A recent study proposed a prior knowledge guided regression model, using the group information to enforce the intra-group similarity with group sparse methods. In recent work, these existing ideas have been combined in Group-sparse Multitask Regression and Feature Selection (G-SMuRFS) (Yan et al., 2015, Wang et al., 2012) which takes into account coupled feature and group sparsity across tasks and uses vertex-based cortical surface measures in an anatomically meaningful manner. Since brain structures tend to work together to achieve a certain function, brain imaging measures are often correlated with each other. It assumes (1) possible partition exists among predictors, and (2) predictors within one partition should have similar weights. However, there exists three limitations of G-SMuRFS: (1) G-SMuRFS allows to learn a common subset of brain regions across all the tasks simultaneously with a Group ℓ2,1-norm. This assumption is too restrictive since different tasks may prefer different brain regions. It is desirable to select the specific ROIs for different tasks. (2) All scores are modeled with Gaussian (least squares) regression in G-SMuRFS, whereas it is not appropriate for all the scores. From Fig. 1, it can be seen that the distribution of scores of TOTAL and ADAS are Gaussian and three scores (T30, RECOG and MMSE) are Poisson. (3) The optimization of G-SMuRFS was done based on an iterative alternative optimization (AO) algorithm, which is an approximate gradient (not sub-gradient) descent method to handle sparse coefficient blocks and results in an inaccurate solution. In order to solve these limitations, we propose a multi-task sparse group lasso (MT-SGL) method which encourages individual feature selection coupled with group selection with sparsity-inducing norm. Instead of learning a shared representation from the level of feature and region across all the tasks simultaneously, the MT-SGL formulation which encourages (a) individual feature selection based on the utility of the features across all tasks and (b) task specific group selection based on the utility of the group to decouple the ROIs sharing across tasks allowing for more flexibility. Moreover, the proposed MT-SGL framework can use general loss functions, including losses derived from generalized linear models (GLMs). In our experiments, we consider MT-SGL models corresponding to Gaussian regression (least squares) as well as Poisson regression, inspired by the response profiles of some cognitive scores. Fig. 2 illustrates a schematic diagram of the proposed framework for cognitive score prediction and biomarker discovery.