There are many active research projects accessing and applying shared ADNI data. Use the search above to find specific research focuses on the active ADNI investigations. This information is requested annually as a requirement for data access.
| Principal Investigator | |
| Principal Investigator's Name: | Esa Pitkänen |
| Institution: | University of Helsinki |
| Department: | Institute for Molecular Medicine Finland (FIMM) |
| Country: | |
| Proposed Analysis: | We propose to develop novel deep machine learning methods to jointly model PET and MRI imaging data, and genetics data, in order to define and analyze associations between biologically and clinically defined Alzheimer’s disease (AD) groups. Application of deep learning methods has resulted in breakthroughs in many fields such as digital histopathology and predictive modeling using electronic medical records (Fu et al. 2020; Rajkomar et al. 2018). Deep learning also shows promise in the analysis of AD pathologies (Jo, Nho, and Saykin 2019; Tang et al. 2019; Qiu et al. 2020). A key factor driving these developments has been the availability of ADNI data to train large-scale deep neural network models. A recent methodological innovation has been the introduction of Transformer models, a class of deep neural network models aimed at analysis of sequential data (Vaswani et al. 2017). Transformers are also well-suited to integrate data with multiple modalities. Two major breakthroughs have been achieved with Transformers: state-of-the-art natural language modeling (GPT-3) (Brown et al. 2020) and protein folding prediction (AlphaFold 2; CASP14 challenge). To our knowledge, no Transformer-based model exists to integrate data modalities relevant to AD with the purpose of delineating amyloid and tau pathologies. NIA-AA research framework groups biomarkers of AD into β amyloid deposition, pathologic tau, and neurodegeneration [AT(N)] (Jack et al. 2018). This biological approach has been suggested to yield three biologically defined groups: Alzheimer continuum (abnormal amyloid regardless of tau status), Alzheimer pathologic change (abnormal amyloid but normal tau), and Alzheimer disease (abnormal amyloid and tau), and these may or may not be related to clinical symptoms (Jack et al. 2018). Biologically defined AD has reported to be more common than clinically defined probable AD, and indicates that some people with biologically defined AD are asymptomatic (Jack et al. 2019). Deep learning techniques could help to clarify associations between biologically and clinically defined AD. In our study we aim to develop and utilize models based on Transformer and convolutional neural networks (CNN) in order to understand how amyloid and tau pathologies arise. In order to do this, we will create deep neural network models to process PET and MRI imaging data, and to predict the biologically defined AD groups. We aim to detect subgroups with different patterns of AD related pathology in terms of AT(N) framework. We will further study differences in neuropsychological functioning, neurodegeneration (hippocampal and cortical atrophy) and polygenic risk score of AD (AD-PRS) between subgroups of different patterns of cortical amyloid and tau pathology. These analyzes have two main aims: 1) to detect cortical amyloid and tau patterns that are related to clinical (episodic memory impairment), neurodegenerative and genetic markers (AD-PRS) of AD; and 2) to predict progression from MCI to AD based on cortical amyloid and tau patterns. Validation of our model predictions would be carried out by comparison to amyloid and tau levels measured in cerebrospinal fluid samples. A key strength of deep neural networks is the ability to perform feature learning i.e. learning higher-level features from raw data such as MRI or PET images. Concordantly, we will aim at maximizing the interpretability of our models in order to be able to extract learnt features and derive explanations for model predictions. Keywords: alzheimer’s, machine learning, deep learning, genetics, genomics, MRI, PET References Brown, Tom B., Benjamin Mann, Nick Ryder, Melanie Subbiah, Jared Kaplan, Prafulla Dhariwal, Arvind Neelakantan, et al. 2020. “Language Models Are Few-Shot Learners.” arXiv [cs.CL]. arXiv. http://arxiv.org/abs/2005.14165. Fu, Yu, Alexander W. Jung, Ramon Viñas Torne, Santiago Gonzalez, Harald Vöhringer, Artem Shmatko, Lucy R. Yates, Mercedes Jimenez-Linan, Luiza Moore, and Moritz Gerstung. 2020. “Pan-Cancer Computational Histopathology Reveals Mutations, Tumor Composition and Prognosis.” Nature Cancer 1 (8): 800–810. Jack, Clifford R., Jr, David A. Bennett, Kaj Blennow, Maria C. Carrillo, Billy Dunn, Samantha Budd Haeberlein, David M. Holtzman, et al. 2018. “NIA-AA Research Framework: Toward a Biological Definition of Alzheimer’s Disease.” Alzheimer’s & Dementia: The Journal of the Alzheimer's Association 14 (4): 535–62. Jack, Clifford R., Jr, Terry M. Therneau, Stephen D. Weigand, Heather J. Wiste, David S. Knopman, Prashanthi Vemuri, Val J. Lowe, et al. 2019. “Prevalence of Biologically vs Clinically Defined Alzheimer Spectrum Entities Using the National Institute on Aging-Alzheimer’s Association Research Framework.” JAMA Neurology, July. https://doi.org/10.1001/jamaneurol.2019.1971. Jo, Taeho, Kwangsik Nho, and Andrew J. Saykin. 2019. “Deep Learning in Alzheimer’s Disease: Diagnostic Classification and Prognostic Prediction Using Neuroimaging Data.” Frontiers in Aging Neuroscience 11 (August): 220. Qiu, Shangran, Prajakta S. Joshi, Matthew I. Miller, Chonghua Xue, Xiao Zhou, Cody Karjadi, Gary H. Chang, et al. 2020. “Development and Validation of an Interpretable Deep Learning Framework for Alzheimer’s Disease Classification.” Brain: A Journal of Neurology 143 (6): 1920–33. Rajkomar, Alvin, Eyal Oren, Kai Chen, Andrew M. Dai, Nissan Hajaj, Michaela Hardt, Peter J. Liu, et al. 2018. “Scalable and Accurate Deep Learning with Electronic Health Records.” NPJ Digital Medicine 1 (May): 18. Tang, Ziqi, Kangway V. Chuang, Charles DeCarli, Lee-Way Jin, Laurel Beckett, Michael J. Keiser, and Brittany N. Dugger. 2019. “Interpretable Classification of Alzheimer’s Disease Pathologies with a Convolutional Neural Network Pipeline.” Nature Communications 10 (1): 2173. Vaswani, Ashish, Noam Shazeer, Niki Parmar, Jakob Uszkoreit, Llion Jones, Aidan N. Gomez, Ł. Ukasz Kaiser, and Illia Polosukhin. 2017. “Attention Is All You Need.” In Advances in Neural Information Processing Systems 30, edited by I. Guyon, U. V. Luxburg, S. Bengio, H. Wallach, R. Fergus, S. Vishwanathan, and R. Garnett, 5998–6008. Curran Associates, Inc. |
| Additional Investigators | |
| Investigator's Name: | Eero Vuoksimaa |
| Proposed Analysis: | We propose to develop novel deep machine learning methods to jointly model PET and MRI imaging data, and genetics data, in order to define and analyze associations between biologically and clinically defined Alzheimer’s disease (AD) groups. Application of deep learning methods has resulted in breakthroughs in many fields such as digital histopathology and predictive modeling using electronic medical records (Fu et al. 2020; Rajkomar et al. 2018). Deep learning also shows promise in the analysis of AD pathologies (Jo, Nho, and Saykin 2019; Tang et al. 2019; Qiu et al. 2020). A key factor driving these developments has been the availability of ADNI data to train large-scale deep neural network models. A recent methodological innovation has been the introduction of Transformer models, a class of deep neural network models aimed at analysis of sequential data (Vaswani et al. 2017). Transformers are also well-suited to integrate data with multiple modalities. Two major breakthroughs have been achieved with Transformers: state-of-the-art natural language modeling (GPT-3) (Brown et al. 2020) and protein folding prediction (AlphaFold 2; CASP14 challenge). To our knowledge, no Transformer-based model exists to integrate data modalities relevant to AD with the purpose of delineating amyloid and tau pathologies. NIA-AA research framework groups biomarkers of AD into β amyloid deposition, pathologic tau, and neurodegeneration [AT(N)] (Jack et al. 2018). This biological approach has been suggested to yield three biologically defined groups: Alzheimer continuum (abnormal amyloid regardless of tau status), Alzheimer pathologic change (abnormal amyloid but normal tau), and Alzheimer disease (abnormal amyloid and tau), and these may or may not be related to clinical symptoms (Jack et al. 2018). Biologically defined AD has reported to be more common than clinically defined probable AD, and indicates that some people with biologically defined AD are asymptomatic (Jack et al. 2019). Deep learning techniques could help to clarify associations between biologically and clinically defined AD. In our study we aim to develop and utilize models based on Transformer and convolutional neural networks (CNN) in order to understand how amyloid and tau pathologies arise. In order to do this, we will create deep neural network models to process PET and MRI imaging data, and to predict the biologically defined AD groups. We aim to detect subgroups with different patterns of AD related pathology in terms of AT(N) framework. We will further study differences in neuropsychological functioning, neurodegeneration (hippocampal and cortical atrophy) and polygenic risk score of AD (AD-PRS) between subgroups of different patterns of cortical amyloid and tau pathology. These analyzes have two main aims: 1) to detect cortical amyloid and tau patterns that are related to clinical (episodic memory impairment), neurodegenerative and genetic markers (AD-PRS) of AD; and 2) to predict progression from MCI to AD based on cortical amyloid and tau patterns. Validation of our model predictions would be carried out by comparison to amyloid and tau levels measured in cerebrospinal fluid samples. A key strength of deep neural networks is the ability to perform feature learning i.e. learning higher-level features from raw data such as MRI or PET images. Concordantly, we will aim at maximizing the interpretability of our models in order to be able to extract learnt features and derive explanations for model predictions. Keywords: alzheimer’s, machine learning, deep learning, genetics, genomics, MRI, PET References Brown, Tom B., Benjamin Mann, Nick Ryder, Melanie Subbiah, Jared Kaplan, Prafulla Dhariwal, Arvind Neelakantan, et al. 2020. “Language Models Are Few-Shot Learners.” arXiv [cs.CL]. arXiv. http://arxiv.org/abs/2005.14165. Fu, Yu, Alexander W. Jung, Ramon Viñas Torne, Santiago Gonzalez, Harald Vöhringer, Artem Shmatko, Lucy R. Yates, Mercedes Jimenez-Linan, Luiza Moore, and Moritz Gerstung. 2020. “Pan-Cancer Computational Histopathology Reveals Mutations, Tumor Composition and Prognosis.” Nature Cancer 1 (8): 800–810. Jack, Clifford R., Jr, David A. Bennett, Kaj Blennow, Maria C. Carrillo, Billy Dunn, Samantha Budd Haeberlein, David M. Holtzman, et al. 2018. “NIA-AA Research Framework: Toward a Biological Definition of Alzheimer’s Disease.” Alzheimer’s & Dementia: The Journal of the Alzheimer's Association 14 (4): 535–62. Jack, Clifford R., Jr, Terry M. Therneau, Stephen D. Weigand, Heather J. Wiste, David S. Knopman, Prashanthi Vemuri, Val J. Lowe, et al. 2019. “Prevalence of Biologically vs Clinically Defined Alzheimer Spectrum Entities Using the National Institute on Aging-Alzheimer’s Association Research Framework.” JAMA Neurology, July. https://doi.org/10.1001/jamaneurol.2019.1971. Jo, Taeho, Kwangsik Nho, and Andrew J. Saykin. 2019. “Deep Learning in Alzheimer’s Disease: Diagnostic Classification and Prognostic Prediction Using Neuroimaging Data.” Frontiers in Aging Neuroscience 11 (August): 220. Qiu, Shangran, Prajakta S. Joshi, Matthew I. Miller, Chonghua Xue, Xiao Zhou, Cody Karjadi, Gary H. Chang, et al. 2020. “Development and Validation of an Interpretable Deep Learning Framework for Alzheimer’s Disease Classification.” Brain: A Journal of Neurology 143 (6): 1920–33. Rajkomar, Alvin, Eyal Oren, Kai Chen, Andrew M. Dai, Nissan Hajaj, Michaela Hardt, Peter J. Liu, et al. 2018. “Scalable and Accurate Deep Learning with Electronic Health Records.” NPJ Digital Medicine 1 (May): 18. Tang, Ziqi, Kangway V. Chuang, Charles DeCarli, Lee-Way Jin, Laurel Beckett, Michael J. Keiser, and Brittany N. Dugger. 2019. “Interpretable Classification of Alzheimer’s Disease Pathologies with a Convolutional Neural Network Pipeline.” Nature Communications 10 (1): 2173. Vaswani, Ashish, Noam Shazeer, Niki Parmar, Jakob Uszkoreit, Llion Jones, Aidan N. Gomez, Ł. Ukasz Kaiser, and Illia Polosukhin. 2017. “Attention Is All You Need.” In Advances in Neural Information Processing Systems 30, edited by I. Guyon, U. V. Luxburg, S. Bengio, H. Wallach, R. Fergus, S. Vishwanathan, and R. Garnett, 5998–6008. Curran Associates, Inc. |
| Investigator's Name: | Arina Tagmazian |
| Proposed Analysis: | We propose to develop novel deep machine learning methods to jointly model PET and MRI imaging data, and genetics data, in order to define and analyze associations between biologically and clinically defined Alzheimer’s disease (AD) groups. Application of deep learning methods has resulted in breakthroughs in many fields such as digital histopathology and predictive modeling using electronic medical records (Fu et al. 2020; Rajkomar et al. 2018). Deep learning also shows promise in the analysis of AD pathologies (Jo, Nho, and Saykin 2019; Tang et al. 2019; Qiu et al. 2020). A key factor driving these developments has been the availability of ADNI data to train large-scale deep neural network models. A recent methodological innovation has been the introduction of Transformer models, a class of deep neural network models aimed at analysis of sequential data (Vaswani et al. 2017). Transformers are also well-suited to integrate data with multiple modalities. Two major breakthroughs have been achieved with Transformers: state-of-the-art natural language modeling (GPT-3) (Brown et al. 2020) and protein folding prediction (AlphaFold 2; CASP14 challenge). To our knowledge, no Transformer-based model exists to integrate data modalities relevant to AD with the purpose of delineating amyloid and tau pathologies. NIA-AA research framework groups biomarkers of AD into β amyloid deposition, pathologic tau, and neurodegeneration [AT(N)] (Jack et al. 2018). This biological approach has been suggested to yield three biologically defined groups: Alzheimer continuum (abnormal amyloid regardless of tau status), Alzheimer pathologic change (abnormal amyloid but normal tau), and Alzheimer disease (abnormal amyloid and tau), and these may or may not be related to clinical symptoms (Jack et al. 2018). Biologically defined AD has reported to be more common than clinically defined probable AD, and indicates that some people with biologically defined AD are asymptomatic (Jack et al. 2019). Deep learning techniques could help to clarify associations between biologically and clinically defined AD. In our study we aim to develop and utilize models based on Transformer and convolutional neural networks (CNN) in order to understand how amyloid and tau pathologies arise. In order to do this, we will create deep neural network models to process PET and MRI imaging data, and to predict the biologically defined AD groups. We aim to detect subgroups with different patterns of AD related pathology in terms of AT(N) framework. We will further study differences in neuropsychological functioning, neurodegeneration (hippocampal and cortical atrophy) and polygenic risk score of AD (AD-PRS) between subgroups of different patterns of cortical amyloid and tau pathology. These analyzes have two main aims: 1) to detect cortical amyloid and tau patterns that are related to clinical (episodic memory impairment), neurodegenerative and genetic markers (AD-PRS) of AD; and 2) to predict progression from MCI to AD based on cortical amyloid and tau patterns. Validation of our model predictions would be carried out by comparison to amyloid and tau levels measured in cerebrospinal fluid samples. A key strength of deep neural networks is the ability to perform feature learning i.e. learning higher-level features from raw data such as MRI or PET images. Concordantly, we will aim at maximizing the interpretability of our models in order to be able to extract learnt features and derive explanations for model predictions. Keywords: alzheimer’s, machine learning, deep learning, genetics, genomics, MRI, PET References Brown, Tom B., Benjamin Mann, Nick Ryder, Melanie Subbiah, Jared Kaplan, Prafulla Dhariwal, Arvind Neelakantan, et al. 2020. “Language Models Are Few-Shot Learners.” arXiv [cs.CL]. arXiv. http://arxiv.org/abs/2005.14165. Fu, Yu, Alexander W. Jung, Ramon Viñas Torne, Santiago Gonzalez, Harald Vöhringer, Artem Shmatko, Lucy R. Yates, Mercedes Jimenez-Linan, Luiza Moore, and Moritz Gerstung. 2020. “Pan-Cancer Computational Histopathology Reveals Mutations, Tumor Composition and Prognosis.” Nature Cancer 1 (8): 800–810. Jack, Clifford R., Jr, David A. Bennett, Kaj Blennow, Maria C. Carrillo, Billy Dunn, Samantha Budd Haeberlein, David M. Holtzman, et al. 2018. “NIA-AA Research Framework: Toward a Biological Definition of Alzheimer’s Disease.” Alzheimer’s & Dementia: The Journal of the Alzheimer's Association 14 (4): 535–62. Jack, Clifford R., Jr, Terry M. Therneau, Stephen D. Weigand, Heather J. Wiste, David S. Knopman, Prashanthi Vemuri, Val J. Lowe, et al. 2019. “Prevalence of Biologically vs Clinically Defined Alzheimer Spectrum Entities Using the National Institute on Aging-Alzheimer’s Association Research Framework.” JAMA Neurology, July. https://doi.org/10.1001/jamaneurol.2019.1971. Jo, Taeho, Kwangsik Nho, and Andrew J. Saykin. 2019. “Deep Learning in Alzheimer’s Disease: Diagnostic Classification and Prognostic Prediction Using Neuroimaging Data.” Frontiers in Aging Neuroscience 11 (August): 220. Qiu, Shangran, Prajakta S. Joshi, Matthew I. Miller, Chonghua Xue, Xiao Zhou, Cody Karjadi, Gary H. Chang, et al. 2020. “Development and Validation of an Interpretable Deep Learning Framework for Alzheimer’s Disease Classification.” Brain: A Journal of Neurology 143 (6): 1920–33. Rajkomar, Alvin, Eyal Oren, Kai Chen, Andrew M. Dai, Nissan Hajaj, Michaela Hardt, Peter J. Liu, et al. 2018. “Scalable and Accurate Deep Learning with Electronic Health Records.” NPJ Digital Medicine 1 (May): 18. Tang, Ziqi, Kangway V. Chuang, Charles DeCarli, Lee-Way Jin, Laurel Beckett, Michael J. Keiser, and Brittany N. Dugger. 2019. “Interpretable Classification of Alzheimer’s Disease Pathologies with a Convolutional Neural Network Pipeline.” Nature Communications 10 (1): 2173. Vaswani, Ashish, Noam Shazeer, Niki Parmar, Jakob Uszkoreit, Llion Jones, Aidan N. Gomez, Ł. Ukasz Kaiser, and Illia Polosukhin. 2017. “Attention Is All You Need.” In Advances in Neural Information Processing Systems 30, edited by I. Guyon, U. V. Luxburg, S. Bengio, H. Wallach, R. Fergus, S. Vishwanathan, and R. Garnett, 5998–6008. Curran Associates, Inc. |
