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: | CHAEHYUN LEE |
| Institution: | Gyeongsang National University |
| Department: | Computer Science |
| Country: | |
| Proposed Analysis: | < Project Name > Industry-Academic Joint Technology Development Project < Assignment Name > Diagnosing Alzheimer’s by Analyaing Brain MRI Images Using AI < Necessity of research project > It is known that there are about 50 million Alzheimer's patients around the world. The American Alzheimer's Association announced in its 2020 statistics that one in three elderly people in the United States is dying from Alzheimer's disease or related diseases, and social costs reach 1,000 trillion won. In addition, the number of patients in the United States will reach 13 million by 2050, more than double the current number. The growth rate of dementia patients in Korea is also steep, reaching an annual average of 16%, and the prevalence of dementia is also high, with one in 10 elderly people aged 65 or older suffering. Recently, due to the rapid aging of the population, concerns about dementia are also increasing. However, despite numerous studies, dementia is still a field that needs to be researched and developed to the extent that the cause of the outbreak has not been accurately identified. In particular, a paper that is the basis for the relationship between amyloid beta and Alzheimer's disease, which was recently published 16 years ago, has been embroiled in allegations of manipulation. As such, Alzheimer's is still a field that needs research to the extent that it cannot even clarify the cause of the outbreak. Of course, the cure does not yet exist either. Nevertheless, if there is any research result so far, it is the fact that treatment through early diagnosis is important for dementia, just like cancer. Early detection of dementia is important because early detection can alleviate symptoms and slow the progress. Moreover, as the national dementia management cost in Korea is estimated to increase from 16.5 trillion won per year in 2019 to about 63.1 trillion won in 2040, it is urgent to prepare countermeasures. However, dementia is a disease that is very difficult to diagnose early. Our brain begins to suffer brain atrophy due to neurodegeneration even before cognitive problems occur, and we often miss the "golden time" of diagnosis because we cannot detect it in advance. Brain regression is mainly checked by doctors for brain atrophy in brain MRI, but brain atrophy occurs even during normal aging, so it is difficult to visually check whether there is atrophy compared to age, especially in early patients. The doctor is also a person, so the accuracy of the diagnosis can be reduced. AI can identify dementia by grasping complex brain image data and structures at once, which are not easy to analyze with the human eye. Recently, researchers at Exeter University in the UK also found that about 8% of existing dementia diagnoses were misdiagnosed. In addition, doctors, who are limited in manpower, spend less time on diagnosis, so they can focus on patients' treatment and surgery, which can help a much larger number of patients, increasing overall efficiency and gaining economic benefits. Learning AI through brain MRI images of patients with Alzheimer's disease, a representative degenerative dementia disease, can be helpful in many ways, including technical, economic, and industrial. < The Objectives of the Research Project > The goal of this study is to develop a system that helps early diagnosis of Alzheimer's by learning the differences between the two classes using the brain MRI image dataset of Alzheimer's disease patients and normal people. Brain MRI can determine dementia based on brain atrophy. There are more than 20 types of dementia, and it is possible to identify what kind of dementia you suffer from based on the brain atrophy. Among them, Alzheimer's disease, a degenerative disease suffered by the majority of dementia patients, can be judged based on the atrophy of the hippocampus. In this study task, a classification study on the presence or absence of a disease will be conducted using brain MRI images of Alzheimer's disease patients and normal people based on the degree of atrophy of the hippocampus using these properties. In existing studies that classify images that are good to judge in 2D dimension in advance, it is difficult to prevent information loss obtained from MRI by reducing one dimension. In order to overcome these limitations, this research project will conduct a classification study on the 3D MRI image itself. < Contents of the Research Project > In this study, we will refer to Katabathula, S., Wang, Q. & Xu, R. Predict Alzheimer’s disease using hippocampus MRI data: a lightweight 3D deep convolutional network model with visual and global shape representations. Alz Res Therapy 13, 104 (2021). In this project, we would like to classify AD based on hippocampus magnetic resonance imaging (MRI) segments. We are going to use DenseCNN to conduct research and development. The description of the DenseCNN to be referred to is as follows. DenseCNN has 3 dense layers, with each layer consisting of 2 convolutional layers, combined with Batch normalization (BN) layers and Relu activation layers. Transition layers end with a max pooling layer to decrease the size of input data. DenseCNN has two streams for left and right hippocampus segments correspondingly. Each stream has an initial 3D convolutional layer followed by a BN layer and a Relu activation layer, extracting low-level image features. Then a max pooling was used to ignore 0 voxels on the edges of the input data and reduce the data size. Two dense blocks and a transition layer were stocked in each stream, using 8 and 16 filters correspondingly. At the end of each stream is a global average pooling (GAP) layer, which compresses high dimensional image features to 1-dimensional features. After the GAP layer, two streams were merged followed by a dropout layer. Finally, a fully connected layer and a SoftMax layer were used for generating prediction. The output of the last GAP layer is the CNN features considered here. For each left and right hippocampus, deep visual features were obtained after the last GAP layer of the DenseCNN. |
| Additional Investigators | |
| Investigator's Name: | DABIN OH |
| Proposed Analysis: | < Project Name > Industry-Academic Joint Technology Development Project < Assignment Name > Diagnosing Alzheimer’s by Analyaing Brain MRI Images Using AI < Necessity of research project > It is known that there are about 50 million Alzheimer's patients around the world. The American Alzheimer's Association announced in its 2020 statistics that one in three elderly people in the United States is dying from Alzheimer's disease or related diseases, and social costs reach 1,000 trillion won. In addition, the number of patients in the United States will reach 13 million by 2050, more than double the current number. The growth rate of dementia patients in Korea is also steep, reaching an annual average of 16%, and the prevalence of dementia is also high, with one in 10 elderly people aged 65 or older suffering. Recently, due to the rapid aging of the population, concerns about dementia are also increasing. However, despite numerous studies, dementia is still a field that needs to be researched and developed to the extent that the cause of the outbreak has not been accurately identified. In particular, a paper that is the basis for the relationship between amyloid beta and Alzheimer's disease, which was recently published 16 years ago, has been embroiled in allegations of manipulation. As such, Alzheimer's is still a field that needs research to the extent that it cannot even clarify the cause of the outbreak. Of course, the cure does not yet exist either. Nevertheless, if there is any research result so far, it is the fact that treatment through early diagnosis is important for dementia, just like cancer. Early detection of dementia is important because early detection can alleviate symptoms and slow the progress. Moreover, as the national dementia management cost in Korea is estimated to increase from 16.5 trillion won per year in 2019 to about 63.1 trillion won in 2040, it is urgent to prepare countermeasures. However, dementia is a disease that is very difficult to diagnose early. Our brain begins to suffer brain atrophy due to neurodegeneration even before cognitive problems occur, and we often miss the "golden time" of diagnosis because we cannot detect it in advance. Brain regression is mainly checked by doctors for brain atrophy in brain MRI, but brain atrophy occurs even during normal aging, so it is difficult to visually check whether there is atrophy compared to age, especially in early patients. The doctor is also a person, so the accuracy of the diagnosis can be reduced. AI can identify dementia by grasping complex brain image data and structures at once, which are not easy to analyze with the human eye. Recently, researchers at Exeter University in the UK also found that about 8% of existing dementia diagnoses were misdiagnosed. In addition, doctors, who are limited in manpower, spend less time on diagnosis, so they can focus on patients' treatment and surgery, which can help a much larger number of patients, increasing overall efficiency and gaining economic benefits. Learning AI through brain MRI images of patients with Alzheimer's disease, a representative degenerative dementia disease, can be helpful in many ways, including technical, economic, and industrial. < The Objectives of the Research Project > The goal of this study is to develop a system that helps early diagnosis of Alzheimer's by learning the differences between the two classes using the brain MRI image dataset of Alzheimer's disease patients and normal people. Brain MRI can determine dementia based on brain atrophy. There are more than 20 types of dementia, and it is possible to identify what kind of dementia you suffer from based on the brain atrophy. Among them, Alzheimer's disease, a degenerative disease suffered by the majority of dementia patients, can be judged based on the atrophy of the hippocampus. In this study task, a classification study on the presence or absence of a disease will be conducted using brain MRI images of Alzheimer's disease patients and normal people based on the degree of atrophy of the hippocampus using these properties. In existing studies that classify images that are good to judge in 2D dimension in advance, it is difficult to prevent information loss obtained from MRI by reducing one dimension. In order to overcome these limitations, this research project will conduct a classification study on the 3D MRI image itself. < Contents of the Research Project > In this study, we will refer to Katabathula, S., Wang, Q. & Xu, R. Predict Alzheimer’s disease using hippocampus MRI data: a lightweight 3D deep convolutional network model with visual and global shape representations. Alz Res Therapy 13, 104 (2021). In this project, we would like to classify AD based on hippocampus magnetic resonance imaging (MRI) segments. We are going to use DenseCNN to conduct research and development. The description of the DenseCNN to be referred to is as follows. DenseCNN has 3 dense layers, with each layer consisting of 2 convolutional layers, combined with Batch normalization (BN) layers and Relu activation layers. Transition layers end with a max pooling layer to decrease the size of input data. DenseCNN has two streams for left and right hippocampus segments correspondingly. Each stream has an initial 3D convolutional layer followed by a BN layer and a Relu activation layer, extracting low-level image features. Then a max pooling was used to ignore 0 voxels on the edges of the input data and reduce the data size. Two dense blocks and a transition layer were stocked in each stream, using 8 and 16 filters correspondingly. At the end of each stream is a global average pooling (GAP) layer, which compresses high dimensional image features to 1-dimensional features. After the GAP layer, two streams were merged followed by a dropout layer. Finally, a fully connected layer and a SoftMax layer were used for generating prediction. The output of the last GAP layer is the CNN features considered here. For each left and right hippocampus, deep visual features were obtained after the last GAP layer of the DenseCNN. |
| Investigator's Name: | Dongho Jang |
| Proposed Analysis: | < Project Name > Industry-Academic Joint Technology Development Project < Assignment Name > Diagnosing Alzheimer’s by Analyaing Brain MRI Images Using AI < Necessity of research project > It is known that there are about 50 million Alzheimer's patients around the world. The American Alzheimer's Association announced in its 2020 statistics that one in three elderly people in the United States is dying from Alzheimer's disease or related diseases, and social costs reach 1,000 trillion won. In addition, the number of patients in the United States will reach 13 million by 2050, more than double the current number. The growth rate of dementia patients in Korea is also steep, reaching an annual average of 16%, and the prevalence of dementia is also high, with one in 10 elderly people aged 65 or older suffering. Recently, due to the rapid aging of the population, concerns about dementia are also increasing. However, despite numerous studies, dementia is still a field that needs to be researched and developed to the extent that the cause of the outbreak has not been accurately identified. In particular, a paper that is the basis for the relationship between amyloid beta and Alzheimer's disease, which was recently published 16 years ago, has been embroiled in allegations of manipulation. As such, Alzheimer's is still a field that needs research to the extent that it cannot even clarify the cause of the outbreak. Of course, the cure does not yet exist either. Nevertheless, if there is any research result so far, it is the fact that treatment through early diagnosis is important for dementia, just like cancer. Early detection of dementia is important because early detection can alleviate symptoms and slow the progress. Moreover, as the national dementia management cost in Korea is estimated to increase from 16.5 trillion won per year in 2019 to about 63.1 trillion won in 2040, it is urgent to prepare countermeasures. However, dementia is a disease that is very difficult to diagnose early. Our brain begins to suffer brain atrophy due to neurodegeneration even before cognitive problems occur, and we often miss the "golden time" of diagnosis because we cannot detect it in advance. Brain regression is mainly checked by doctors for brain atrophy in brain MRI, but brain atrophy occurs even during normal aging, so it is difficult to visually check whether there is atrophy compared to age, especially in early patients. The doctor is also a person, so the accuracy of the diagnosis can be reduced. AI can identify dementia by grasping complex brain image data and structures at once, which are not easy to analyze with the human eye. Recently, researchers at Exeter University in the UK also found that about 8% of existing dementia diagnoses were misdiagnosed. In addition, doctors, who are limited in manpower, spend less time on diagnosis, so they can focus on patients' treatment and surgery, which can help a much larger number of patients, increasing overall efficiency and gaining economic benefits. Learning AI through brain MRI images of patients with Alzheimer's disease, a representative degenerative dementia disease, can be helpful in many ways, including technical, economic, and industrial. < The Objectives of the Research Project > The goal of this study is to develop a system that helps early diagnosis of Alzheimer's by learning the differences between the two classes using the brain MRI image dataset of Alzheimer's disease patients and normal people. Brain MRI can determine dementia based on brain atrophy. There are more than 20 types of dementia, and it is possible to identify what kind of dementia you suffer from based on the brain atrophy. Among them, Alzheimer's disease, a degenerative disease suffered by the majority of dementia patients, can be judged based on the atrophy of the hippocampus. In this study task, a classification study on the presence or absence of a disease will be conducted using brain MRI images of Alzheimer's disease patients and normal people based on the degree of atrophy of the hippocampus using these properties. In existing studies that classify images that are good to judge in 2D dimension in advance, it is difficult to prevent information loss obtained from MRI by reducing one dimension. In order to overcome these limitations, this research project will conduct a classification study on the 3D MRI image itself. < Contents of the Research Project > In this study, we will refer to Katabathula, S., Wang, Q. & Xu, R. Predict Alzheimer’s disease using hippocampus MRI data: a lightweight 3D deep convolutional network model with visual and global shape representations. Alz Res Therapy 13, 104 (2021). In this project, we would like to classify AD based on hippocampus magnetic resonance imaging (MRI) segments. We are going to use DenseCNN to conduct research and development. The description of the DenseCNN to be referred to is as follows. DenseCNN has 3 dense layers, with each layer consisting of 2 convolutional layers, combined with Batch normalization (BN) layers and Relu activation layers. Transition layers end with a max pooling layer to decrease the size of input data. DenseCNN has two streams for left and right hippocampus segments correspondingly. Each stream has an initial 3D convolutional layer followed by a BN layer and a Relu activation layer, extracting low-level image features. Then a max pooling was used to ignore 0 voxels on the edges of the input data and reduce the data size. Two dense blocks and a transition layer were stocked in each stream, using 8 and 16 filters correspondingly. At the end of each stream is a global average pooling (GAP) layer, which compresses high dimensional image features to 1-dimensional features. After the GAP layer, two streams were merged followed by a dropout layer. Finally, a fully connected layer and a SoftMax layer were used for generating prediction. The output of the last GAP layer is the CNN features considered here. For each left and right hippocampus, deep visual features were obtained after the last GAP layer of the DenseCNN. |
| Investigator's Name: | BONGSEOK CHA |
| Proposed Analysis: | < Project Name > Industry-Academic Joint Technology Development Project < Assignment Name > Diagnosing Alzheimer’s by Analyaing Brain MRI Images Using AI < Necessity of research project > It is known that there are about 50 million Alzheimer's patients around the world. The American Alzheimer's Association announced in its 2020 statistics that one in three elderly people in the United States is dying from Alzheimer's disease or related diseases, and social costs reach 1,000 trillion won. In addition, the number of patients in the United States will reach 13 million by 2050, more than double the current number. The growth rate of dementia patients in Korea is also steep, reaching an annual average of 16%, and the prevalence of dementia is also high, with one in 10 elderly people aged 65 or older suffering. Recently, due to the rapid aging of the population, concerns about dementia are also increasing. However, despite numerous studies, dementia is still a field that needs to be researched and developed to the extent that the cause of the outbreak has not been accurately identified. In particular, a paper that is the basis for the relationship between amyloid beta and Alzheimer's disease, which was recently published 16 years ago, has been embroiled in allegations of manipulation. As such, Alzheimer's is still a field that needs research to the extent that it cannot even clarify the cause of the outbreak. Of course, the cure does not yet exist either. Nevertheless, if there is any research result so far, it is the fact that treatment through early diagnosis is important for dementia, just like cancer. Early detection of dementia is important because early detection can alleviate symptoms and slow the progress. Moreover, as the national dementia management cost in Korea is estimated to increase from 16.5 trillion won per year in 2019 to about 63.1 trillion won in 2040, it is urgent to prepare countermeasures. However, dementia is a disease that is very difficult to diagnose early. Our brain begins to suffer brain atrophy due to neurodegeneration even before cognitive problems occur, and we often miss the "golden time" of diagnosis because we cannot detect it in advance. Brain regression is mainly checked by doctors for brain atrophy in brain MRI, but brain atrophy occurs even during normal aging, so it is difficult to visually check whether there is atrophy compared to age, especially in early patients. The doctor is also a person, so the accuracy of the diagnosis can be reduced. AI can identify dementia by grasping complex brain image data and structures at once, which are not easy to analyze with the human eye. Recently, researchers at Exeter University in the UK also found that about 8% of existing dementia diagnoses were misdiagnosed. In addition, doctors, who are limited in manpower, spend less time on diagnosis, so they can focus on patients' treatment and surgery, which can help a much larger number of patients, increasing overall efficiency and gaining economic benefits. Learning AI through brain MRI images of patients with Alzheimer's disease, a representative degenerative dementia disease, can be helpful in many ways, including technical, economic, and industrial. < The Objectives of the Research Project > The goal of this study is to develop a system that helps early diagnosis of Alzheimer's by learning the differences between the two classes using the brain MRI image dataset of Alzheimer's disease patients and normal people. Brain MRI can determine dementia based on brain atrophy. There are more than 20 types of dementia, and it is possible to identify what kind of dementia you suffer from based on the brain atrophy. Among them, Alzheimer's disease, a degenerative disease suffered by the majority of dementia patients, can be judged based on the atrophy of the hippocampus. In this study task, a classification study on the presence or absence of a disease will be conducted using brain MRI images of Alzheimer's disease patients and normal people based on the degree of atrophy of the hippocampus using these properties. In existing studies that classify images that are good to judge in 2D dimension in advance, it is difficult to prevent information loss obtained from MRI by reducing one dimension. In order to overcome these limitations, this research project will conduct a classification study on the 3D MRI image itself. < Contents of the Research Project > In this study, we will refer to Katabathula, S., Wang, Q. & Xu, R. Predict Alzheimer’s disease using hippocampus MRI data: a lightweight 3D deep convolutional network model with visual and global shape representations. Alz Res Therapy 13, 104 (2021). In this project, we would like to classify AD based on hippocampus magnetic resonance imaging (MRI) segments. We are going to use DenseCNN to conduct research and development. The description of the DenseCNN to be referred to is as follows. DenseCNN has 3 dense layers, with each layer consisting of 2 convolutional layers, combined with Batch normalization (BN) layers and Relu activation layers. Transition layers end with a max pooling layer to decrease the size of input data. DenseCNN has two streams for left and right hippocampus segments correspondingly. Each stream has an initial 3D convolutional layer followed by a BN layer and a Relu activation layer, extracting low-level image features. Then a max pooling was used to ignore 0 voxels on the edges of the input data and reduce the data size. Two dense blocks and a transition layer were stocked in each stream, using 8 and 16 filters correspondingly. At the end of each stream is a global average pooling (GAP) layer, which compresses high dimensional image features to 1-dimensional features. After the GAP layer, two streams were merged followed by a dropout layer. Finally, a fully connected layer and a SoftMax layer were used for generating prediction. The output of the last GAP layer is the CNN features considered here. For each left and right hippocampus, deep visual features were obtained after the last GAP layer of the DenseCNN. |
| Investigator's Name: | Yugyeong Choi |
| Proposed Analysis: | < Project Name > Industry-Academic Joint Technology Development Project < Assignment Name > Diagnosing Alzheimer’s by Analyaing Brain MRI Images Using AI < Necessity of research project > It is known that there are about 50 million Alzheimer's patients around the world. The American Alzheimer's Association announced in its 2020 statistics that one in three elderly people in the United States is dying from Alzheimer's disease or related diseases, and social costs reach 1,000 trillion won. In addition, the number of patients in the United States will reach 13 million by 2050, more than double the current number. The growth rate of dementia patients in Korea is also steep, reaching an annual average of 16%, and the prevalence of dementia is also high, with one in 10 elderly people aged 65 or older suffering. Recently, due to the rapid aging of the population, concerns about dementia are also increasing. However, despite numerous studies, dementia is still a field that needs to be researched and developed to the extent that the cause of the outbreak has not been accurately identified. In particular, a paper that is the basis for the relationship between amyloid beta and Alzheimer's disease, which was recently published 16 years ago, has been embroiled in allegations of manipulation. As such, Alzheimer's is still a field that needs research to the extent that it cannot even clarify the cause of the outbreak. Of course, the cure does not yet exist either. Nevertheless, if there is any research result so far, it is the fact that treatment through early diagnosis is important for dementia, just like cancer. Early detection of dementia is important because early detection can alleviate symptoms and slow the progress. Moreover, as the national dementia management cost in Korea is estimated to increase from 16.5 trillion won per year in 2019 to about 63.1 trillion won in 2040, it is urgent to prepare countermeasures. However, dementia is a disease that is very difficult to diagnose early. Our brain begins to suffer brain atrophy due to neurodegeneration even before cognitive problems occur, and we often miss the "golden time" of diagnosis because we cannot detect it in advance. Brain regression is mainly checked by doctors for brain atrophy in brain MRI, but brain atrophy occurs even during normal aging, so it is difficult to visually check whether there is atrophy compared to age, especially in early patients. The doctor is also a person, so the accuracy of the diagnosis can be reduced. AI can identify dementia by grasping complex brain image data and structures at once, which are not easy to analyze with the human eye. Recently, researchers at Exeter University in the UK also found that about 8% of existing dementia diagnoses were misdiagnosed. In addition, doctors, who are limited in manpower, spend less time on diagnosis, so they can focus on patients' treatment and surgery, which can help a much larger number of patients, increasing overall efficiency and gaining economic benefits. Learning AI through brain MRI images of patients with Alzheimer's disease, a representative degenerative dementia disease, can be helpful in many ways, including technical, economic, and industrial. < The Objectives of the Research Project > The goal of this study is to develop a system that helps early diagnosis of Alzheimer's by learning the differences between the two classes using the brain MRI image dataset of Alzheimer's disease patients and normal people. Brain MRI can determine dementia based on brain atrophy. There are more than 20 types of dementia, and it is possible to identify what kind of dementia you suffer from based on the brain atrophy. Among them, Alzheimer's disease, a degenerative disease suffered by the majority of dementia patients, can be judged based on the atrophy of the hippocampus. In this study task, a classification study on the presence or absence of a disease will be conducted using brain MRI images of Alzheimer's disease patients and normal people based on the degree of atrophy of the hippocampus using these properties. In existing studies that classify images that are good to judge in 2D dimension in advance, it is difficult to prevent information loss obtained from MRI by reducing one dimension. In order to overcome these limitations, this research project will conduct a classification study on the 3D MRI image itself. < Contents of the Research Project > In this study, we will refer to Katabathula, S., Wang, Q. & Xu, R. Predict Alzheimer’s disease using hippocampus MRI data: a lightweight 3D deep convolutional network model with visual and global shape representations. Alz Res Therapy 13, 104 (2021). In this project, we would like to classify AD based on hippocampus magnetic resonance imaging (MRI) segments. We are going to use DenseCNN to conduct research and development. The description of the DenseCNN to be referred to is as follows. DenseCNN has 3 dense layers, with each layer consisting of 2 convolutional layers, combined with Batch normalization (BN) layers and Relu activation layers. Transition layers end with a max pooling layer to decrease the size of input data. DenseCNN has two streams for left and right hippocampus segments correspondingly. Each stream has an initial 3D convolutional layer followed by a BN layer and a Relu activation layer, extracting low-level image features. Then a max pooling was used to ignore 0 voxels on the edges of the input data and reduce the data size. Two dense blocks and a transition layer were stocked in each stream, using 8 and 16 filters correspondingly. At the end of each stream is a global average pooling (GAP) layer, which compresses high dimensional image features to 1-dimensional features. After the GAP layer, two streams were merged followed by a dropout layer. Finally, a fully connected layer and a SoftMax layer were used for generating prediction. The output of the last GAP layer is the CNN features considered here. For each left and right hippocampus, deep visual features were obtained after the last GAP layer of the DenseCNN. |
