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: | Sophia Frangou |
Institution: | Icahn School of Medicine at Mount Sinai |
Department: | Psychiatry |
Country: | |
Proposed Analysis: | Background: Neuropsychiatric disorders show significant inter-individual variability that is poorly understood but may reflect differences in disease stage. Machine learning approaches are increasingly being used to derive neurosignatures with diagnostic and prognostic value for neuropsychiatric disorders. Deep-learning methods have recently been used to identify distinct disease subtypes that may correspond to different disease progression patterns. The aim of this study is to conduct a comparative evaluation of three widely used algorithms that have been proposed for the detection of disease stages based on neuroimaging data, namely the Smile-GAN (SeMI-supervised cLustEring via Generative Adversarial Network) and GANCMLAE(Generative Adversarial Network constrained Multiple loss autoencoder), which are both semi-supervised deep-learning algorithms, and SuStaIn (Subtype and Stage Inference), a supervised clustering and pathway modeling algorithm. Performance of these algorithms will be compared in terms of accuracy and interpretability using structural neuroimaging data from the ADNI dataset and simulated data. Clinical and cognitive data from the ADNI dataset will be used for external validation of the disease stages identified by each of the three algorithms. |
Additional Investigators | |
Investigator's Name: | Shalaila Haas |
Proposed Analysis: | Background: Neuropsychiatric disorders show significant inter-individual variability that is poorly understood but may reflect differences in disease stage. Machine learning approaches are increasingly being used to derive neurosignatures with diagnostic and prognostic value for neuropsychiatric disorders. Deep-learning methods have recently been used to identify distinct disease subtypes that may correspond to different disease progression patterns. The aim of this study is to conduct a comparative evaluation of three widely used algorithms that have been proposed for the detection of disease stages based on neuroimaging data, namely the Smile-GAN (SeMI-supervised cLustEring via Generative Adversarial Network) and GANCMLAE(Generative Adversarial Network constrained Multiple loss autoencoder), which are both semi-supervised deep-learning algorithms, and SuStaIn (Subtype and Stage Inference), a supervised clustering and pathway modeling algorithm. Performance of these algorithms will be compared in terms of accuracy and interpretability using structural neuroimaging data from the ADNI dataset and simulated data. Clinical and cognitive data from the ADNI dataset will be used for external validation of the disease stages identified by each of the three algorithms. |
Investigator's Name: | Ruiyang Ge |
Proposed Analysis: | Background: Neuropsychiatric disorders show significant inter-individual variability that is poorly understood but may reflect differences in disease stage. Machine learning approaches are increasingly being used to derive neurosignatures with diagnostic and prognostic value for neuropsychiatric disorders. Deep-learning methods have recently been used to identify distinct disease subtypes that may correspond to different disease progression patterns. The aim of this study is to conduct a comparative evaluation of three widely used algorithms that have been proposed for the detection of disease stages based on neuroimaging data, namely the Smile-GAN (SeMI-supervised cLustEring via Generative Adversarial Network) and GANCMLAE(Generative Adversarial Network constrained Multiple loss autoencoder), which are both semi-supervised deep-learning algorithms, and SuStaIn (Subtype and Stage Inference), a supervised clustering and pathway modeling algorithm. Performance of these algorithms will be compared in terms of accuracy and interpretability using structural neuroimaging data from the ADNI dataset and simulated data. Clinical and cognitive data from the ADNI dataset will be used for external validation of the disease stages identified by each of the three algorithms. |
Investigator's Name: | Yuetong Yu |
Proposed Analysis: | Background: Neuropsychiatric disorders show significant inter-individual variability that is poorly understood but may reflect differences in disease stage. Machine learning approaches are increasingly being used to derive neurosignatures with diagnostic and prognostic value for neuropsychiatric disorders. Deep-learning methods have recently been used to identify distinct disease subtypes that may correspond to different disease progression patterns. The aim of this study is to conduct a comparative evaluation of three widely used algorithms that have been proposed for the detection of disease stages based on neuroimaging data, namely the Smile-GAN (SeMI-supervised cLustEring via Generative Adversarial Network) and GANCMLAE(Generative Adversarial Network constrained Multiple loss autoencoder), which are both semi-supervised deep-learning algorithms, and SuStaIn (Subtype and Stage Inference), a supervised clustering and pathway modeling algorithm. Performance of these algorithms will be compared in terms of accuracy and interpretability using structural neuroimaging data from the ADNI dataset and simulated data. Clinical and cognitive data from the ADNI dataset will be used for external validation of the disease stages identified by each of the three algorithms. |
Investigator's Name: | Kevin Yu |
Proposed Analysis: | Background: Neuropsychiatric disorders show significant inter-individual variability that is poorly understood but may reflect differences in disease stage. Machine learning approaches are increasingly being used to derive neurosignatures with diagnostic and prognostic value for neuropsychiatric disorders. Deep-learning methods have recently been used to identify distinct disease subtypes that may correspond to different disease progression patterns. The aim of this study is to conduct a comparative evaluation of three widely used algorithms that have been proposed for the detection of disease stages based on neuroimaging data, namely the Smile-GAN (SeMI-supervised cLustEring via Generative Adversarial Network) and GANCMLAE(Generative Adversarial Network constrained Multiple loss autoencoder), which are both semi-supervised deep-learning algorithms, and SuStaIn (Subtype and Stage Inference), a supervised clustering and pathway modeling algorithm. Performance of these algorithms will be compared in terms of accuracy and interpretability using structural neuroimaging data from the ADNI dataset and simulated data. Clinical and cognitive data from the ADNI dataset will be used for external validation of the disease stages identified by each of the three algorithms. |
Investigator's Name: | Nicole Sanford |
Proposed Analysis: | Background: Neuropsychiatric disorders show significant inter-individual variability that is poorly understood but may reflect differences in disease stage. Machine learning approaches are increasingly being used to derive neurosignatures with diagnostic and prognostic value for neuropsychiatric disorders. Deep-learning methods have recently been used to identify distinct disease subtypes that may correspond to different disease progression patterns. The aim of this study is to conduct a comparative evaluation of three widely used algorithms that have been proposed for the detection of disease stages based on neuroimaging data, namely the Smile-GAN (SeMI-supervised cLustEring via Generative Adversarial Network) and GANCMLAE(Generative Adversarial Network constrained Multiple loss autoencoder), which are both semi-supervised deep-learning algorithms, and SuStaIn (Subtype and Stage Inference), a supervised clustering and pathway modeling algorithm. Performance of these algorithms will be compared in terms of accuracy and interpretability using structural neuroimaging data from the ADNI dataset and simulated data. Clinical and cognitive data from the ADNI dataset will be used for external validation of the disease stages identified by each of the three algorithms. |