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: | Taravat Ghafourian |
| Institution: | Nova Southeastern University |
| Department: | College of Pharmacy |
| Country: | |
| Proposed Analysis: | At Nova Southeastern University, we are interested in modeling the progression of Alzheimer’s disease using longitudinal data and identifying patient factors that may contribute to the course of disease progression by using machine learning methods. Disease progression in this dataset has been tracked using various neuroimaging methods, CSF proteins, synaptic biomarkers, and cognitive test scores such as ADAS-cog and CDR-SOB. Previous literature has reported computational models and data analysis to identify and/or validate disease progression scores from various biomarkers (Jedynak et al 2015, Beheshti et al 2022, Galasko et al 2019). Our aim is to merge the available datasets and pursue two objectives: 1. To develop new models of Alzheimer’s disease progression by using (variants of) decision trees algorithms and ensembles; this is different from previous models that assume constant disease progression rates (e.g. Jedynak et al 2015) or models that use logistic curves to describe the non-linearity in data (Samtani et al 2012, 2013, Delor et al 2013). The application of decision tree algorithms here is advantageous as they do not postulate any particular type of data distribution or model structure, with the ability to cope with complex relationships between the features and the output variables. 2. To investigate the effect of various patient factors such as lifestyle attributes, existing comorbidities (and biochemical features of comorbidities such as inflammation biomarkers), medications used for these comorbidities, and pharmacological interventions for Alzheimer’s disease on the disease progression. Furthermore, we will also investigate the impact of the interaction of these ‘patient factors’ with the available genotype to derive a model that can predict the disease progression rate. We will apply various feature selection algorithms to investigate the role of these “patient factors” on the disease progression rate, measured by various clinical, biochemical, and imaging methods. When developing computational methods, we will use already identified biomarkers of the disease, such as those developed by Beheshti et al (2022) and Galasko et al (2019), as well as the established clinical measures of cognitive decline from this dataset. In terms of machine learning and feature selection methods, in addition to decision trees, we will use a number of linear and other non-linear algorithms and compare the accuracy of the resulting models with the existing literature models. References: Beheshti et al 2022, J Alzheimers Dis. 89, 1493-1502, https://doi.org/10.3233/JAD-220585 Delor et al 2013, CPT: Pharmacometrics & Systems Pharmacology, 2: 1-10 78, https://doi.org/10.1038/psp.2013.54 Galasko et al 2019, Alzheimers Dement (N Y) 5, 871-882. https://doi.org/10.1016/j.trci.2019.11.002 Jedynak et al 2015, Neurobiology of Aging, Suppl 1, Pages S178-S184, https://doi.org/10.1016/j.neurobiolaging.2014.03.043 Samtani et al 2012, J Clin Pharmacol. 52, 629-44, https://doi.org/10.1177/0091270011405497 Samtani et al 2013, Br. J. Clin. Pharmacol. 75, 146–161, https://doi.org/10.1111/j.1365-2125.2012.04308.x |
| Additional Investigators |
