Toolbox for Alzheimer's disease prediction. This toolbox is available online: http://alzpredict.org. 

 

The predicted trajectories are computed based on ubiquitous non-invasive data and do not include imaging (e.g., MRI, PET Subscale (ADAS-COG)  and Clinical Dementia Rating–Sum of Boxes (CDR-SB), in the context of late-onset sporadic Alzheimer’s Disease. The predictions are computed using a Bayesian machine learning strategies. Please cite our paper if you use this model in your publication.  Scale-Cognitive State Examination (MMSE) , blood or CSF-derived variables.) , The Alzheimer’s Disease Assessment Mini-Mental research tool that computes and visualizes individual-level predictions of the trajectories of three clinical scores (as a function of age): easy-to-useThis website provides an

 

The necessary input variables of the prediction tool include:

  1. APOE4 allele count,

  2. APOE2 allele count,

  3. Sex (1: male, 0: female),

  4. Education (in years); 

 

 

Additionally, the user can provide MMSE, ADAS-COG and CDR-SB scores from historical (past longitudinal) visits. There are no restrictions on the amount of past visits that can be input.These observed (past) test scores will need to be input along with the age of the subject at the time of visit. The more historical data you provide, the more accurate is the model’s predictions expected to be.

 

 

Format of Input File:

 

Each input file should correspond to a single person.  Please type all input information in a comma separated variable text file, e.g., named as test_data.txt with following format. First row: APOE4 (allele count: {0, 1, 2}), gender (male=1, female=0), education (years, e.g., 12), APOE2(allele count: {0, 1, 2}) This first row is necessary, and captures all the so-called AD risk factors that our model relies on.  From second row onwards, each line should correspond to a single past visit, where the following variables are listed MMSE, ADAS-COG,CDR-SB, age (years). As in the first line, all numbers should be separated by a comma. All other rows should be similar to the second row, corresponding to different visits The number of past visits can vary for each individual. We assume all three test scores MMSE, ADAS-COG,CDR-SB are available for each visit. E.g., Two Past Visit Records.

 

The test_data.txt file of this testing patient would be formatted as,

 

0,1,16,1

29,7,0,77.5

30,5.5,0,78.1

 

Please upload the data file: test_data.txt using the button upload on the left top of the webpage. After that, the input data will be displayed on the top and the backend prediction algorithm will be called automatically. Please wait for 5 to 10 minutes for the backend algorithm to finish and show the final prediction results. The prediction time horizon ranges from baseline age to 30 years after it. Three figures will be displayed to show the predicted MMSE (red), ADAS-COG (green) and CDR-SB (blue) with standard deviation (shadows and dash line) and mean (bold line). The horizontal axis shows the ages and vertical axis shows the clinical scores. Here are two examples of the predicted clinical scores for a healthy control subject and AD subject.

 

have any further questions, please feel free to contact: yz2377@cornell.edu.yourThis is a preliminary version of our prediction model. We are still improving it. If

 

 

This site was designed with the
.com
website builder. Create your website today.
Start Now