{"id":6328,"date":"2015-10-08T05:36:03","date_gmt":"2015-10-08T05:36:03","guid":{"rendered":"http:\/\/revoscience.com\/en\/?p=6328"},"modified":"2015-10-08T05:36:03","modified_gmt":"2015-10-08T05:36:03","slug":"predicting-change-in-the-alzheimers-brain","status":"publish","type":"post","link":"https:\/\/www.revoscience.com\/en\/predicting-change-in-the-alzheimers-brain\/","title":{"rendered":"Predicting change in the Alzheimer\u2019s brain"},"content":{"rendered":"

Combining MRI and other data helps machine-learning systems predict effects of neurodegenerative disease.<\/strong><\/em><\/span><\/p>\n

\"MIT-Predicting_0\"<\/a>CAMBRIDGE, Mass.<\/strong> —\u00a0MIT researchers are developing a computer system that uses genetic, demographic, and clinical data to help predict the effects of disease on brain anatomy.<\/span><\/p>\n

In experiments, they trained a machine-learning system on MRI data from patients with neurodegenerative diseases and found that supplementing that training with other patient information improved the system\u2019s predictions. In the cases of patients with drastic changes in brain anatomy, the additional data cut the predictions\u2019 error rate in half, from 20 percent to 10 percent.<\/span><\/p>\n

\u201cThis is the first paper that we\u2019ve ever written on this,\u201d says Polina Golland, a professor of electrical engineering and computer science at MIT and the senior author on the new paper. \u201cOur goal is not to prove that our model is the best model to do this kind of thing; it\u2019s to prove that the information is actually in the data. So what we\u2019ve done is, we take our model, and we turn off the genetic information and the demographic and clinical information, and we see that with combined information, we can predict anatomical changes better.\u201d<\/span><\/p>\n

[pullquote]The researchers\u2019 first step is to produce a generic brain template by averaging the voxel values of hundreds of randomly selected MRI scans.[\/pullquote]<\/p>\n

First author on the paper is Adrian Dalca, an MIT graduate student in electrical engineering and computer science and a member of Golland\u2019s group at MIT\u2019s Computer Science and Artificial Intelligence Laboratory. They\u2019re joined by Ramesh Sridharan, another PhD student in Golland\u2019s group, and by Mert Sabuncu, an assistant professor of radiology at Massachusetts General Hospital, who was a postdoc in Golland\u2019s group.<\/span><\/p>\n

The researchers are presenting the paper at the International Conference on Medical Image Computing and Computer Assisted Intervention this week. The work is a project of the Neuroimage Analysis Center, which is based at Brigham and Women\u2019s Hospital in Boston and funded by the National Institutes of Health.<\/span><\/p>\n

Common denominator<\/strong><\/span><\/p>\n

In their experiments, the researchers used data from the Alzheimer\u2019s Disease Neuroimaging Initiative, a longitudinal study on neurodegenerative disease that includes MRI scans of the same subjects taken months and years apart.<\/span><\/p>\n

Each scan is represented as a three-dimensional model consisting of millions of tiny cubes, or \u201cvoxels,\u201d the 3-D equivalent of image pixels.<\/span><\/p>\n

The researchers\u2019 first step is to produce a generic brain template by averaging the voxel values of hundreds of randomly selected MRI scans. They then characterize each scan in the training set for their machine-learning algorithm as a deformation of the template. Each subject in the training set is represented by two scans, taken between six months and seven years apart.<\/span><\/p>\n

The researchers conducted two experiments: one in which they trained their system on scans of both healthy subjects and those displaying evidence of either Alzheimer\u2019s disease or mild cognitive impairment, and one in which they trained it only on data from healthy subjects.<\/span><\/p>\n

In the first experiment, they trained the system twice, once using just the MRI scans and the second time supplementing them with additional information. This included data on genetic markers known as single-nucleotide polymorphisms; demographic data, such as subject age, gender, marital status, and education level; and rudimentary clinical data, such as patients\u2019 scores on various cognitive tests.<\/span><\/p>\n

The brains of healthy subjects and subjects in the early stages of neurodegenerative disease change little over time, and indeed, in cases where the differences between a subject\u2019s scans were slight, the system trained only on MRI data fared well. In cases where the changes were more marked, however, the addition of the supplementary data made a significant difference.<\/span><\/p>\n

Counterfactuals<\/strong><\/span><\/p>\n

In the second experiment, the researchers trained the system just once, on both the MRI data and the supplementary data of healthy subjects. But they instead used it to predict what the brains of Alzheimer\u2019s patients would have looked like had they not been disfigured by disease.<\/span><\/p>\n

In this case, there are no clinical data that could validate the system\u2019s predictions. But the researchers believe that exploring this sort of counterfactual could be scientifically useful.<\/span><\/p>\n

\u201cIt would illuminate how changes in individual subjects \u2014 for example, with mild cognitive impairment, which is a precursor to Alzheimer\u2019s \u2014 evolve along this trajectory of degeneration, as compared to what normal degeneration would be,\u201d Golland says. \u201cWe think that there are very interesting research applications of this. But I have to be honest and say that the original motivation was curiosity about how much of anatomy we could predict from genetics and other non-image data.\u201d<\/span><\/p>\n