Neuroscientists Xu L, Wu X, et al. at the College of Information Science and Technology, Beijing Normal University, Beijing, China provide evidence that discrimination between patients with Alzheimer’s disease (AD) and mild cognitive impairment (MCI) can be improved via extending a multi-modality algorithm associated with multi-modality use.

In their study published in the journal: Comput Methods Programs Biomed, 2015, a well-developed method in pattern recognition and machine learning, termed ‘sparse representation-based classification (SRC) was applied to a multi-modality classification framework named the ‘weighted multi-modality SRC (wmSRC).  The data from the latter system includes three modalities of volumetric magnetic resonance imaging (MRI), fluorodeoxyglucose (FDG) positron emission tomography (PET) and florbetapir PET previously used to classify AD/MCI patients.

What was found was that using this expanded algorithm an accuracy of 94.8 was achieved for AD versus controls, 74.5% for MCI versus controls, and 77.8 percent for progressive MCI versus stable MCI.  These values are reported to be superior or comparable with results of other current models used in other recent multi-modality research.

The complete article is available at:  Elsevier Ireland Ltd. And Comput Methods Programs Biomed. 2015 Aug 10. pii: S0169-2607(15)00201-1. doi: 10.1016/j.cmpb.2015.08.004. [Epub ahead of print]