Changes in the Degree of Connectivity Between Multidimensional Characteristics in Longitudinal Studies

• Activity Number: 613
• Type: Contributed
• Date/Time: Thursday, August 2, 2012 : 8:30 AM to 10:20 AM
• Sponsor: Section on Statistical Learning and Data Mining
• Author(s): Arnold Mitnitski*+ and Matthew Richard and Michael Gold and Kenneth Rockwood
• Companies: Dalhousie University and DGI Clinical and Allon Therapeutics Inc. and Dalhousie University
• Keywords: connectivity network ; visualization ; subsampling ; longitudinal studies ; degree distribution ; pair-wise mutual information

Abstract

The multidimensional characterization of complex biomedical systems usually demands a large number of cases in order to obtain reliable inferences. Even so, the number of participants in many studies is relatively small as, for example, in typical clinical trials. Here we suggest an approach based on network visualization, combined with resampling, to discern the patterns of relationships among variables. We illustrate how this can be applied to analyze changes in multiple outcomes in people with dementia. The relationships between 60 variables were represented by connectivity graphs drawn using, for example, pairwise mutual information (PMI). Statistically significant values of the PMI were found using resampling to generate the distribution of the PMIs for each pair of variables. If the null hypothesis (PMI=0) was rejected, the vertices on the graph representing the variables, were connected by the edge. The number of edges (connectivity degree) was used to characterize the staging of the impairment. We found that the degree changed characteristically and showed erosion of connectivity with dementia progression, therefore well summarizes complex information in a single number.