Localizing Synchronized Dipole Sources Using a Modified RAP-MUSIC Algorithm

Synchronization across different brain regions is suggested to be a possible mechanism for functional integration. Analyzing the synchronization of cortical activity is possible if the source activity can be estimated by solving the inverse problem. RAP-MUSIC and R-MUSIC are well-known inverse algorithms that employ an independent topography (IT) source model which allows for synchronous sources. However, these two algorithms have difficulty distinguishing highly correlated sources because of their dependence on a correlation threshold. In this paper we modify RAP-MUSIC to a multi-stage process which analyzes the correlation of candidate sources and searches ITs among pre-correlated groups. This algorithm is not dependent on a subspace correlation threshold to identify synchronous sources and also avoids an exhaustive multidimensional search for such sources among numerous possible combinations. A comparative study was carried out on simulated data, and the results demonstrate superior performance with the modified algorithm compared to the original RAP-MUSIC in recovering synchronous sources. The modified algorithm thus has potential in the study of brain synchronization