A graphical time series analysis tool to assess synchrony: Application on neuronal activity of Parkinson’s disease

PATRICIO YANKILEVICH; LUIS RIQUELME; GUSTAVO MURER

Congreso; XVI Congreso Español sobre Tecnologías y Lógica Fuzzy; 2012

In order to better understand Parkinson disease, the electrophysiological changes of neurons in the Substantia Nigra Pars Reticulata (SNpr) on rat experimental models were characterized. Previous work by Tseng et al. demonstrated that periodic discharges occur in a synchronize manner in neurons affected by the disease. Motivated by a problem raised from neurophysiologists about the difficulty to systematically assess neuronal behavior and detect synchrony between firing patterns of different neurons affected by Parkinson disease, we developed a computational tool that makes a qualitative and quantitative assessment of the neural activity signals by integrating classical and novel graphical statistical methods corresponding to the non-linear time series analysis. The analyzed signals were recorded simultaneously in vivo from the SNpr on rat models induced to get Parkinson by 6-OHDA toxin. The graphical statistical methods implemented are: autocorrelation diagram, crossed correlation diagram, inter-spikes intervals (ISI) histogram, spike raster-grams and real time dynamic graphs to compare inter-spikes intervals. In addition, and innovating, we include the variability diagrams, which are a graphical method to represent the existing higher order local correlation among the successive values of a time series. The application of these methods on Parkinson neuronal time series detected an inhibitory synchrony among neurons, a negative correlation pattern where some neurons are inhibited while others exited. This inhibitory synchrony pattern was not clearly identified in previous studies. The resulting software is able to easily detect the existence of underlying patterns (determinism) and synchrony in the neuronal dynamics.