This study investigates the evolution of the AO, NAO, and PNA teleconnection patterns preceding severe convective activity in the eastern and southeastern regions of the United States. Severe reports spanning 1950 to 2016 for the months of March through August were obtained and consolidated into unique severe weather days. Teleconnection time series preceding each severe day were retrieved and sorted based on whether the day itself was non-tornadic, weakly tornadic (EF0 – EF2), or significantly tornadic (EF3 and above). A self-organizing map (SOM) technique was implemented to sort the time series for each teleconnection into nine distinct nodes based on their intrinsic variability. The storm characteristics of the days grouped into each node were analyzed, including storm type, time of day, and time of year. Preliminary results suggest that the analyzed teleconnection patterns exhibit more variability preceding significantly tornadic days, relative to non-tornadic and weakly tornadic days. Furthermore, multiple SOM nodes bear semblance to these significantly tornadic teleconnection time series, suggesting that they have a physical significance, rather than just a statistical one. Based on these findings, six nodes (AO/NAO/PNA, for both the east and southeast) were selected as the most favorable teleconnection patterns for tornadic activity. Utilizing NARR data, composite anomaly plots of CAPE, SRH, low-level vertical shear and other pertinent variables were developed for the days grouped into each of these favorable nodes, providing a dynamic and thermodynamic explanation for the observed storm characteristics.