Introduction. Regional ischemia resulting from coronary occlusion is widely known to be arrhythmogenic; however, the mechanisms are not well understood. The goal of this study is to examine the role of the ischemic border zone (BZ) in the mechanisms that lead to establishment of reentry in regional ischemia. Methods. A novel geometrically accurate finite element model of a cross section of the rabbit ventricles that incorporates a zone of ischemia (see figure) is used. Ischemic zone geometry mimics LAD occlusion and incorporates a central ischemic zone (CIZ) and BZs (epi/endocardial and lateral) with a linear transition to the surrounding normoxic tissue. Ischemia results from hypoxia, acidosis, and hyperkalemia (in figure, black to white color represents change in [K⁺]_o from 5.4 to 15mM) of varying degrees (0-10min of occlusion). The width of the lateral BZ is 5.5mm, while the epi/endocardial BZ is varied (0.25-0.5mm). The model is paced (see figure) at 175-300ms cycle lengths (CLs). A premature stimulus is applied in the LV 155ms following the last paced beat in an attempt to initiate reentry. Results. Reentry occurred only when CIZ was severely ischemic; propagation failed in the CIZ (site A) as well as in the lateral BZ. Conduction continued in the epi/endocardial BZs, however, it exhibited alternans (site B). The alternans increased with decreasing CL and decreasing epi/endocardial BZ width. At CLs below 200ms and BZ widths below 0.5mm, propagation failed in LV and septum, and reentered through the RV epicardial BZ (see arrows). Further decrease in epi/endocardial BZ width led to conduction failure eliminating the reentrant pathway. Conclusions. This study provides insight into the arrhythmogenic effects of the ischemic BZ width for various degrees of ischemia severity in CIZ. It demonstrates that formation of a reentry around the ischemic zone is preceded by alternans in the epi/endocardial BZ.