The paper introduces a methodology designed for the investigation of rock matrix heterogeneities and their effect on pre- and post- fracture behavior. Specifically, a grain edge interaction length distribution (EILD) is constructed by Cathodoluminescnece image analysis. The EILD augments traditional Discrete Element Method (DEM) models by stochastically strengthening or weakening bonds, which simulates the presence of defects and locally tough regions. These heterogeneities cause the development of an intrinsic process zone (IPZ), which is a material property that is experimentally observable by Acoustic Emissions (AE). This paper compares the development of the IPZ within numerical and experimental three point bending tests. Similar to experimental observations, EILD-augmented DEM three point bending tests yield IPZs with variable widths. In comparison, the traditional DEM is unable to generate an IPZ. The paper concludes that the physically informed EILD contains the necessary physical distribution of grain contacts to augment DEM rock fracture models. Further analysis of the numerical AE activity reveals that larger AE events are located directly along the rupture and they are linearly related to their number of constituent interactions. As such, an AE interaction count threshold is identified to distinguish between rupture and damage AE activity. These results demonstrate the ability of the presented augmented DEM model to investigate the rock volumes associated with large rupture events for various levels of heterogeneity.