Political thought and discussion are key to shaping the course of society (Conover, Searing & Crewe, 2002), and many fields of study have sought to explain these interactions. Past research suggests that individuals with homogenous political thought are more likely to engage in discussions with each other, (Huckfeldt &Sprague, 1995; Mutz, 2006), and evidence also supports emotional processing in such discussions (Knutson, Wood, & Spampinato, 2006). Much of this work, however, has relied on the limited approach of survey-based research which cannot indicate the specific cognitive and brain mechanisms involved in political discussion (McPherson & Smith-Lovin, 1987), especially regarding these discussions in heterogeneous settings (Eliasoph, 1998; Festinger, 1957; Walsh, 2004). This study assessed willingness to discuss politics in an experimental setting, and operationalized similarity/difference in two different ways: partisanship and issue position. These methods allowed for the examination of willingness of a person to talk to individuals who are completely like themselves (i.e. same party, same stance), different from themselves (i.e. different party, different stance), or both similar and different (i.e. different party, same stance; same party, different stance). Social neuroscience has used the Feedback-Related Negativity (FRN) Event-Related Potential (ERP) component to study social conflict. Based on this research, it was hypothesized that robust FRN responses would differentiate between three possible models of social conflict as involved in political discussion (i.e., the reinforcement learning hypothesis, the motivational significance hypothesis, or the expectation violation hypothesis). This presentation will focus on the extent to which the present findings provide insight into the fundamental brain mechanisms underlying political discord, and the extent that these findings can expand the current thinking in both political science and social neuroscience using ERP-technology.