Although non-suicidal self-injury (NSSI) is a pernicious and increasingly prevalent behavior, why people start and continue to engage in NSSI still is poorly understood. To elucidate these issues, the present study utilized a sample of 73 undergraduates (33 control; 24 affect dysregulation; 16 NSSI) and employed psychophysiological measures of affect (startle-alone reactivity) and quality of information processing (prepulse inhibition), and experimental methods involving a NSSI-proxy to mimic the NSSI process. Consistent with theory, it was predicted that the NSSI group would display cognitive-affective regulation after the NSSI-proxies whereas the control group would display dysregulation after the NSSI-proxy. Additionally, consistent with theory about initial reinforcement of NSSI, it was predicted that the affect dysregulation group would display dysregulation to the first, but regulation to the second NSSI-proxy. Results supported hypotheses, providing the best evidence yet for why people start and continue to engage in NSSI.