Contingency discrimination training and resurgence: Effects of nonzero alternative reinforcement rates during “off” sessions

Abstract

Resurgence is an increase in a previously reduced target behavior with worsening conditions for a more recently reinforced alternative behavior. Alternating sessions in which alternative reinforcement is available (i.e., “on” sessions) versus extinction (i.e., “off” sessions) during treatment (i.e., contingency discrimination training [CDT]) reduces resurgence. This experiment examined whether CDT could mitigate resurgence without complete extinction during “off” sessions. Target lever pressing of rats was reinforced in baseline and then extinguished during treatment. In treatment, all groups experienced “on” sessions with high-rate alternative reinforcement delivered on a variable-interval (VI) 10-s schedule. “Off” sessions arranged either extinction, the same (VI 10 s), or lower rates of alternative reinforcement (VI 12.5 s, VI 16.6 s, VI 25 s, or VI 50 s). CDT eliminated resurgence only when extinction was arranged during “off” sessions, but resurgence was reduced when the VI 50 s was arranged. Resurgence as choice in context theory described the data well, assuming that the biasing effects of discriminating the relevant contingencies are a function of the ratio of alternative reinforcement rates for “off” versus “on” sessions. These findings suggest CDT can reduce resurgence without requiring extinction of the alternative response if alternating reinforcement rates differ enough across sessions.