The experiments reported in this thesis studied the mechanisms of expression of overexpectation of conditioned fear, as measured by freezing. In Stage I, rats were conditioned to fear a tone and a flashing light conditioned stimulus (CS) through pairings with a 0.5 mA, 1 s shock. In Stage II, overexpectation was trained by the reinforcement of a compound of these CSs with a shock of the same magnitude. Two compound ?? shock pairings produced an overexpectation effect, as measured by freezing to presentations of the tone alone, while further Stage II training caused over-training of overexpectation. Expression of the overexpectation effect produced by two compound ?? shock pairings could be prevented by pre-test injection of the benzodiazepine partial inverse agonist FG7142. This effect was dose-dependent and not due to state-dependent memory. Control experiments suggested that it was also not due to any general effect of FG7142 on the Pavlovian freezing response. Freezing to a tone that had been conditioned, but not subjected to any decremental training procedures, was unaffected by administration of FG7142 before either the conditioning or test session. FG7142 also did not affect freezing to a tone that had been subjected to an associative blocking procedure. The hypothesis that overexpectation of conditioned fear may be context-dependent was also tested. However, renewal was not observed. Rats that received Stage II training in a context distinct from the Stage I training context showed equivalent expression of overexpectation regardless of whether testing was conducted in the Stage I or Stage II training context. These results are consistent with the hypothesis that overexpectation, like extinction, leads to the imposition of a GABAA receptor-mediated mask on the fear CR. Moreover, they suggest that this masking of fear is the specific consequence of negative predictive error.

The experiments reported in this thesis studied the mechanisms of expression of overexpectation of conditioned fear, as measured by freezing. In Stage I, rats were conditioned to fear a tone and a flashing light conditioned stimulus (CS) through pairings with a 0.5 mA, 1 s shock. In Stage II, overexpectation was trained by the reinforcement of a compound of these CSs with a shock of the same magnitude. Two compound ?? shock pairings produced an overexpectation effect, as measured by freezing to presentations of the tone alone, while further Stage II training caused over-training of overexpectation. Expression of the overexpectation effect produced by two compound ?? shock pairings could be prevented by pre-test injection of the benzodiazepine partial inverse agonist FG7142. This effect was dose-dependent and not due to state-dependent memory. Control experiments suggested that it was also not due to any general effect of FG7142 on the Pavlovian freezing response. Freezing to a tone that had been conditioned, but not subjected to any decremental training procedures, was unaffected by administration of FG7142 before either the conditioning or test session. FG7142 also did not affect freezing to a tone that had been subjected to an associative blocking procedure. The hypothesis that overexpectation of conditioned fear may be context-dependent was also tested. However, renewal was not observed. Rats that received Stage II training in a context distinct from the Stage I training context showed equivalent expression of overexpectation regardless of whether testing was conducted in the Stage I or Stage II training context. These results are consistent with the hypothesis that overexpectation, like extinction, leads to the imposition of a GABAA receptor-mediated mask on the fear CR. Moreover, they suggest that this masking of fear is the specific consequence of negative predictive error.