Blocking, Unblocking, and Overexpectation of Fear: A Role for Opioid Receptors in the Regulation of Pavlovian Association Formation Gavan P. McNally, Michael Pigg, and Gabrielle Weidemann University of New South Wales Injection of the opioid receptor antagonist naloxone facilitated acquisition of fear to contextual and auditory conditioned stimuli (CSs) in Experiments 1A and 1B. Experiment 2 showed that prior conditioning to a distinctive context blocked conditioning to an auditory CS. Blocking of CS fear was prevented by administrations of naloxone or increases in footshock intensity. Blocking of CS fear was facilitated by decreases in footshock intensity in a naloxone-reversible manner. Experiment 3 showed that compound conditioning of two CSs, each previously and separately paired with shock, produced overexpectation of fear that was reversed by naloxone. These results are consistent with a role for opioid receptors controlling Pavlovian association formation by regulating the discrepancy ( -V) described by R. A. Rescorla and A. R. Wagner (1972). Placed in a novel context (a distinctive chamber) and exposed to pairings of a discrete conditioned stimulus (CS) with brief but aversive footshock (unconditioned stimulus [US]), rats learn about this relation between the context, the CS, and US. They exhibit this learning when reexposed to either the CS or context in a diverse but correlated range of behavioral, endocrine, and autonomic re- actions that include freezing, hypoalgesia, potentiated startle, cor- ticosteroid release, and increased arterial blood pressure (Davis, 1992; Fanselow & LeDoux, 1999; LeDoux, 2000; Maren, 2001). The contents of this learning are commonly viewed in terms of the formation of excitatory associations between a representation of the context, the CS, and a fear motivational system aroused by the aversive US (Konorski, 1967; but see Bolles & Fanselow, 1980). Considerable evidence supports the view that these excitatory Pavlovian associations are mediated by glutamatergic neurotrans- mission in the amygdala. Specifically, this evidence suggests that activation of N-methyl-D-aspartate (NMDA) receptors in the amygdala basolateral nucleus (ABl) detects the CS–US conjunc- tion. This activation, in turn, initiates a variety of signal transduc- tion cascades (e.g., Ca 2+ and cyclic AMP-dependent signaling) to result in synaptic plasticity and the long-term representation of the CS–US association (for reviews, see Fanselow & LeDoux, 1999; Maren, 1996, 2001). Among the evidence consistent with this view are lesion data showing that ABl destruction abolishes fear con- ditioning (Campeau & Davis, 1995; Goosens & Maren, 2001; Sananes & Davis, 1992), neuropharmacological data showing that infusions of NMDA receptor antagonists prevent the acquisition of fear conditioning (e.g., Campeau, Miserendino, & Davis, 1992; Kim, DeCola, Landeira-Fernandez, & Fanselow, 1991; Miseren- dino, Sananes, Melia, & Davis, 1990), and electrophysiological data showing that fear conditioning produces synaptic plasticity in the ABl that is correlated with conditioned freezing (e.g., Rogan, Staubli, & LeDoux, 1997). Moreover, this NMDA-receptor- dependent plasticity is often further related to Hebb’s (1949) learning rule, which states that the increment in learning on a conditioning trial is a function of a constant, activity in a presyn- aptic neuron, and activity in a postsynaptic neuron. A difficulty faced when attempting to relate these neurobiolog- ical findings to associative learning theory is that learning does not proceed according to the simple CS–US temporal contiguity re- quired for NMDA receptor activation. Temporally contiguous presentations of the CS and US can cause increments; no change; or, indeed, decrements in conditioned fear depending on the asso- ciative status of other conditioned stimuli present during those presentations. For example, Kamin (1968) subjected rats to pair- ings of a light CS and a footshock US. In Stage 2, rats received a compound CS of light + noise followed by the footshock. Kamin’s seminal finding was that prior conditioning of the light CS blocked conditioning from accruing to the noise CS, despite the fact that the rats received numerous noise–shock pairings in compound with the light. Kamin (1968, 1969) interpreted these findings to mean that conditioning proceeds to the extent to which the US is surprising or unexpected. The rats failed to learn about the added element (the noise CS) in Stage 2 because it was redundant or noninformative with respect to US occurrence. Rescorla (1970) has reported a similarly striking failure of excitatory conditioning using a variant of this blocking design. In Stage 1, Rescorla (1970) trained rats to fear a light CS and a tone CS by pairing them separately with shock. In Stage 2, rats in the experimental group received compound presentations of the light and tone CS with shock, whereas rats in the control groups received either additional tone–shock pairings or no additional training. Rescorla’s important observation was that the Stage 2 compound training with the two CSs reduced the amount of fear provoked by either CS alone on a subsequent test. That is, there was overexpectation of fear in Stage 2, prompting a reduction in associative strength to each CS that Gavan P. McNally, Michael Pigg, and Gabrielle Weidemann, School of Psychology, University of New South Wales, Sydney, New South Wales, Australia. These experiments were supported by Grant DP0343808 from the Australian Research Council. We thank Fred Westbrook for helpful discussions. Correspondence concerning this article should be addressed to Gavan P. McNally, School of Psychology, University of New South Wales, Sydney 2052, NSW Australia. E-mail: g.mcnally@unsw.edu.au Behavioral Neuroscience Copyright 2004 by the American Psychological Association, Inc. 2004, Vol. 118, No. 1, 111–120 0735-7044/04/$12.00 DOI: 10.1037/0735-7044.118.1.111 111