Psychopharmacology (2001) 156:92–97 DOI 10.1007/s002130100757
O R I G I N A L I N V E S T I G AT I O N
Justin A. Harris · R. Frederick Westbrook
Contextual control over the expression of fear in rats conditioned under a benzodiazepine
Received: 28 October 2000 / Accepted: 24 February 2001 / Published online: 28 April 2001 © Springer-Verlag 2001
Abstract Rationale: Benzodiazepines disrupt fear conditioning, but this disruption is context-specific; if rats have been conditioned under a benzodiazepine, their fear is recovered if they are tested in a different context. The present experiments investigated how the conditioning context controls fear in rats conditioned under a benzodiazepine. Objectives: The experiments had three aims: (1) to replicate the finding that fear is recovered when rats are tested in a different context, (2) to test whether the conditioning context reduces fear generally or only for the specific stimulus conditioned in that context and (3) to test whether latent inhibition of the conditioning context reduces its control over fear. Methods: Rats were injected with the benzodiazepine midazolam (1.25 mg/kg) or saline and exposed to a conditioned stimulus (CS) and shock in a distinctive chamber. Latent inhibition of the chamber was induced by extensively preexposing the rats to the chamber. The day after conditioning, fear was assessed by presenting the CS while rats were in either the conditioning chamber or a different chamber. Results: The midazolam-induced reduction of fear was reversed (i.e. fear was partially recovered) if rats were tested in the different context, and was completely prevented if the conditioning context had been latently inhibited. These two effects were not additive since, when the conditioning context had been latently inhibited, rats showed less fear in the different context than in the conditioning context. Conclusions: We argue that midazolam does not disrupt conditioning, but imbues the conditioning context with control over retrieval of the CSshock association. In this regard, the effects of midazolam closely parallel those of extinction. Keywords Midazolam · Amnesia · Freezing · Extinction · Latent inhibition · Rat J.A. Harris · R.F. Westbrook (✉) School of Psychology, The University of New South Wales, Sydney 2052, Australia e-mail:
[email protected] Tel.: +61-2-93853033, Fax: +61-2-93853641
Introduction In addition to their anxiolytic and hypnotic effects, benzodiazepines are renowned for their capacity to interfere with new learning. This “amnestic” effect of benzodiazepines has been reported in studies with both animal (Thiébot 1985; Venault et al. 1986) and human subjects (Lister and File 1984; Polster et al. 1993). For example, compared to rats injected with saline, rats injected with a benzodiazepine before being exposed to pairing of a conditioned stimulus (CS) and a noxious unconditioned stimulus (US) subsequently show weaker fear responses (freezing, analgesia, passive avoidance, suppression of drinking) in the presence of that CS (Cole and Jones 1995; Jensen et al. 1979; Oishi et al. 1972; Scobie and Garske 1970; Westbrook et al. 1991). Thus, it has been argued that this effect of benzodiazepines results from an inhibition of the neural processes that mediate fear conditioning (Davis et al. 1994). There have been studies in which benzodiazepines did not prevent fear conditioning (Davis 1979; Tenen 1967). Recently, we noted a procedural difference between these studies and those in which benzodiazepines were shown to impair fear conditioning (Harris and Westbrook 1999). Specifically, the studies that failed to find evidence for an effect of benzodiazepines trained rats with a CS and shock in one chamber but tested them in a different chamber (Davis 1979; Tenen 1967). In contrast, studies using similar experimental paradigms but which trained and tested rats in the same chamber did report an effect of benzodiazepines on fear conditioning (Feldon and Weiner 1989; Scobie and Garske 1970). This suggests that the effect of benzodiazepines on fear conditioning may be specific to the environmental context in which the rats are conditioned. We have obtained evidence that supports this suggestion (Harris and Westbrook 1999). We injected rats with the benzodiazepine midazolam before exposing them to pairing of a CS and shock in a particular chamber. When tested the next day, the rats showed little fear of the CS if it was presented in the shock chamber, but they showed significantly more fear if the CS was presented in a different chamber.
93
Our findings indicate that benzodiazepines do not prevent rats from learning CS–US associations, rather the drugs regulate where that learning is subsequently expressed. The present experiments further investigate the role of the context in regulating the expression of midazolam's effect on fear conditioning. Experiment 1 aimed to replicate our previous demonstration that the effect of midazolam on fear conditioning is context-specific, and to examine which aspect of the CS–US association is controlled by the conditioning context. Rats were conditioned to a clicker CS in one context and to a light CS in another context; the rats were injected with midazolam prior to both conditioning sessions. They were then tested for fear of one of the CSs either in the context where it had been conditioned (Same) or in the context where the other CS had been conditioned (Different). If the context where rats are conditioned under midazolam acts to inhibit retrieval of the shock memory, then the rats should show equally low levels of fear to either CS in either context. However, if the rats show less fear when the CS is presented in its conditioning context than in the other context, this would indicate that the locus of contextual control involves the representation of the CS. Experiments 2 and 3 investigated the impact of preexposing rats to the conditioning context before they were trained with a CS-shock pairing while under the influence of midazolam. Preexposure should latently inhibit the context (Keirnan and Westbrook 1993), either preventing the development of new associations involving the context (McLaren et al. 1989) or interfering with the retrieval of those associations on test (Bouton 1993, 1994). Therefore, context preexposure should reduce the extent to which the context can come to control conditioned fear. Experiment 2 used a 2×2 factorial design: factor 1 was whether or not the rats were preexposed to the conditioning chambers and factor 2 was whether the rats were injected with midazolam or saline before being exposed to the CS-shock pairing. Experiment 3 used the same design but with an additional factor: rats were tested either in the conditioning context or a different context. Thus, experiment 3 aimed to test both the effects of preexposure and of context shift, and thereby determine whether these manipulations interact.
Materials and methods
tion, and were approved by the Animal Care and Ethics Committee at the University of New South Wales. Apparatus Two sets of four experimental chambers were located in different rooms in the laboratory. The background noise level in each room was 65 dB [Bruel and Kjaer (Nerum, Denmark) sound-level meter type 2235]. Each chamber of the first set was 24 cm deep × 24 cm wide × 19 cm high. The aluminium back and side walls and the roof were painted white, and the front wall was constructed of clear plastic and could be opened to place the rats inside the chamber. The floor consisted of stainless steel rods, 10 mm in diameter, spaced 22.5 mm apart (centre to centre). One millilitre of concentrated rose oil (Cara-Mia, Sydney, Australia) was added to the bedding under each chamber to create a distinctive odour. The chambers were illuminated by a white fluorescent light located on the ceiling above the chamber in the centre of the room. The second set consisted of four chambers, each 30 cm deep × 30 cm wide × 27 cm high. The side walls and ceiling were made of aluminium, painted black, and the back and front walls were made of clear plastic, covered on the outside by black cardboard. The floor consisted of stainless steel rods, 2 mm in diameter, spaced 10 mm apart, centre to centre. One millilitre of concentrated banana essence (Rockman, Sydney, Australia) was sprayed onto the bedding material under each chamber to provide a distinctive odour. These chambers were illuminated with a red fluorescent tube located in the ceiling of the room. Unscrambled AC 50 Hz shock from a constant current generator could be delivered to the floor of each chamber in either set, and the current available to each floor was adjusted by reference to an in-line milliamp meter. All chambers were isolated in separate compartments of a wooden cabinet, and the door of each compartment was kept open to permit observation of the rats. The clicker CS consisted of a 72-dB 10-Hz spike (rise time