Variability of drug self-ad ministration in rats

Rationale: Although temporal patterns of drug self-administration in animals are known to be highly regular, this regularity has rarely been quantified or systematically compared across reinforcers. Objectives: Over a range of unit doses, this study assessed: (1) the within-subject variability of inter-infusion intervals (latencies); (2) the estimated whole-body drug level at the time of self-infusion; (3) the within-subject variability of these drug levels; and (4) the statistical dependence between successive latencies, to determine whether regularity is maintained by compensatory, moment-to-moment adjustment of latencies. Methods: Rats were trained with cocaine (10-1000 mug/kg per infusion, IV), remifentanil (an ultra-short acting opioid; 0.25-32 mug/kg per infusion, IV), or food (20-180 mg/delivery). Results: Within subjects, latencies were most consistent from infusion to infusion at unit doses on the descending limb of the dose-response curve. However, the drug level at the time an infusion was initiated was actually least consistent at these doses. Sequential latencies showed only a weak autocorrelation for both drugs. Conclusion: These results suggest that highly consistent response patterns are not simply a product of precise titration of drug levels. The weak autocorrelation between sequential latencies suggests that temporal regularity of responding is not maintained through compensatory adjustments of post-infusion pauses.