Subjects.
Four adult rhesus monkeys (Macaca mulatta), two male and two female, were studied in the conflict study. Six adult male squirrel monkeys (Saimiri sciureus) were used for the observational study, and four adult male squirrel monkeys were used in the drug discrimination study. Monkeys used in the conflict and discrimination studies were maintained at 90 to 95% of their free-feeding weight. Monkeys used in the observational study were maintained under free-feeding conditions. All monkeys were housed individually and maintained under a 12-h light/dark cycle in a temperature- and humidity-controlled room. All procedures were conducted with the approval and under the supervision of the Harvard University Institutional Animal Care and Use Committees. Animals in this study were maintained in accordance with the guidelines of the Committee on Animals of the Harvard Medical School and the "Guide for Care and Use of Laboratory Animals" National Research Council, Department of Health, Education and Welfare Publication No. (NIH 85-23), revised 1996.

Monkeys in the conflict and discrimination studies were prepared with chronic indwelling venous catheters using the general surgical procedures described by Carey and Spealman (1998 ). Under isoflurane anesthesia and aseptic conditions, one end of a polyvinyl chloride catheter (rhesus monkey = i.d., 0.64 mm; o.d., 1.35 mm; squirrel monkey = i.d., 0.38 mm; o.d., 0.76 mm) was passed to the level of the right atrium by way of a brachial, femoral, or jugular vein. The distal end of the catheter was passed subcutaneously and exited in the mid-scapular region. Rhesus monkey catheters were flushed daily with heparinized saline (150–200 U/ml), whereas squirrel monkey catheters were flushed daily with saline that did not contain heparin. All catheters were sealed with stainless steel obturators when not in use. Monkeys wore custom-made nylon mesh jackets (Lomir Biomedical, Toronto, ON, Canada) at all times to protect the catheter.

Conflict Study.
Rhesus monkeys were trained to sit in a custom-designed restraint chair (Crist Instruments, Hagerstown, MD) located in a sound-attenuating chamber. A response lever, stimulus lights (Med Associates, Georgia, VT), and a receptacle into which food pellets (1 g of marshmallow-flavored pellets; BioServ, French-town, NJ) were delivered were positioned in front of the monkey. Monkeys were trained on a multiple schedule of food reinforcement consisting of two components: 1) a schedule of food pellet delivery, and 2) a schedule of food pellet delivery plus a schedule of foot shock delivery (0.25-s duration, 1–3 mA depending on the individual monkey). Four components were available in a session, separated by 10-min timeout periods in which responding had no programmed consequences. Responding was maintained in each component under an 18 response, fixed-ratio (FR) schedule of food pellet delivery. Each component consisted of the schedule of food pellet delivery signaled by red stimulus lights, followed immediately by the same schedule of food delivery combined with a 20 response FR schedule of foot shock delivery signaled by green stimulus lights. Each response requirement was followed by a 10-s timeout. Drugs were administered during the 5th min of the 10-min timeout that preceded each component.

Training sessions were conducted 5 days per week until performance (measured as rates of responding, see below) in both "food only" and "food + shock" components was stable (i.e., no upward or downward trends for 3 consecutive days). In addition, if rates of responding in a component during a training session varied by more than 20% of the corresponding response rates in the previous training session, additional training sessions were conducted until responding was again stable. Once training criteria were met, test sessions were conducted once or twice a week, separated by at least 2 days. Dose-response functions were determined for test drugs using a cumulative dosing procedure similar to the one described by Rowlett et al. (2001 ). Four-point cumulative dose-response functions were determined within a single test session as a result of incremental increases in drug (one-half log units) administered at the 5th min of the 10-min timeout period. Dose-response functions were determined for SL651498 (0.1–3 mg/kg), the classical nonselective BZ agonist chlordiazepoxide (1–30 mg/kg), as well as triazolam (0.001–0.03 mg/kg).

The number of responses in a component, minus responding during pellet delivery and the 10-s timeouts, was divided by the total component time minus the 10-s timeouts to obtain rates of responding (responses per second). Data for multiple determinations were averaged for an individual monkey, and then these response rates were averaged across monkeys (N = 4, unless otherwise noted). The effects of the individual doses of drug were compared with the control rates of responding using Dunnett's test comparing the average response rate engendered by each dose to the average response rate after vehicle administration ( level = p < 0.05).
Observation Study. The observable behavioral effects of BZ ligands were assessed in squirrel monkeys according to the procedures described by Platt et al. (2002 ). Each monkey was habituated to a ventilated, transparent Plexiglas arena (114 cm x 122 cm x 213 cm). This observation arena was equipped with perches, suspended plastic chains, and a wood chip foraging substrate to allow the monkeys to express a range of species-typical behaviors. A video camera was positioned 1 m in front of the chamber and operated throughout the 30-min session.
Drug testing was conducted once or twice per week with control sessions preceded by saline injections on intervening days. Doses of SL651498 (0.3–10 mg/kg i.m., administered 10 min prior to the start of the session) or chlordiazepoxide (3.0–56 mg/kg i.m., administered 30 min prior to start of the session) were administered on separate test days. Trained observers, unaware of the drug being studied, scored the videotapes by recording the presence or absence of each of eight behaviors (Table 1) at 15-s intervals during three 5-min observation periods across the session (0–5, 12–17, and 24–29 min). For each subject, frequency scores (defined as the total number of 15-s intervals in which a particular behavior was observed; maximum score = 20) for each behavior were averaged across the three observation periods of a session because no reliable differences in scores were identified by separated repeated measures analysis of variance. Scores were then averaged across subjects to obtain group means. To determine statistical reliability of treatment effects on each behavior, the effect of dose was determined for each drug by separate repeated measures analyses of variance. Treatment effects were assessed further using Bonferroni t tests, in which the effects of different doses of each drug were compared with vehicle.

Muscle relaxation and ataxia were measured in the same animals during the 6th, 18th, and 30th min of each 30-min observation session. The monkeys were removed briefly from the observation arena by a trained handler and evaluated for degree of muscle relaxation, which was defined as decreased resistance to extension of a hind limb. A score of 0 indicated a normal resistance to extension, a score of (–1) indicated a decreased resistance to extension, and a score of (–2) indicated no resistance to extension (i.e., the monkey was flaccid and completely relaxed). Ataxia was defined as the inability to balance on the 1-cm diameter stainless steel transport pole held in the horizontal plane. A score of 0 indicated that the monkey was able to balance normally on the pole, a score of (+1) indicated that the monkey was able to hold onto the pole but unable to maintain balance (e.g., suspended by limbs below pole), and a score of (+2) indicated that the monkey could neither balance on nor hold on to the pole.

In an additional study, SL651498 (10 mg/kg i.m.), chlordiazepoxide (56 mg/kg i.m.), or triazolam (0.1 mg/kg i.m., administered 30 min prior to the start of the session) were evaluated alone and combined with the 1GABAA-preferring antagonist -carboline-3-carboxylate-t-butyl ester ( -CCT) (Platt et al., 2002 ; 3.0 mg/kg i.m., administered 10 min prior to the start of the session). Monkeys were evaluated for muscle relaxation and ataxia.

Drug Discrimination Study.
Squirrel monkeys were trained to discriminate the conventional BZ agonist triazolam from saline under an FR 10 schedule of food delivery using the procedure described by Lelas et al. (2001 , 2002 ). The monkeys sat in a restraint chair located in a sound-attenuating chamber (Med Associates). In front of the monkey, response levers and stimulus lights were available as well as a receptacle into which food pellets (190 mg of sucrose pellets; BioServ) were delivered. Training sessions consisted of one to four components. A 10-min timeout period preceded each FR component, and either saline or triazolam (0.03 mg/kg) was administered intravenously at the 5th min. Each FR component, during which the stimulus lights were illuminated and the FR 10 schedule of food delivery was in effect, lasted 5 min or until 10 pellets were delivered, whichever occurred first. Ten consecutive responses on the correct lever resulted in food delivery and extinguished the stimulus lights for a 10-s timeout period. Responses on the incorrect lever did not result in pellet delivery and reset the response requirement on the correct lever.

Drug testing sessions were conducted once or twice per week with training sessions scheduled on intervening days. Test sessions were conducted if 80% or more of the total responses occurred on the lever designated correct for that component for at least four of the five preceding training sessions. Test sessions consisted of i.v. injections of saline, triazolam (0.003–0.1 mg/kg), or SL651498 (0.1–1 mg/kg). Overlapping cumulative dose-response functions were administered on different test sessions until at least one dose of drug produced 80% drug-lever responding, decreased response rates to 25% of control, or resulted in two or more animals not completing a response requirement. Antagonism studies also were conducted in which selected doses of -CCT (0.3 and 1.0 mg/kg i.v.) were studied in combination with SL651498. -CCT and its vehicle control were administered i.v. in the first component similar to previous studies (Lelas et al., 2002 ).