One solution to minimizing such events is to keep head-restraint periods short (<1 s). A second solution, which we used for head-restraint periods of up to 8 s, was to deliver intermittent water reward (0.5–1 Hz) during head restraint. Gamma-secretase inhibitor A third solution, which
we used for 6 s long head-restraint periods without intermittent water reward, was to provide a rat-activated release switch. We observed that rats pushed on the floor of the cage when they attempted to withdraw their head from the headport. In this approach, the floor of the cage was mounted on a low-friction linear slide with a 2.5 mm travel. Movement of the floor toward the kinematic clamp, caused by the animal pushing with its hind legs, would depress a 1.67 N force snap action switch, which was used to trigger release of the clamp. The release switch appeared to be successful in preventing aversion to the clamp and allowed successful training for long head-restraint
periods: in sessions with 6 s long head-restraint periods and without any water reward during head restraint, an average check details of less than one trial per session was aborted by early release. To determine whether the voluntary head-restraint system could be used with newly developed methods for high-throughput behavioral training, we incorporated a second generation, fully automated, head-restraining system into a semiautomated rat training facility (Erlich et al., 2011 and Brunton et al., 2013). In this facility, rats are placed into operant chambers for a 1.5–2 hr behavioral training session by husbandry isothipendyl staff blind to the experiment being performed. During the behavioral training session, fully automated custom software controls the progression of rats across the stages of training. At the end of the session, the rat is removed from the chamber and is
replaced by the next rat to be trained. In this way, six to nine rats per box can be trained daily while husbandry staff monitor the rats’ health and weights and provide food and supplementary water. Human intervention is required only for animal transport and husbandry, allowing the facility to be readily scaled to many automated boxes running in parallel. To automate training stage 1, we mounted the center nose poke on a linear translation stage driven by a stepper motor driver and robotically controlled by signals from a computer running behavioral training software. After each successful trial, the nose poke was moved 200 μm away from the inside of the chamber. To automate training stage 2, we provided piston pressure by a voltage-controlled pneumatic regulator, which was in turn controlled by the behavioral training software. Computer control over piston pressure enabled the gradual ramping increase of piston pressure at the beginning of each head-restraint trial. This prevented loud noises or jerking movements during piston deployment, which facilitated rapid acclimation of rats to the kinematic clamp.