This control mouse has learned to press a bar for food reward.
Because we also use this procedure in dopamine-deficient mice, who do not eat normally, the food reward is delivered by intraoral (IO) cannula. The IO cannula is implanted into the cheek, so that food can be infused directly into the mouth. The mouse can choose to eat the food, or spit it out. The IO cannula is coated in an orange protective sheath, so that the mouse does not damage it while grooming.
Eating via an IO cannula is not aversive, and well-fed mice will willingly press the bar to obtain a small infusion of the sweet, fatty liquid food. While the mouse is in the self-administration chamber, the IO cannula is hooked up to an infusion tube, the stainless steel tip of which is visible here. The infusion tubing is mounted to a swivel, so that the mouse is able to move freely. When he is not in the self-administration chamber, the cannula is unhooked from the infusion tube. The IO cannula does not interfere with the ability of the mouse to eat normally.
In this chamber, there are two bars, or levers, and one nose-poke. One of the levers, and the nose-poke, are "active". Pressing the active lever, or sniffing the nose-poke, resulted in a short infusion of food. The signal light over the active lever, or inside of the nose-poke hole, was illuminated during the infusion. In this video, the nose-poke is not illuminated because the response occurred shortly after pressing the active lever, and therefore was within the time-out period, as described below.
The other lever, which is to the right of the nose-poke and is obscured by the chamber wall from this angle, was inactive. Pressing the inactive lever resulted in no infusion, and the signal light was not illuminated. Mice quickly learned which lever would result in an infusion of food and, by the end of the training period, made far more responses on the active lever than at the inactive one.
Our experiment also included a time-out period. After an active response and during the infusion, the signal light remained on and no further rewards were given for responses at either the active lever, or the nose poke, although these responses were recorded. Normal mice learned this and, over time, stopped responding during the time-out period.