A simple T-maze method for estimating working memory in mice. Effect of ethylcholine mustard aziridinium ion (AF64A).

Mice were housed in a cage containing a T maze. A watering place was located at the entrance of the maze. The right and left arms of the maze each had two exits, one of which led to the home cage where food was placed, while the other led to the watering place via a bypass. The exit leading to the home cage in either the right or left arm was alternately closed every 90 min. One-way swinging doors were inserted at the entrance to each arm and between each bypass and the watering place. The mice were given a cholinergic neurotoxin--ethylcholine mustard aziridinium ion (AF64A)--(8 nmol) or saline as a control into the left ventricle 2 weeks before they were housed in the apparatus. Those mice housed in this apparatus mastered the alternation task at a 5-sec delay on day 3 in the sham group and on day 4 in the AF64A group. When a longer delay (5-90 sec) was introduced for the mice that mastered the alternation task at 5-sec delay, the AF64A group made significantly more errors than did the sham group at 60- and 90-sec delays. These results show that the apparatus is useful in estimating working memory in mice with little effort.