We present a training procedure and maze equipped with sensors and automated feeders for training spatial behavioral tasks in rodents. The maze can be transformed from an enclosed box to a maze of variable dimensions. The modularity of the protocol and setup makes it highly flexible and suitable for training a wide variety of spatial tasks, and facilitates incremental training stages of increasing maze size for more efficient learning. The apparatus, in its software and hardware, is able to adapt to animal performance, adjusting task challenges and difficulty. Two different methods of automatic behavioral scoring are evaluated against manual methods. Sensors embedded in the maze provide information regarding the order of reward locations visited and the time between the activation of the cue via the nose-poke and the activation of the reward location sensors. The distributions of these reaction times differ between correct and incorrect trials, providing an index of behavior and motivation. The automated maze system allows the trainer to operate and monitor the task away from the experimental set-up, minimizing human interference and improving the reproducibility of the experiment. We show that our method succeeds in training a binary forced-choice task in rats.

No abstract

No abstract

We present a training procedure and a T-maze equipped with sensors and automated feeders for training spatial behavioral tasks in rodents. The maze can be transformed from an enclosed box to a maze of variable dimensions. The modularity of the protocol and setup makes it highly flexible and suitable for training a wide variety of spatial tasks, and facilitates incremental training stages of increasing maze size for more efficient learning. The apparatus, in its software and hardware, is able to adapt to animal performance, adjusting task challenges and difficulty. Two different methods of automatic behavioral scoring are evaluated against manual methods. Sensors embedded in the maze provide information regarding the order of reward locations visited and the time between the activation of the cue via the nose-poke and the activation of the reward location sensors. The distributions of these reaction times differ between correct and incorrect trials, providing an index of behavior and motivation. The automated maze system allows the trainer to operate and monitor the task away from the experimental set-up, minimizing human interference and improving the reproducibility of the experiment. We show that our method succeeds in training a binary forced-choice task in rats.The learning curve in figure 2A was acquired using manual scoring to mark a trial as correct or incorrect.