Dr. David Bucci (right) and graduate student Chris Keene (left) look over their 8-Arm Radial Arm Maze at Dartmouth College.

Use of the Radial Arm Maze to Examine Multiple Forms of Learning and Memory

David J. Bucci, Ph.D.
Department of Psychological and Brain Sciences,
Dartmouth College, Hanover, New Hampshire

The radial arm maze, originally made famous by Olton and colleagues in the 1970s (1), has been used extensively to study the neural substrates of learning and memory. The maze consists of elevated runways, usually 8, that radiate from a central hub. The maze was originally designed to examine spatial learning and memory and was surrounded by large visual cues placed around the room housing the maze. In this procedure, rats are released into the central hub and allowed to freely explore the maze and travel down the arms to retrieve food. The optimal strategy is for the rat to visit each arm once, since trials are typically time-limited. It is believed that rats use spatial cues in the room to navigate between arms in the maze. Performance is evaluated through a variety of measures, which may be more or less appropriate depending on the nature of the study. Typical measures include latency to visit all the arms, number of arms revisited (error of commission), and number of arms never visited (errors of omission).

There are several modifications of this procedure that can be used to tap spatial working memory versus reference memory. For example, once a rat returns to the central hub after visiting an arm, the doors to all the arms can be closed and the rat forced to wait for a period of time before the doors open again. Varying the time the rat must wait enables the experimenter to examine delay-dependent working memory (2). In a procedure designed more to tax reference memory (3), the same four arms are baited each day, while the other four do not contain food. The rat must use cues in the environment to learn which four are baited and quickly learns to visit only those four arms. In either case, the experimenter can use a variety of measures to determine what sort of strategy the rat is using to perform the maze. Examples include analysis of angle and arm bias and the order in which the arms are visited (4). In addition, the task can be arranged so that spatial cues cannot be used to navigate the maze, forcing the rat to adopt a different strategy (5). By making the walls of the maze opaque, the rat must resort to more local cues or perhaps a response strategy, such as, always turn left when leaving the hub, to determine which arms are baited with food. The difficulty of the task can be manipulated by changing the number of available arms.

Our laboratory has used the Med Associates rat radial arm maze to examine the effects of different brain lesions and drug manipulations in many versions of the task. The automation available through the Med Associates radial arm maze apparatus has proved valuable in each of our studies. The apparatus is equipped with photocells that detect when the rat has entered or vacated an arm and the arm doors can be opened and closed automatically based on the rat’s position in the maze. This greatly reduces the possibility of experimenter bias as well as the chance that the rat is using the experimenter as an additional environmental cue if he/she has to approach the maze to manually open/close doors to the arms. We have used the spatial working memory version of the task to examine the contributions of the retrosplenial cortex to spatial learning (6). We have also examined the effects of kynurenic acid, an endogenous antagonist that acts at NMDA receptors and nicotinic cholinergic receptors on working memory (2), and assessed angle bias to verify the specific strategy used by control and experimental rats. We have also modified the task to examine the effects of our manipulations on motivation by measuring the latency to retrieve food when only a single arm was made available. Other studies ongoing in our laboratory are examining manipulations of the striatum and effects on performing the non-spatial version of the radial arm maze.

1. Olton D.S. and Samuelson RJ (1976) Remembrance of places past: spatial
          memory in rats: Journal of Experimental Psychology, Animal Behav Process 2,
          97-116.
   
   
2. Chess, A.C., Simoni, M.K., Alling, T.E., and Bucci, D.J. (2006) Elevations
          of endogenous kynurenic acid produce spatial working memory deficits.
          Schizophrenia Bulletin, doi:10.1093/schbul/sbl033.
   
   
3. Jarrard LE. (1983). Selective hippocampal lesions and behavior: effects of
           kainic acid lesions on performance of place and cue tasks. Behavioral
           Neuroscience 97, 873-889.
   
   
4. Loh, E.A., Smith, A.M., Roberts, D.C. (1993). Evaluation of response
           perseveration of rats in the radial arm maze following reinforcing and
           nonreinforcing drugs. Pharmacology, Biochemistry, and Behavior 44,735-740.
   
   
5. Chang, Q. and Gold P.E, (2004) Inactivation of dorsolateral striatum impairs
           acquisition of response learning in cue-deficient, but not cue-available,
           conditions. Behavioral Neuroscience 118, 383-388.
   
   
6. Keene, C.S, Chess, A.C., and Bucci, D.J. (2005) Contributions of the
           retrosplenial cortex to attentional processing of conditioned stimuli and
           spatial learning. Program No. 411.13. 2005 Abstract Viewer/Itinerary
           Planner. Washington, DC: Society for Neuroscience.