Probing Spatial Memory
in the Water Maze with the On-Demand (Atlantis) Platform

Brendan Hare and Gerry Herrera
Med Associates Inc. and Catamount Research and Development

Med-Associates Inc. was proud to announce the completion of our water maze video tracking system for rats and mice during the 2007 Society for Neuroscience meeting in San Diego. Shortly after the unveiling of this system our engineers completed prototypes of an On-Demand (Atlantis) platform (ENV-595) for use with the water maze. The On-Demand platform is designed such that it remains well below the water surface, and the animal must learn to remain above its position for a certain amount of time (dwell time) in order to cause the platform to rise up and rescue the animal. This style of platform has been shown to improve rats’ spatial learning when used in place of standard platforms for training (Spooner et al., 1994). In this short study, the On-Demand platform was used to examine the effect of rescue on search strategy maintenance when compared to a group of rats in which no platform was used (null) during successive probe trials that lacked a platform altogether.

Methods
Two groups of male Sprague Dawley rats were used in this study. All animals were trained for eight days using eight trials per day with a fixed platform position and three random start locations. The quadrant of the maze in which the platform was located was excluded as a start location. On the day after the eighth training day (Day 9), animals were divided into two groups for probe trial testing in which the standard fixed platform was removed from the water maze. One group of rats (n=6) received 8 probe trials using the On-Demand platform, which was located in the same position as the platform used for training. The second group of rats (n=6) received 8 probe trials with no platform in the water maze (Null). Null animals were placed into the maze at random start locations and tracked for 120-seconds, without a platform available for escape. On-Demand animals were placed into the maze at random start locations and tracked until they remained over the On-Demand platform location for three seconds, triggering the On-Demand platform to rise and rescue them, or until the full 120-second trial period elapsed. Animals in the On-Demand group not rescued by the platform were guided to it after the trial period elapsed. This process of probe trial testing was repeated again on Day 10.

Results After eight days of training, all animals had successfully learned the water maze task and were showing stable performance (Figures 1 and 2). Average escape latencies for the Null and On-Demand group were 7.6 ± 0.5 seconds and 8.5 ± 0.7 seconds, respectively. During the first day of successive probe trial testing (Day 9), animals in the Null platform group gradually showed an extinction phenomenon, in which they gradually started taking longer amounts of time to achieve their first platform crossing. During Trials 5–8 on Day 9 (first day of probe trial testing), the average time to first platform crossing was 23.2 ± 9.5 seconds in the On-Demand group. Animals in the Null platform group took 66.8 ± 10.8 seconds to make the first platform crossing (P < 0.001 On-Demand v. Null). Results for time spent in platform section are similar to those of time to first platform crossing. Rats in the On-Demand group spent an average of 66.8 ± 5.7% of trial time in the platform section during trials 6–8 on Day 9, while those in the Null group spent an average of 25.3 ± 2.4% (P < 0.001).

Escape latency of animals in the Null and On-Demand demand groups over the course of an 8-day training period. Representative swim paths for the last training trial (Day 8, trial 8) in (A) the Null, and (B) the On-Demand group.

These results remain consistent during the second day of successive probe testing (Day 10). Animals in the On-Demand group remain in the section of the maze containing the platform for a significantly longer percentage of trial time on Day 10 than animals in the Null group (65.6 ± 1.2% and 33.1 ± 1.6% respectively, Figure 3).

Percentage of time spent in the platform section for Null and On-Demand subjects on:
(A) The final day of training. (B) The second (final) day of successive probe trials
(* = P < 0.001 On-Demand v. Null).

In fact, animals in the On-Demand group spend similar amounts of time in the platform quadrant during the last day of testing (Day 10) as they did during the last day of training (Day 8; 65.6 ± 1.2% and 64.8 ± 2.3% respectively), whereas animals in the Null platform group spent significantly less time in the platform section during testing than they did during training (Figures 3A and 3B). Furthermore, on the second day of probe trial testing (Day 10), both groups of rats swam fairly quickly to the platform location on the first trial of the day (Figures 4 and 5). However, only animals in which an On-Demand platform was used continued to swim relatively directly to the platform location during subsequent trials (Figures 4 and 6). Overall, subjects from the On-Demand group make their way to the platform location significantly faster than animals in the Null group (12.6 ± 2.4 seconds and 55.2 ± 6.4 seconds respectively) on Day 10 (Figure 4).

Time until first platform crossing for subjects in the Null and On-Demand groups on the second (final) day of successive probe trials (* = P < 0.001 On-Demand v. Null).

Representative (A) Null and (B) On-Demand animal swim paths from trial 1, Day 10, showing 15 seconds of swim time. Mean time to first platform cross was 16.7 ± 6.5 seconds and 18.5 ± 13.5 seconds for the Null (n=5) and On-Demand (n=6) groups respectively. (C) Full swim path of Null animal shown in panel (A).

(B) On-Demand animal swim paths from trial 8, Day 10, showing 15 seconds of swim time. Mean time to first platform cross was 74.5 ± 15.5 seconds and 16.4 ± 5.2 seconds for the Null (n=5) and On-Demand (n=6) groups respectively. (C) Full swim path of Null animal shown in panel (A).

Discussion
This study used a short dwell time (3 seconds) and two days of successive probe testing in an attempt to quickly bring out a change in the search strategy of the Null group. The aim of this study was to demonstrate the On-Demand platform’s utility for researchers using probe trials as a means of drawing conclusions about a subject’s spatial memory or search strategy. When using the On-Demand platform, subjects seem to use the same search strategy over successive trials compared to those animals receiving typical platform-absent (null) probe trials, in which they swim to the platform location during the first platform-absent trial, and then gradually adopt more varied search strategies. Subjects in the On-Demand group made their way to the platform location and circled the platform area until the dwell time elapsed, allowing the platform to rise and “rescue” them, maintaining results consistent with the final day of training. The percentage of trial time the Null group spent in the platform section of the pool and time to first platform cross remained consistent with that seen from the On-Demand group until the final three trials of day nine, the first day of successive probe testing. On the first trial of Day 10, Null animals quickly cross the platform location before utilizing a new search strategy on successive probe trials that results in time spent in the platform section that is little more than would be expected by chance. This differs significantly from the percentages seen in the On-Demand group.

The On-Demand platform is a tool to aid researchers in solving many issues that confound water maze testing. Researchers using platform-absent probe trials at the end of each training session may find that animals simply “float,” having learned that there is no platform present on this trial. Using the On-Demand platform and the “raise-after” feature can maintain motivation in animals while allowing researchers to examine search strategy. An animal that uses the pool edge as a starting point and swims concentric circles until it finds the platform (non-spatial memory) will not be successful when an On-Demand platform is in use since the animal will have swam out of the “include radius” before the dwell time elapses. Researchers performing lesion or aging studies will find the On-Demand platform useful for reducing the effects of extinction when using successive probe trials over the course of long term studies.

Please contact Med Associates Inc. with any questions regarding the On-Demand platform (ENV-595) or Water Maze system with video tracking software (MED-VWM).

References
Spooner RI, Thomson A, Hall J, Morris RG, Salter SH. The Atlantis platform: a new design and further developments of Buresova’s on-demand platform for the water maze. Learn Mem 1: 203-211, 1994. •