Abstracts

Rationale: Because it is the most readily accessible measure of memory in rodents, spatial performance in navigation tasks is commonly studied in animal models of neurological disorders. The most widely used task is the Morris water-maze (MWM) task, which consists of training rats to find a submerged platform in a pool of opaque water. In this task, the performance of rat models of temporal lobe epilepsy (TLE) is dramatically impaired as compared to controls. This finding parallels the finding that patients with TLE are commonly affected by pronounced episodic memory impairments. For safety and technical concerns, it is particularly challenging to test water maze performance in humans. In an attempt to parallel the research done in rodents, we developed a virtual reality analogue to the MWM. Methods: The Human WM task was developed using the Source Engine (Valve ). This software is widely used to create custom first person shooting video-game environments. The virtual WM was presented on a laptop computer. A Play Station 3 Controller (SONY ) was provided. Movements were restricted to forward/back and left/right body motion. Five healthy volunteers participated in the study. The training environment consisted of a rectangular room (768*396 ft, h= 342 ft; 1 ft= 16 pixels) and the water maze environment was a cylindrical arena (576 ft diameter, h= 123 ft) surrounded by multiple visual cues (a tree, a wall and a building). In both environments, the floor was covered with opaque water (h= 22.5 ft) under which a platform (24*24 ft, h= 2.25 ft) was submerged. Five control subjects were first asked to navigate for 10 minutes in the training environment to get accustomed to the procedure. When the subjects stepped on the platform, a red light turned on to denote success. After training, subjects were dropped in the arena and asked to reach the platform as fast as possible. Once the subject reached the platform, he was allowed to observe the environment for 1 minute without moving. After this delay, he was teleported to another location in the maze. This process was repeated 4 times per session for 3 sessions. The day after, the platform was removed without informing the subjects and they were asked to perform the task once more. Repeated measures ANOVA and logistic regressions were performed. Results: On the first day, there was a significant reduction in latency to reach the platform by session (p<0.001 - repeated measures ANOVA). Latency was significantly longer in session 1 than session 3 (p<0.001). After adjusting for latency, there was also a significant reduction in swimming distance (p=0.017). During the first 30 seconds of the probe test, subjects spent 70% of their time in the platform quadrant. Conclusions: These preliminary results show a striking parallel to rodent data in the water maze. The task is easy to learn and performance improved within a few trials. This technique will likely be very useful to investigate spatial performance in patients with epilepsy, and can be done while recording electrophysiological activity.