Morris water maze

A blinded water-maze study of spatial learning in groups of virtual mice, from random allocation and training to the probe trial, exclusions and a mixed ANOVA.

Guided experiment: Testing spatial memory in the Morris water maze.

In the Morris water maze an animal swims in a pool of opaque water and learns to escape onto a platform hidden just below the surface, finding it from cues around the room; its path is recorded by a video camera and tracking system (Morris, 1984). Rats with hippocampal lesions were impaired at this place navigation (Morris et al., 1982), and chronic intraventricular infusion of the NMDA-receptor antagonist AP5 impaired place learning without affecting visual discrimination learning, while suppressing long-term potentiation in vivo (Morris et al., 1986). Standard procedures combine training trials from several start positions with a probe trial in which the platform is removed, reversal training, and cued trials with a visible platform that detect performance factors unrelated to place learning, such as differences in vision, swimming or motivation (Vorhees and Williams, 2006). The studies of Morris and colleagues cited here used rats; the simulator runs the same kind of study with virtual mice in a smaller pool.

Design the experiment

Questions to investigate

Run the study with the default settings. Does the treatment group learn more slowly, and is its probe-trial preference for the target quadrant weaker?
Set the impairment to 0%. Run the study several times: how often does a study of 10 mice per group still report a significant difference, and what do the confidence intervals look like?
Reduce the number of mice per group to 4 and then raise it to 16. How do the width of the confidence intervals and the smallest detectable effect change?
Why is swim speed compared between groups, and what would a difference in speed mean for the latency results?
Increase the chance of floating mice. How many are excluded by the pre-specified criterion, and from which groups, once the code is broken?
Add a reversal phase. Which group searches the old platform location first, and why?

In a real laboratory

Mice are handled before testing and tested at the same time of day. The pool contains water made opaque with a non-toxic additive, kept at a constant, recorded temperature and surrounded by fixed distal cues; a camera above the pool feeds a tracking system. The question, primary outcome, sample size (from a power calculation), randomisation scheme and exclusion criteria are written down in advance. Allocation is concealed: a colleague assigns coded identities, and the experimenter does not know the groups until the analysis. Genotype cannot be allocated at random, so littermates are used and testing order is randomised. Each mouse is released facing the wall, guided to the platform if it fails to find it, and dried and warmed after each trial. Reports follow the ARRIVE 2.0 guidelines (Percie du Sert et al., 2020). All procedures require approval by an Institutional Animal Ethics Committee (in India, under the CCSEA and the Prevention of Cruelty to Animals Act 1960; elsewhere, the equivalent national body), and the 3Rs apply: replacement, reduction and refinement.

How the simulation works

The pool is about 120 cm across, with a 10 cm platform in the north-east quadrant; positions are recorded at 25 Hz with a small tracking error. The heading of each mouse combines random exploration, wall-following (thigmotaxis) that fades over the first trials, and a component directed towards a remembered platform location. The memory grows after every trial towards a ceiling, so paths become shorter and more accurate. Mice differ in swim speed (about 15–25 cm/s), learning rate, memory ceiling and thigmotaxis, and a chosen fraction float for long periods. The treatment slows the growth of the memory and lowers its ceiling but leaves swim speed and visible-platform performance unchanged, a pattern loosely based on the impaired place learning after hippocampal lesions or AP5 in rats (Morris et al., 1982; Morris et al., 1986). A trial in which the platform is not found ends at the maximum duration, which is recorded as its latency. All values are approximate model settings. Sex has no effect in the model; handling, stress, water temperature, the inter-trial interval, time of day, litter and cage effects, and strategies other than place learning are not simulated.

Morris, R. (1984). Developments of a water-maze procedure for studying spatial learning in the rat. Journal of Neuroscience Methods, 11(1), 47–60. https://doi.org/10.1016/0165-0270(84)90007-4
Morris, R. G., Garrud, P., Rawlins, J. N., & O’Keefe, J. (1982). Place navigation impaired in rats with hippocampal lesions. Nature, 297(5868), 681–683. https://doi.org/10.1038/297681a0
Morris, R. G., Anderson, E., Lynch, G. S., & Baudry, M. (1986). Selective impairment of learning and blockade of long-term potentiation by an N-methyl-D-aspartate receptor antagonist, AP5. Nature, 319(6056), 774–776. https://doi.org/10.1038/319774a0
Vorhees, C. V., & Williams, M. T. (2006). Morris water maze: procedures for assessing spatial and related forms of learning and memory. Nature Protocols, 1(2), 848–858. https://doi.org/10.1038/nprot.2006.116
Percie du Sert, N., Hurst, V., Ahluwalia, A., et al. (2020). The ARRIVE guidelines 2.0: updated guidelines for reporting animal research. PLoS Biology, 18(7), e3000410. https://doi.org/10.1371/journal.pbio.3000410

About the simulations

Each simulator is a simplified version of a published model of cells, signals, instruments or behaviour, and runs in the web browser. The simulations show how the techniques work and how design choices change the results; they do not replace training in a laboratory.

All data are simulated and generated afresh on each run, so results vary between runs, as in real experiments. Nothing is stored or sent: designs and results are lost when the page is closed, unless the protocol is copied or the data file is downloaded.

Last updated 20 October 2026