Morris water maze

The question

Right, let's get our feet wet. How does a pool of cloudy water tell us whether a rat has a map in its head? You're watching Morris Water Maze Software Guide, over at morriswatermazesoftware.com.

Meet the Morris water maze

The Morris water maze is a round pool, usually 1.2 to 2 metres across for rats and a little smaller for mice, filled with opaque water. A small escape platform sits just under the surface. The animal can't see it, so it has to learn where it is from cues around the room: posters, shelves, the door. That makes it a test of allocentric spatial memory, which leans heavily on the hippocampus.

How it runs

So, what does a session look like? First, set the pool in a room with stable, distinct distal cues, and keep the water at a steady temperature (around 21 to 25 degrees for rats). Next, run acquisition trials over several days: release the animal from varied start points facing the wall, and give it up to about 60 to 90 seconds to find the platform. Then, let it sit on the platform for a few seconds so it can take in the room, then dry it and return it to a warm cage. After that, remove the platform for a probe trial, often 24 hours after the last training trial, and record where the animal searches. Finally, optionally add reversal (move the platform) or a visible-platform session to check vision and motivation.

Strengths and limits

Why do labs keep coming back to it? It needs no food deprivation: escape from water is the motivation. Learning is fast, often within four or five days. Decades of published data make results easy to compare. Probe trials separate memory for the place from simply bumping into the platform. And the catch? Swimming is stressful, and stress can mask or mimic memory effects. Strains and ages differ in swimming, floating and wall-hugging. Latency alone can improve through non-spatial habits such as circling.

What to measure

This is where tracking software earns its keep. Latency and path length: the basic acquisition measures; they fall with learning but also fall with procedural strategies, so treat them as a first look, not the answer. Time and path in the target quadrant, and platform crossings: classic probe measures; they show memory, but they ignore swim speed and miss animals that search tightly around, but not exactly over, the platform spot. Path efficiency, heading error, the corridor test and the cone test: directional accuracy: does the animal set off straight towards the goal, from any start point. Close passes, time close to the target and the Gallagher proximity measures: how near the animal stays to the goal across the trial, sensitive even when it never quite crosses the spot. Thigmotaxis, circling, chaining, floating and active speed: the non-spatial and non-strategic behaviours you need to rule out before claiming a memory effect. Picture a mouse in its third day of training. Its latency has dropped from 50 seconds to 15. Great, right? But the trace shows it swimming a neat ring at exactly the platform's distance from the wall. That's chaining, not a map.

The takeaway

One thing to take away: time and path in the target quadrant, and platform crossings and its neighbours tell you how the task was done, not just how fast. Scripts, chapters and sources live on this page.