ANY-maze Help > The ANY-maze reference > The Protocol page > The elements of a protocol > Behaviour > Zones > An introduction to zones

An introduction to zones

Introduction

ANY-maze will always calculate results for a range of measures for a test, for example the total distance the animal travelled and the test duration. These measures, however, always apply to the apparatus as a whole, so how do you get results for just a certain part of the apparatus - for example, how long the animal spent in the open arms of a plusmaze?

The answer is that you use Zones. A zone defines a certain part of your apparatus for which you would like individual results. You can define any number of different zones, and ANY-maze will calculate the results separately for each one.

Defining a zone

As you will probably recall, when you draw an apparatus map, you don't just define the border of the apparatus; you also divide it into discrete areas, and these areas will now be used to define your zones. Essentially, all you have to do to define a zone is select the area or areas of the apparatus map which will constitute the zone - see figure 1.

Figure 1. Defining the areas which constitute the 'Open arms' zone of a plusmaze. Note that the zone can consist of more than one area, and that the areas don't have to be contiguous.

There's no limit to how many different areas you can select to create a single zone, nor do the areas have to be contiguous. Furthermore, one area can appear in any number of different zones.

Hidden zones

In some apparatus, you may have areas in which the animal can disappear from the camera's view - a tunnel, for example, or a nest box. In such apparatus you can define a 'hidden zone', and ANY-maze will assume that if it can't find the animal anywhere in the apparatus, then it must be in the hidden zone. Thus you will be able to score measures such as Time in the tunnel and Number of entries into the tunnel.

In some circumstances you may have multiple hidden zones, for example, two tunnels connecting different parts of the apparatus. In this situation you can define two distinct tunnels by creating two hidden zones and having the system determine which tunnel the animal is in, based on where it last saw it before it disappeared.

Investigation zones

In apparatus, such as the Novel Object Recognition (NOR) test, you're not only interested in the animal being in specific zones; you'll usually also want to know how much time the animal spent investigating certain zones. Clearly this requires a definition of investigation and in ANY-maze this can be set to include any combinations of the following:

You can choose which of these criteria to apply, so it might be that you do want to consider the animal to be investigating the zone if it is inside it.

Figure 2. The animal is 'investigating' the object because: its head is within 25mm of it (the dotted line shows the investigation area), it is oriented towards the object (the flashlight-like beam shows the orientation) and it is not on top of the object.

Investigation zones can be moveable (see the next section) which is very useful in tests such as NOR where the novel object may be placed in different locations in different tests.

Zones which change location

It's not uncommon to have zones in your apparatus which are sometimes physically located in one part of the apparatus and sometimes in another. For example, a 'Known arm' in a Y-maze may sometimes be on the left and sometimes on the right; or an island in a water-maze might sometimes be in the NW quadrant, sometimes in the SW and sometimes in the NE. Clearly then, there needs to be a way to tell ANY-maze that a zone might move about.

Defining a zone which can change its position is actually very simple - instead of defining a single location for the zone, as described above, you define a number of different positions which the zone can adopt - see figure 3.

Figure 3. The two possible positions of the 'Known arm' zone in a Y-maze. The zone will always adopt one of these two positions in a test.

Clearly, if a zone can adopt different positions, then ANY-maze will need to know which position the zone is in when it runs a test. For example, is the known arm on the left or the right side of the Y-maze? For experiments in some apparatus, the position of a zone may just depend on the animal which is being tested, whereas in others it might depend on the animal and/or the trial. For example, in a water-maze you might locate the island in the NW for animal 1 and in the SW for animal 2, and throughout the animals' trials these positions won't change - for animal 1 the island will always be in the NW, and for animal 2 it will always be in the SW. On the other hand, in a T-maze you might alter the reward arm in a certain sequence so that, for example, in an animal's first trial it's on the left and in the second trial on the right.

The importance of this difference in how a zone alters position, is related to how you tell ANY-maze where the zone will be:

Figure 4. In this water-maze experiment, the position of the 'Island' zone differs both between and within the animals, so ANY-maze asks you to specify the position before each test. The purple areas indicate the possible positions of the island; the blue area indicates the position the user has selected.

ANY-maze also allows you to define relationships between the positions of movable zones. For example, in a Y-maze, you might test an animal twice, once with one of the arms closed off, and a second time with both the arms accessible. Let's assume you call the arm which is always accessible the 'Known arm', and the arm which is only accessible in the second trial the 'New arm'. You then randomly choose which side of the Y-maze will be the 'Known arm' for each animal, and you tell ANY-maze what they are: Animal 1, Known arm is on the left; Animal 2, Known arm is on the right; etc. But this implies that for animal 1, the New arm must be on the right and that for Animal 2, on the left. So rather than also having to tell ANY-maze the position of the New arm for every animal, you can instead just tell it that when the Known arm is on the right, the New arm is on the left and vice versa.

This method of defining the position of one zone based on the position of another is a great time saver, as it reduces the amount of information you have to enter each time you run an experiment - remember, this information is part of the protocol, so you just define it once and reuse it in all your experiments.

Defining zone entries

One of the problems with video tracking systems is that they tend to define an animal's position based on a single point, usually somewhere in the centre of the animal. This is usually fine in apparatus such as a water-maze, where you're interested in such things as the distance the animal travelled, but can create problems in apparatus such as a plusmaze.

Specifically, two problems often occur:

ANY-maze addresses both these issues by including the ability to define a zone entry using the entire area of the animal, rather than just its centre point.

Specifically, you can define how much of the animal must enter the zone before an entry occurs - perhaps 85% - and how much of the animal must remain in the zone to prevent a zone exit from occurring - perhaps 75%. In this example, then, the value of 85% will equate nicely to the plusmaze '4 paws in the arm' rule - i.e. by the time 85% of the animal is in the arm, its 4 paws probably will be too - see figure 4. And the spurious entries problem will also be fixed because we're saying that the animal has to get 85% of its body into the zone to count as an entry but then, provided 75% of it stays in the zone it won't exit. Effectively we have a 10% buffer between an entry and an exit, which means that the animal can shift about across the zone division without causing multiple entries.

Figure 4. In the first image, the centre of the animal (indicated by the orange dot) has entered the open arm, but ANY-maze doesn't count this as an entry because less than the required amount of the animal - 85% in this case - is in the zone. The blue shading indicates the area that the system considers to be the animal. In the second image, ANY-maze does consider the animal to be in the zone, which it indicates by shading the zone in green.

ANY-maze also allows you to define zone entries based on the position of the animal's head.

Editing zones

You can alter anything about zones, including adding new ones at any time, whether before, during or after you have run the tests in your experiment. This can be really useful, as the video below demonstrates.

Figure 5. This video shows how easy, and useful, it is to add or edit zones in an experiment that includes performed tests.

See also: