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

An introduction to sequences

ANY-maze has been designed to be extensible, so if your experiments need other types of sequences beyond those described here, please let us know. Where possible we'll add them to the system - for free.

Introduction

ANY-maze includes many measures that relate to the entire apparatus or to individual zones but, while these are generally very useful, what they don't do is report information about movements between different parts of the apparatus.

For example, in a place preference box, you might want to know how often the animal goes from the left side of the box through the tunnel and into the right side of the box. Although at first this just appears to be the number of entries into the right side, it is in fact slightly different, because an entry into the right side would be counted both in this situation and also in the situation where the animal goes from the right side into the tunnel and then back into the right side again - see figure 1.

To address the need to measure these types of movements between different parts of the apparatus, ANY-maze uses sequences. Essentially, a sequence consists of a number of steps which the animal must complete in order to perform the sequence. An animal completes a step simply by entering a particular part of the apparatus, so, in the above example, the steps would be 'Enter (or just be in) the left side' then 'Enter the tunnel' then 'Enter the right side'. If the animal completes these steps in the right order, and without going anywhere else, then it would perform the sequence.

Figure 1. Both of the situations shown would generate an entry into the right side of the place preference box, but just the first one would satisfy a 'Left side > tunnel > Right side' sequence.

Types of sequences

Simple sequences

The sequence described above is an example of a simple sequence. Essentially, to complete the sequence the animal simply has to move through a certain number of steps in the correct order; in this case, Left side > tunnel > Right side.

Bi-directional sequences

The example described above will only be completed if the animal moves from the left side, through the tunnel and into the right side. But you might also be interested in movement from the right side, through the tunnel and into the left side - indeed you might not care which side the animal starts in and which side he ends in; what you want to know is how often he changes sides. In this case, you could specify that the sequence is bi-directional. ANY-maze will then consider the sequence to be performed whether the animal goes through the steps forwards or backwards.

Sequences which can start at any step

In the place preference box sequence, we would always want the animal to start at one end or the other of the sequence, but this might not always be the case. For example, in a water-maze, the sequence shown in figure 2 (below) could be used to detect rotations around the maze. Here, it wouldn't matter where the animal started in the sequence, provided it worked through all the steps in the right order.

As you may notice, there's an implicit requirement for this sequence to be 'circular' - in other words, for the last step to lead back to the first one - so in this case, you would want the sequence to be completed when the animal returns to the starting step. However, this may not always be the case. For example, in a Y-maze, you might be interested in how often the animal moves through all three arms in a certain order (for example, moving from arm A to B to C), but in this case you wouldn't require the animal to return to the first step in order to complete the sequence; rather, the sequence would be completed when the animal enters the last step.

Figure 2. Using a sequence to detect rotations around a water-maze. The purple arrow represents the implicit link between the last and first steps that ANY-maze can add to 'close' a sequence which can start at any step.

You can mix sequences which can start at any step with bi-directional sequences, which is what you would probably want to do in this example - otherwise, the sequence would only count clockwise rotations around the maze.

Sequences which are not broken by moves that are not in any step

In the water-maze example shown above, the animal must move from step to step, and any move to a position which is not in a step will break the sequence – in the case of detecting water-maze rotations, this is exactly what we want. But consider the Y-maze example shown in figure 3; here, we wish to detect moves between the arms, but to get from one arm to another the animal must move through the centre of the maze, and this is not part of any arm (and therefore not part of any step). So if we wish to detect moves from arm A to arm B to arm C, the animal will have to actually go from A to centre to B to centre to C, and thus will never be able to complete the sequence.

Figure 3. In the Y-maze, the animal will necessarily pass through the centre in order to move from one arm to another. In this case, in order to detect a sequence of movements between the arms, it is necessary to define that the sequence is not broken by moves that are not in any step (i.e. moves into the centre).

To address this, you can specify that a sequence is not broken by moves that are not in any step. Thus in the example, any move to the centre will not break the sequence. It is important to understand that a move to the centre is still a move, just that the sequence processing waits to see what happens next; so A to centre to B to centre to C is just like A to B to C. In the same way, a move from A to centre and then back to A is considered to be a move from A to A. This is important because, for example, the animal might move from B to A to A to C, and this would mean that it has NOT performed the sequence B to A to C.

Non-return sequences

A non-return sequence is one in which the steps don't actually have to be performed in order; what's important is that no step is repeated. For example, in a radial arm maze, you might want to see whether the animal visits certain arms just once each, without returning to an arm it's already visited - see figure 3.

Figure 4. A non-return sequence of 'Arm 1, Arm 2, Arm 4, Arm 6' would be satisfied in this example, despite the fact that the animal actually visited the arms in a different order.

Defining a sequence

Sequences are defined in ANY-maze as a series of steps, where each step defines a part of the apparatus that the animal must enter to complete that step. An important aspect of steps is that they can either be defined by selecting one or more areas in the apparatus map, or they can be defined by specifying a zone.

The ability to define a step by specifying a zone is very useful when you want to include a movable zone in a sequence. For example, in the place preference box described above, we used a sequence to detect movement from the left side through the tunnel and into the right side. But perhaps you would be more interested to know how often the animal moves from a drug-paired side through the tunnel and into a non-drug-paired side. In this case the actual side which is drug-paired (left or right) would probably differ from animal to animal, but defining the steps of the sequence using zones, rather than directly selecting areas of the apparatus map, would automatically take this into account.

On the other hand, in the water-maze rotations example, the areas which the animal would have to enter to complete each step would be of little interest in their own right and you'd be unlikely to have defined them as individual zones. In this case, then, you'd want to define the steps by directly selecting areas in the apparatus map.

Note that when deciding whether to use zones or areas for a sequence step, you'll need to consider whether you'll want to compare sequence and zone measures. This is because a sequence step that uses an area of the apparatus will simply use the animal's centre point to determine entry to that area, whereas a step that uses a zone will use the zone's entry settings.

To see how this could make a difference, consider the following example: imagine you have set up a sequence in which the animal must move from area A to area B. Also imagine you have created a zone for area A, and you've specified that the animal is in the zone when its head is in the zone. Now, you compare the latency to enter Zone A with the latency to start the sequence. Despite the fact that to start the sequence the animal must enter area A, you'll find that the latency to enter Zone A is less than the latency to start the sequence! The reason will be because when the animal's head enters the area, a zone entry will be scored, but the sequence will only begin when the animal's centre point enters the area (which will happen later, and in fact may not even happen at all!). If you set up the sequence so that the first step required the animal to enter Zone A (rather than area A), then the latencies would agree.

See also:

© Copyright 2003-2026 Stoelting Co. All rights reserved

ANY-maze help topic T0349