When the dogs explore their surroundings, they discriminate scents all the time. They know who of their favourites or enemies has just walked by the street, and they can smell where their owner has been while gone. It has been said that a human can smell pizza and think “Mmm…Pizza!”, but a dog can smell pizza and think “Mmm…pepperoni, garlic, tomato sauce, mozzarella cheese, wheat, olive oil,…”.
When we speak scent discrimination we typically mean a specific exercise where a dog is sent to a set up that is called a scent line up, a scent wheel or a scent wall. These are different types of set ups that are used to train dogs to find and indicate a specific target scent. These set ups are not made for dogs to learn to discriminate scents in general, but the scents that are important for their future job as detection dogs. For the dogs (and other animals) working in laboratory circumstances, the scent discrimination is their main working tool and they don’t necessary need to be trained for searching in other surroundings.
There are three typical discrimination exercises 1) reinforcing the target scent, 2) discriminating the target scent from other scents and 3) stimulus control.
Reinforcing the target scent is an exercise where only the target scent is placed in one or more spots of the setting. The dog searches for the target scent that can be placed anywhere in the setting. No distraction scents are used in this exercise, but the target scent can be made for instance stronger or milder. The other spots can be controls or they can just be left empty.
Control scent is usually a scent that is often or always present with the target scent, in practise and/or in real situations, but by itself is not relevant. One typical control scent is the scent of rubber gloves. If the gloves are used while handling the target scent, the dogs might start reacting to the scent of the rubber gloves. Since it is problematic that a dog starts to alert rubber gloves, it is useful to make that one of the control scents early in the beginning.
When the dog has learned the target scent, discrimination exercises can contain any of the other scents that might distract the dog while its doing a search or tracking. These could be for instance food, scent of human beings, scents of other animals or a toy. When these distractions are made non-profitable by scent discrimination, it is easier for a dog to ignore them while working too.
The last but an important use of scent discrimination is stimulus control. When the dog is trained long enough and thoroughly enough it should only alert the target scent. Stimulus control in scent work means that the dog alerts only the trained target scent/s, only when the target scent/s are present and never when it is not. Some trainers may even want a dog to alert only when it is allowed to work, after the search cue, but that is rather rare since in most cases there is no harm if the dog alerts the target scent without the cue. When training stimulus control, the target scent should be taken off the setting once in a while to test how the dog reacts. If the target scent is not present in the setting, the dog should not alert. It should either keep on searching or give an alert to “clear” depending on how its been taught.
Scent discrimination is an effective way to teach a dog the target scent and make sure that it only reacts to it and nothing else. It gives an opportunity to keep reward density high and make simple but informative searching exercises for the dog.
But what could go wrong then?
There are couple of common training problems that occur with scent discrimination with these kinds of set ups. If the final goal is to do searching or tracking and not only discrimination in laboratory circumstances one should remember that scent discrimination is only a tool to teach the dog the basics and to test its skills regularly. It’s something that should be done alongside with search and tracking training and not something that one should train constantly. When the dog knows the basics, it should focus more on the job itself. And here comes why:
One big problem with scent discrimination is us, dog trainers. As discussed in my previous blog post (How do you know, if the dog knows?) dogs read us people like open book. No matter how hard we try not to influence their work, if we know where the target scent lies, it’s extremely likely that the dog can read it from our body language. It is often seen how the handlers train scent discrimination by standing by the scent line or scent wheel. Or maybe even sit by the cans to change the order between the repetitions, like in this video clip.
It should be remembered that as long as the handler is in the sight of her dog the dog can read the handler! The more often this happens, the more the dogs behavior gets reinforced for reading the handler. Thus, the dog may look like it is working very well on a line or on a wheel but it really has no clue.
Another problem occurs when the dog is sent to a real search. It starts asking for help from the handler very easily since it’s used to seeing the handler and getting help from him/her. The dog doesn’t learn how to work independently and how to work in distance. This on the other hand makes the handler want to pressure the dog by either showing directions or by giving a command to search. And if the handler starts helping the dog, thus intervening its work, the dog learns that it can always ask the handler and the vicious cycle is ready.
To avoid the problems with scent discrimination, the set up should always be made so that the dog doesn’t pay attention to the handler while searching. or Blind or double blind exercises may also be helpful. The handler should either only walk the dog trough the set up with even pace or stand aside, away from the set up, while the dog searches independently. And if possible, another person should be present to mark the right behaviour for the dog. If the handler needs to work alone, an automated rewarding machine is a great help! That leaves the dog to focus on the task and the machine, and leaves the handler out of the sight. It gives no signs to the dog what so ever, no matter how much the dog tries to ask for help or read the body language of a trainer.
If the dog needs to do searching or tracking as its main task, it should not work with scent discrimination set ups for too long before starting to train independent searching. If the scent is strongly connected to a set up, it is harder for the dog to learn how to work in a bigger space without that familiar set up. To be efficient, one should first reinforce the target scent and then immediately start hiding it bit by bit to teach the dog to seek its way to it. After the dog knows that it is supposed to find the scent by sniffing and not only look for a container, it makes it easier for it to learn the scent discrimination set up as well.
This post is also available in: Suomi (Finnish)