Train Too Much and a Dog Won't Remember

Training must leave time for the dog's memory to consolidate

Posted Apr 19, 2017

Dennis Hill photo — Creative Commons License
Source: Dennis Hill photo — Creative Commons License

So you are training your dog and spending lots of time educating him, but then a week later he acts as though he only remembers fragments of what you taught him. If you talk to dog trainers they are apt to tell you that you should just add more training sessions. Some of them will tell you that you should train a dog for a period of time on one task, and then move on to train him on another task. The trick, they say, is to continue going through various topics for as long of a training session as you and the dog can sustain without feeling fatigued.

Most dog trainers are aware of issues that can affect the rate at which a dog learns. If you eavesdrop on a group of trainers discussing their methods you are almost certain to hear them talk about the relative effectiveness of rewards versus punishments (click here for more on that topic). You are also likely to hear them discuss the use of markers in training, such as clickers versus voice (click here for more on that). Much less frequently you will hear them talk about schedules of training (click here for more). The emphasis among both professional and pet dog trainers seems to be on how quickly and efficiently the dog performs during the training session.

However, psychologists are beginning to recognize that using good training methods is not enough. They are now beginning to understand that what a dog does after training affects what he will remember later. If a memory is to be useful in guiding behaviors after the training session ends it has to be processed and stored in the brain in what psychologists call long-term memory. The process by which short-term memories are converted to long-term memories is called "consolidation". Data has shown that getting some sleep after learning something can greatly improve consolidation. This is because it is during the REM or dream state of sleeping that memories are sorted through and finally stored in our long-term or permanent memory.

A team of researchers headed by Anna Kis of the Institute of Cognitive Neuroscience and Psychology at the Hungarian Academy of Sciences in Budapest, Hungary set out to explore how what happens after a training session affects how well a dog remembers what he has learned. They began with the notion that, since it is during sleep that memories are consolidated, perhaps the best strategy would be to have a dog take a nap after a training session. So they first had to establish that the process of being trained actually affects what happens during a dog's sleep.

Their initial study was quite straightforward. One group of dogs was given a very simple learning task, namely to respond to a new command label on a task which they had a already mastered. The researchers began with a group of pet dogs who had already learned the basic tasks of sitting or lying down when given a command in Hungarian. For the purposes of this study they were next trained to respond with these same actions when given the command in English. The dogs were allowed to take a nap for three hours, while their brain waves were recorded. The electrical patterns in their brains were compared to another group of dogs who simply practiced the commands that they had already learned in Hungarian (no new learning for this group). What the researchers found was that there were significant differences between the electrical patterns in the sleeping brains of the dogs who had learned something new when compared to those who were simply practicing already learned actions. The researchers felt that what was causing the difference in these brain activity patterns was that the dogs who had been trained recently were consolidating what they had been taught into a long-term memory.

So now the research team decided to see if they could put this information into practice. They went on to conduct a second experiment with the expectation that dogs who are given a chance to take a nap after a training session will have a better memory of what they had learned when later tested. Of course, to be fair, they needed to compare the performance of dogs who were learning something new and then sleeping afterwards to dogs who were being trained in the intensive manner that many competitive dog trainers use, namely teaching the dog something in a training session, and then moving on to teach the dog something else immediately after. They also decided to see if simply having a session of physical activity or perhaps simply a relaxed play session after training, would affect their ability to remember what they had learned.

In this second experiment a group of 53 pet dogs were all trained on the same task that had been used before (shifting commands from Hungarian to English). Following the training one group got to take a nap in their owner's car for an hour. A second group went on to another training session where they learned a new task, unrelated to the first (it was based on lure training). The third group took a walk across campus, while the fourth group got to play with a Kong that had been stuffed with some treats.

The researchers originally expected that the dogs which had the nap after the training session would have perform better when retested. This was not case, and in fact what the dogs had been doing after the training session seem to make no difference when they were retested an hour later — all seemed to be performing at about the same level that they were when training ended. However, recognizing that the effects of memory consolidation often take a while to show up, the investigative team sent the dogs home and then brought them back approximately a week later for another set of tests to see how much of their original training they still remembered. That is when they got a bit of a surprise.

First of all the group which had had the nap after training now actually performed better than it had when tested immediately after their hour-long sleep. Although this seemed to confirm their original predictions the data additionally showed that both the group that had had a walk after training and the group of dogs who had played after training also performed better than they had in their first retest. These other two groups remembered the task at about same level of proficiency as the group which took a nap. The only group which did not show improvement over the interval was the group whose initial training was followed by another training session where they had to learn a new task.

To state this another way, what this research was showing is that a dog who had gone through a training session, and then immediately after got another training session to learn a new task, was less likely to remember that original training. In comparison the dogs that had gotten a break of some sort, either to nap, exercise, or play, actually had better memory and performance a week later. The best guess is that the additional training after the first session had actually interfered with the consolidation process for the earlier learned exercises. To see why this might be the case you might imagine that there is a narrow gateway through which the short-term memories from training into long-term memory must pass in order to get to the long-term storage. If you fill that narrow gateway with too many new memories which need to be consolidated you get a log jam. That means that the processing of some of the short-term memories established during training will be slowed meaning that they will fade before they get a chance to be consolidated into a more permanent form. Other activities which do not involve active learning don't require the same type of processing and therefore don't interfere with the establishing of longer-term memories.

I think that everyone will agree that training is useless if the dogs do not later remember what they have learned. The conclusion to be reached from this experiment is that back to back training sessions actually decrease a dog's later memories rather than improving their performance. Just like kids, dogs need a "recess" between classes.

Stanley Coren is the author of many books including: Gods, Ghosts and Black Dogs; The Wisdom of Dogs; Do Dogs Dream? Born to Bark; The Modern Dog; Why Do Dogs Have Wet Noses? The Pawprints of History; How Dogs Think; How To Speak Dog; Why We Love the Dogs We Do; What Do Dogs Know? The Intelligence of Dogs; Why Does My Dog Act That Way? Understanding Dogs for Dummies; Sleep Thieves; The Left-hander Syndrome

Copyright SC Psychological Enterprises Ltd. May not be reprinted or reposted without permission


Anna Kis, Sára Szakadát, Márta Gácsi, Enikő Kovács, Péter Simor, Csenge Török, Ferenc Gombos, Róbert Bódizs, József Topál1, (2017). The interrelated effect of sleep and learning in dogs (Canis familiaris), an EEG and behavioural study. Scientific Reports, 7, 41873, doi: 10.1038/srep41873