Presumably this effect was first noticed without any training, when dogs who'd witnessed their owners having seizures made the connection and started to get agitated when another was imminent.

There is a scent, I believe it is a hormone but I don’t think it is what the body actually produces. If a dog reacts to it they can try to train it. If it doesn’t, they train the dog for something else or find it a good home. They also have people fake seizures. You can probably find some YouTube videos of it.

We have a dog that could do it. She could warn us about my partner’s seizures and fainting a few minutes before it happened. She could fetch help and apply pressure to the abdomen. (I say could because my partner had surgery a year ago and hasn’t had any seizures or fainting since then.)

I definitely noticed a change in my partner before seizures (but not fainting). The dog alerted on me once during a moment of high anxiety. I wouldn’t begin to argue WHAT the dog is actually reacting to, but the training methods people are using seem to work.

I wonder if the fake seizures confuse the dogs about what signs to look for... Might be a quantity over quality problem though (how would you train enough dogs without faking it?)

I’m not an expert on this so I’m just guessing but I think the fake seizures are more for training what to do during the seizure rather than the detection part.

> More specifically, the training for SAD required three main steps. First a positive association is created between the scent of a seizure and treats in order to induce a conditioned emotional response.

The paper goes into a bit of detail on training, but it's a fair and interesting point: How do they get the seizure scent? I guess you could maybe take a sample just after a seizure, or a second person could do it if they were present and prepared to take it.

They took samples after/during seizures.

Samples of what, though? Sweat? Clothing?

Sweat.

Stimulus generalization comes to mind. IIRC Pavlows dogs at some point reacted not just to the food, but even earlier on him entering the room.

I doubt that they train dogs like "Steve's having a seizure. Good! Get that dog here."

I imagine that you do a preliminary training with faked seizures to get responses right and then hope (meaning: Train under real conditions) that the dog will react on a real seizure as well.

I imagine you train them to react to seizures, and they figure the predictive bit out themselves, if they are able to.

How do you rule out the possibility that dogs cause seizures?

The patient already had seizures before getting a seizure dog.