Get Ready: AI, rPPG, and Migraine Prediction

The prediction and diagnosis of migraine is changing fast, although in some ways it’s not changing at all.

Since the dawn of time, patients and doctors have been able to predict headache attacks to a certain extent, and to notice what prevents them or triggers them.

But then along came more sophisticated computers. The computer can analyze large amounts of data all at the same time, in a way a human can’t. As long as it’s fed numbers, it can search for patterns. Huge medical studies have shown us new patterns that we never would have noticed as humans, even if we had collected all the information together.

Then along came the ubiquitous smart and surveillance devices. The ease with which we can take certain measurements of ourselves and others means more data for the computers, and so the discovery of new patterns.

So ever since the smartphone came along, developers have been trying to come up with apps that could predict migraine attacks, identify triggers, and even recommend personalized treatments. An early success what N1-Headache, formerly Curelator, which began to question some common beliefs about migraine triggers.

Subtle changes in the body can tell us a lot about headache. After all, many patients have noted that their dogs seem to be able to predict migraine!

One area of interest that you need to know about is rPPG – that is, (brace yourself), remote photoplethysmography.

So imagine this. You’re on a video chat with someone, and their computer is actively analyzing the video feed. On the other end, information starts popping up on the other person’s screen:

• This person is nervous about what you just said.
• This person will probably have a migraine attack within the next 24 hours.
• This person is at high risk of [a certain disease]

Or perhaps you’re not even on a video chat. Maybe there just happens to be another camera in the room, watching you. Or maybe a computer is going over your social media posts and photos.

Remote photoplethysmography is one method of measuring visible changes in your body, such as in your cardiovascular system.

Now I’ve started out presenting this in a rather negative light. The positives are obvious for you personally, and for the medical community. Gathering huge amounts of data about migraine patients will hopefully lead to some leaps forward in migraine treatments.

Collection of personal data can help you customize your own treatment, and sharing that data – even anonymously – can help you compare your situation to other similar patients. This is much more powerful that an online forum where you meet someone with similar symptoms!

The use of AI to go through the huge amounts of data out there is a developing area of research, as you can see for yourself with a quick search online. Although in many ways it’s not so new, it is becoming much more powerful. How will you use it? How might you try to limit the sharing of your data, or do you care? It’s time to think all this through, because the future is here.

Here’s a quick video from the Mayo Clinic, talking about AI in migraine treatment.