I was reading the article Symptoms perception, placebo effects and the Bayesian brain by Ongaro and Kaptchuk, and I am having some difficulty sussing out what exactly the cliff notes are on the Bayes theorem and how it ties into the brain and its predictions. I watch a video or 5 on Bayes theorem and I understand at a basic level the equation and its use for prediction. From my reading in this article I took away from it that:
the brain makes hypotheses or predictions about what will be felt based on past experience
sensory input and contextual clue are sent âupâ to the brain from the external world
the brain attempts to create a perception for the individual based on the two
the brain tries as best as it can to minimize error between the 2 and 3 above to create an accurate prediction of the world
I can understand that if there is a new input (possible acute traumatic injury) that the hypothesis and prediction cannot be I guess correct or close enough to match this sensory input and thus we feel mechanical based pain, and that if there is a strong (not sure if the right use of adjective) prediction that the body will shape the sensory input match the prediction thus again minimizing the error.
While I have gotten all of this, I feel as though there are a ton of holes and things that I am not putting together in as far as a solid understanding of this and its application to pain and practice go. Do I need to take a look at more videos or descriptions on the theorem and re-read this article a couple more times or is there a better way to make this click in my head.
This was such an interesting article! The way I understood it was that the brain is making predictions all the time based on current environment, past experience and contextual factors but it can get it wrong too. According to the article, we have pain because we predict we have pain. This can be accurate or an error. For me, it lines up with Moseley and Butlerâs idea of danger vs safety in me (DIMS and SIMS). Pain is expressed when the brain decides there is more danger than safety. This can be actual damage or perceived damage. I would look at pain with perceived damage as a prediction error. The brain predicted something was wrong when it actually wasnât. What I found far more interesting was the idea of conditioning to treatments and responses. Weâve talked about this for awhile but the way the article explained it really clicked with me. How many times have patients come in saying they were in pain but after they made the appointment, the pain went away. Thatâs happened to me personally. I made an appointment and by the time it came around, I wasnât in pain. I look at this as the brain predicting that I would be better because I was seeing a trusted practitioner and getting a treatment I deemed helpful. I conditioned myself to predict that I would get better when I saw that person and got that treatment.
Hereâs what I found interesting in the new paper I linked. They state that ârealâ treatment and placebo both work off the same inferential process for symptom relief so both are seen equally real from the patientâs point of view. They go on to suggest that to be truly patient centered, we need to address these predictive processes to determine the best course of action to lead the brain to predict the bodyâs health. If the brain predicts the body is healthy, it is less likely to predict pain.
The question is do we try to change the patientâs point of view that the placebo isnât real or let them continue to think it is? The argument can go in both directions. If it doesnât harm the patient, then let them believe. Where do we draw the line here? Manual therapy, supplements, surgery, lying to them just to get their money? We all know clinicians who say as long as their patient gets better, they donât care but many of them seem to not address their own biases and take full credit for the improvement. Is it better to realize all the complexities surrounding pain and treatment but only address certain nocebos etc with patients?