Wednesday, October 17, 2018

Ongaro and Kaptchuk 2.1

This blogpost deals with the first reference in the introduction of the paper I'm currently reading and absorbing (Ongaro, Giulio, Kaptchuk, Ted J., 2018, Symptom perception, placebo effects, and the Bayesian brain. PAIN, Aug 6, 2018. (To read the full paper, click on "Article as pdf").

The references are cited in the last sentence of the intro:

"It expands on recent high-quality empirical work on predictive processing (1,7,19,24) and outlines, more broadly, how Bayesian models offer an altogether different picture of how the brain perceives symptoms and relief."

The first reference is 
Anchisi D, Zanon MA. Bayesian perspective on sensory and cognitive integration in pain perception and placebo analgesia. PLoS One 2015. First, the authors
Davide Anchisi and Marco Zanon are both in Italy. Italy is a hotbed of placebo research (Fabrizio Benedetti* is in Italy as well). Both authors are at Department of Medical and Biological Sciences, Universit√† degli Studi di Udine, Udine, Italy, according to PubMed, although Zanon may be in Bologna now.
Here is a 
list of publications by Anchisi;  Anchisi is an MD and PhD.Here is a list of publications by Zanon;  Zanon is a post-doc research fellow with the psychology department. 


1. "How can an inert treatment cause a response?"
This is the first sentence. Good question. My follow-up question is, who decided that any given treatment was "inert"? What were their criteria? They certainly left out a lot of info or maybe their info was incomplete. Right? An awake alert human brain is always going to respond to context of said treatment, so nothing can ever truly be "inert." But I digress.  

2. "... living organisms and their perceptive systems deal with uncertainty and face transposed conditional probability problems: they have to infer the features of activating stimuli (related to the state of the world) from nervous signals elicited by those stimuli, and come to a reliable perception (i.e., make an effective decision) in spite of the noisy and incomplete information that sensory signals provide about the world."
Yup, that's pretty much what the nervous system is doing from the moment it switches itself on as an embryo to the day it dies.

3. ".. experimental and theoretical work showed that perception is a multisensory task and support the hypothesis that, in perception, different pieces of information are near optimally combined in a Bayesian way."
Note to self: learn whatever I can grok about "Bayesian." This link states: 

"Bayesian statistics is a system for describing epistemological uncertainty using the mathematical language of probability. In the 'Bayesian paradigm,' degrees of belief in states of nature are specified; these are non-negative, and the total belief in all states of nature is fixed to be one. Bayesian statistical methods start with existing 'prior' beliefs, and update these using data to give 'posterior' beliefs, which may be used as the basis for inferential decisions."
Apparently there was a guy in 1763, Thomas Bayes, who came up with this.  

4. "
In this study we focused on the placebo effect because it is one of the best examples of experimentally controllable modulation of pain experience, and has been extensively investigated in recent decades." OK...

5. "Our aim was to develop a Bayesian framework which could describe and explain pain perception and its modulation....As we argue here, the nociceptive signals are not the only source of information used to compute the inference: past experience and cognitive information also play a role."
As Melzack's Neuromatrix framework states.

6. "
A total of 55 healthy human volunteers (mean age ± s.d.: 21.40 ± 1.03; 29 females) were recruited by advertising at the University of Udine (Italy) and randomly divided into two groups: Experiment 1 (n = 24); and Experiment 2 (n = 31)." And then they "conditioned" the subjects (gave a bunch of twentyish-year-olds the opportunity to organize their expectations), and applied TENS.

7. "
The main findings of this study concern not only the placebo effect but a wider range of effects also due to past experience." I wonder what results would look like if the subjects were in their 60's or 70's? Just a thought.

8. "
The Bayesian decision model we developed comprises three key elements: the prior probability, which conveys previous experiences and expectancy (e.g. through information derived from the context); the likelihood function, which implements the sensory inputs and also information from multiple sources (whether sensory, cognitive or psychological); and the decision process, which eventually determines if and to what extent pain is perceived."

9. "
The model we developed shows how the placebo effect results from the evaluation and integration of nociceptive stimuli with context information, and how the relevance of the context (through expectation/conditioning) comes, at least in part, from past experience. The process of information integration would be at the very base of pain perception, and would lead to the placebo effect and to other phenomena such as those predicted by the fBD model.

Overall, our findings support the hypothesis that pain perception can be described according to the rules of Bayesian probabilistic reasoning."

Ongaro and Kaptchuk 1: Intro 
Ongaro and Kaptchuk 2: Intro

*I spent a great deal of time with Benedetti's book, The Patient's Brain, read it cover to cover, made lots of notes; I also had the pleasure of hearing him speak about his work at the San Diego Pain Summit in 2016.

No comments: