Pain phenotyping has surfaced as an important task to enable personalised medicine in chronic pain populations. In this project we pursued two aims: (1) testing the potential of a new pain phenotyping tool and (2) the improvement of existing phenotyping approaches. To fulfil aim one, we explored the potential of the Bayesian estimates of the weight placed on somatic input and expectations to be used for phenotyping (studies one, two and three). In study one we tested the influence of these weights on pain outcomes (placebo analgesia) by exploring whether the weight placed on these two factors could be used to differentiate between placebo responders and non-responders. Participants were classified as placebo responders or non-responders based on their responses in an experimental placebo paradigm. Then, a Bayesian approach was used to extract the weight each group (responder/non-responder) placed on somatic input and expectations. Our results showed that the weight placed on different factors varied between placebo responders and non-responders supporting the potential use of these estimates for capturing inter-individual/group differences. In study two, to establish whether these weights represented stable traits, and therefore, had the potential to be used as a diagnosing tool, we evaluated their test-retest reliability. To do this, a group of healthy participants completed a cued pain paradigm twice, with a two-week interval between sessions. Then, the stability in scores across sessions was calculated. Results indicated that the weights placed on different factors have a good test-retest reliability. In study three, we explored the concurrent validity of the weights to ensure they represented what they were intended to measure (attentional focus to somatic and expectation factors). To do this, participants completed the same cued pain task as in study two whilst their brain activity was measured through EEG. Then, the different weights were correlated with brain activity at areas identified as relevant by previous research. The concurrent validity tests demonstrated that, as expected, somatic weight represented the tendency to direct attention to somatic sources of information, and expectation weight represented the tendency to direct attention to semantic sources. To fulfil aim two (improve existing pain phenotyping approaches) in study four we conducted a systematic literature search and critical evaluation of current chronic pain vulnerability models. Our results showcased that these models excel at their clarity and are supported by correlational evidence. However, they also present gaps in their lack of depth of aetiological factors. In study five we attempted to fill some of these gaps. To do so, we conducted a cross sectional study in which we measured potential aetiological factors (trauma during childhood and adulthood) as well as some key vulnerability variables identified by the models (pain catastrophizing, anxiety sensitivity) in a group of chronic pain patients. Our results identified emotional abuse as a potential aetiological factor to include in models of chronic pain vulnerability. In summary, through this project, we found support for the use of somatic input and expectation weight estimates as potential pain phenotyping variables and identified the need to include aetiological factors in current chronic pain vulnerability models as a way to improve the phenotyping through psychological variables.
Pain phenotyping based on somatic input, expectations and psychological characteristics
Delgado Sanchez, A. (Author). 31 Dec 2023
Student thesis: Phd