THE EMOTIONAL CONTEXT MAINTENANCE AND RETRIEVAL MODEL (ECMR)
The most important experiences of our life are often emotionally intense – be they positive (successes, family celebrations) or negative (failures, funerals). More mundane experiences in our daily life often also harbour an emotional resonance. These tend to be the experiences we remember the best.
eCMR is a cognitive-computational model that aims to account for our enhanced ability to recall emotionally-evocative experiences.
In the lab, researchers typically study this phenomenon by using complex, novel pictures. Pictures have an incredible power to evoke strong emotions. eCMR explains why we remember emotionally-evocative images so well.
Emotional pictures are prevalent in news reports and films, as well as in the pictures we take ourselves to preserve the experiences that are important for us (holidays, births, weddings). They are also used extensively for many other purposes. For example, the Extinction Rebellion marches have featured many images that have penetrated the national imagination. The picture above depicts a “funeral procession” to highlight all the species that have gone extinct (© Tony Gussin).
A short, accessible explanation of how eCMR accounts for memory for emotional pictures is available in an article published in the BPS Cognitive Psychology Bulletin (2020). For the complete account see the article in Psychological Review. Please see bioRxiv for the preprint.
For a trove of publications on retrieved context theories and useful code for the analysis of free recall please see the Computational Memory Lab website.
BAYESIAN MODEL OF SUBJECTIVE PAIN EXPERIENCE
Dan Acosta-Kane developed the models used in Hoskin et al. (2019) Cognition paper. He has now revised them to make them more accessible.
To Run the code:
devtools::install_github("dan-ak/bayesian-pain")
library(BayesianPain)
data_loc <- system.file("extdata", "exp_1_data.csv", package="BayesianPain")
exp_data <- BayesianPain::center_scale(data.frame(read.csv( data_loc , header = TRUE)))
stan_data <- BayesianPain::create_stan_data(exp_data)
models <- BayesianPain::get_models(c(1,3,4,5,6))
output <- BayesianPain::fit_models(models, stan_data, iter=700, chains = 4)
output['DICs']
BayesianPain::get_parameters(output['fits'])