Bradley C Love

Funded by

Tweets
View on Twitter

Model-Based fMRI Workshop

Model-Based fMRI Workshop for Math Psych Folks (application deadline August 25th)

Dear All,

With the help of others, I hope to be organizing a workshop on Model-based fMRI this Summer in London, which will be timed to coincide with the Math Psych and CogSci conferences. The workshop is aimed at modelers in psychology who are interested in model-based fMRI. The format of the workshop will be unusual (details below) as it will be patterned on a www.kaggle.com competition in which participants compete on common challenge problems. The fMRI data will be heavily pre-processed and constrained such that no prior experience is required — it should be no more difficult than dealing with several cell means.

Are you (or someone you work with) interested? If so, please email the following ASAP:

  1. Your CV.
  2. Are you a Psychonomic Society member? Are you willing to join?

I aim for a balanced group in terms of career stage, gender, location, etc.

Should funding go through, lodging and travel will be covered for the 10-20 participants.

I apologise in advance if I get more interest than spots.

Best, Brad

Over the past decade, I’ve become excited about linking cognitive models from Psychology to brain data. This area is of increasing importance, but unfortunately many talented people from communities such as the Psychonomic Society are on the sidelines because of the joint expertise necessary to make a contribution.

For example, a postdoc trained in cognitive modeling would lack the background, knowledge, and skills necessary to collect and analyze raw fMRI data. This is a tremendous waste as junior academics with these skills are desperately needed in this burgeoning area of research. To break into this area and publish credible work would require a couple years of training and access to high quality data. Of course, once a young researcher published in this area, they would be an attractive collaborator and competitive for funding.

The aim of this workshop is to reduce this barrier of entry for junior researchers and select senior researchers seeking a change. The workshop will be organized as a competition (along the lines of www.kaggle.com) in which workshop participants (and anyone on the web) compete to see who has the best model-based analysis of several open science brain imaging datasets. These datasets will be heavily pre-processed and put in a form approachable to cognitive psychologists. I work in categorization and memory and have already secured several datasets that form a coherent theme. For example, a dataset might involve a plot of the hippocampus’s response to two types of stimuli over the course of learning. The challenge would be to specify the cognitive model that best accounted for the data (in a cross-validated sense).

The workshop itself would consist of participants presenting their results and ample discussion. Unlike most workshops, everyone will be tackling the same challenges so interaction and interest should be high. At least half the slots for participants will be allocated to junior researchers. Rather than only reward researchers with the best fitting model, discussion would also acknowledge researchers whose work was highly explanatory and advanced psychological theory. A special issue, such as at CABN, could be fascinating given the competitive and highly comparable nature of the contributions. We’d also aim for a symposium at the 2017 meeting for the Psychonomic Society.

The location is London in the Summer of 2017. The challenge problems and datasets will be posted on a dedicated website.

Funded by