Predictive Analytics Competition 2019 Reward Session #10
Description
Predictive Analytics Competition 2019
PAC 2019 Website
This year, we invite teams from all over thte world to develop a model predicting brain age from healthy individuals based on structural Magnetic Resonance Imaging (sMRI) data (n > 3000) .
All teams are asked to engineer a machine learning model using their expertise and either raw nifti data and or analogous to Cole et al. fully pre-processed data as described in the paper. Each team has the opportunity to upload predictions for a given testset and we will evaluate model performance by comparing the uploaded predictions to actual chronological age for each individual. This year, there are two distinct objectives:
-
The team submitting the model with the smallest Mean Absolute Error for the test dataset will win this year’s PAC Best Model Award.
-
The team submitting the model with the smallest Mean Absolute Error for the test dataset while keeping the Spearman correlation between brain predicted age difference and chronological age below r = .10 will win this year’s PAC Bias Reduction Award.
As we find the resulting performances to be an interesting benchmark as well as the chosen approaches to be exciting oppurtunities for inspiration, we would like to finish the Predictive Analytics Competition with a reward session at the OHBM 2019 in Rome. In addition, both the number of participants ( right now being > 200 person in > 60 teams) as well as the number of training data are extraordinary so that we expect the outcome to be of major interest for the general neuromaging community.
Main organiser
Tim Hahn
Preferred time slots
Any
Tag other attendees
James Cole, Christian Gaser, Ramona Leenings, Nils Winter,
Open to all?
Yes
Additional Context
The brain changes as we age, and these changes are associated with cognitive decline, neurodegenerative disease and dementia. Although brain ageing is universal, rates of brain ageing differ markedly; some people suffer cognitive decline in later middle-adulthood, while others remains cognitively normal into their tenth decade. The process of brain ageing includes morphological and functional changes to the brain, which can be assessed using neuroimaging. This raises the possibility that the variability in brain ageing can be measured, and research has focused on developing such a neuroimaging biomarker of brain ageing; the so-called ‘brain-age’ paradigm.
The idea with brain-age is that if statistical models can be developed to accurately predict chronological age in healthy people (using neuroimaging data), then the apparent age of a new individual’s brain can be calculated. Where someone’s brain-age is older than their real age, this is thought to reflect poorer brain health, relative to their age. Older-appearing brains have been associated with psychiatric and neurological diseases, with greater risk of developing dementia and a shorter lifespan. Younger-appearing brains have been found in people who exercise more, have greater years of education, meditate or play musical instruments.
The hope is that brain-age can provide a sensitive, if unspecific, global measure of brain health, that could be used in many contexts. These include clinical trials of neuroprotective therapies, screening groups of people at-risk of poorer cognitive ageing, and providing mechanistic insights into the downstream consequences of different diseases.
Critical to the success of brain-age models, is the accuracy of the healthy training model. Hence, the goal of this year’s PAC is to build the most accurate model, using the training data supplied. Specifically, we would like to minimize brain predicted age difference (brain-PAD) which is calculated as brain-predicted age minus chronological age.