Stopping criteria for boosting automatic experimental design using real-time fMRI with Bayesian optimization
This work addresses a specific bottleneck in closed-loop neuroimaging for researchers, but it is incremental as it builds on existing Bayesian optimization frameworks.
The paper tackles the challenge of determining when to stop Bayesian optimization in real-time fMRI experiments, proposing and evaluating two stopping criteria to address high scanning costs and limited subject attention.
Bayesian optimization has been proposed as a practical and efficient tool through which to tune parameters in many difficult settings. Recently, such techniques have been combined with real-time fMRI to propose a novel framework which turns on its head the conventional functional neuroimaging approach. This closed-loop method automatically designs the optimal experiment to evoke a desired target brain pattern. One of the challenges associated with extending such methods to real-time brain imaging is the need for adequate stopping criteria, an aspect of Bayesian optimization which has received limited attention. In light of high scanning costs and limited attentional capacities of subjects an accurate and reliable stopping criteria is essential. In order to address this issue we propose and empirically study the performance of two stopping criteria.