Neurodesign-plus: design optimization for constrained event-level fMRI paradigms

Task-based functional magnetic resonance imaging (fMRI) experiments are implemented by scheduling events of interest. The schedule specifies their order, durations, and intervening intervals. Together with the acquisition and first-level model specifications, these choices determine the design matrices used for analysis and therefore how precisely task-related blood-oxygen-level-dependent (BOLD) responses can be detected or estimated. Constructing an appropriate schedule is therefore a crucial part of task design, requiring trade-offs among statistical efficiency, sequence predictability, and scan duration. Neurodesign searches for efficient schedules by scoring candidates on four optimality criteria and using genetic-algorithm or simulation-based methods to identify high-scoring designs. However, Neurodesign represents each trial as a single stimulus event with global duration and pre-event, post-event, and inter-trial timing parameters. It therefore cannot directly encode trials containing multiple ordered events or intervals whose durations depend on the events they separate. Neurodesign-plus is a maintained fork that extends the optimization framework to support multi-event trials, probabilistic templates, transition-dependent intervals, and fixed event orders and durations. These additions widen the space of schedules that researchers can generate, represent, and score, thereby improving task-fMRI design optimization.