How Much is Brain Data Worth for Machine Learning?

2026-05-09

AI Summary

This study addresses the value of neuroimaging data in enhancing machine learning model performance and delineates the conditions under which collecting such data is justified. By constructing a linear Gaussian multimodal model, the authors theoretically analyze an estimator that fuses neural recordings with task labels, deriving scaling laws that characterize performance as a function of sample size. They establish, for the first time, an equivalence ratio quantifying the exchangeability between neural and task data. The work further elucidates how brain data improves model robustness under distribution shift, demonstrating that its utility hinges on task–brain alignment, noise levels, latent dimensionality, and sample size. These insights collectively define the regimes in which incorporating neural data is advantageous under a fixed data acquisition budget.

Abstract

If a person can solve a task, can measuring their brain make it easier to train a model to solve that task too? Recent NeuroAI work suggests that supplementing task training with neural recordings can modestly improve model performance and robustness. However, it is unclear when there should be a benefit from using neural data and how much benefit to expect. We formulate this question mathematically, and begin to address it theoretically using a simple, analytically tractable linear gaussian model of task targets and neural recordings. For a multimodal estimator trained on both brain data and task labels, we derive scaling laws for how performance scales with the numbers of brain and task samples. From these laws we derive relative value and exchange rates between brain samples and task samples, quantifying how much extra task samples neural data is worth as a function of task-brain alignment, neural and task noise, latent dimension, and brain data sample size. We also analyze test distribution shift, to identify conditions where brain-regularized learning can produce substantial robustness gains through learned invariances. Finally, under a fixed collection budget, we characterize the regimes in which brain data is worth collecting. Our results provide a foundation for understanding how valuable brain data could be for improving machine learning.

Problem

• NeuroAI
• brain data value
• machine learning
• scaling laws
• distribution shift

Innovation

• NeuroAI
• scaling laws
• brain-data value
• multimodal learning
• distributional robustness