The blood oxygen level dependent (BOLD) signal as measured by functional magnetic resonance imaging (fMRI) has become a standard marker of neural activity. The relationship between neural activity and its hemodynamic response is characterized by complex spatiotemporal dynamics. Most analyses rely on simple models of neurovascular coupling and thus underestimate the neural information in the BOLD signal. We use machine learning methods to obtain a model free estimate of the neural information in BOLD contrast data from simultaneous recordings of high resolution fMRI signals and intracranially measured neural bandpower signals. These estimates can help to guide parameter selection in fMRI studies for optimal decoding of stimulus information and might serve as a baseline to which studies without intracranial neural measurements can be compared.