Measures of brain activity through functional magnetic resonance imaging (fMRI) or Electroencephalography (EEG), two complementary modalities, are ground solutions in the context of neuro-feedback (NF) mechanisms for brain rehabilitation protocols. While NF-EEG (real-time neurofeedback scores computed from EEG signals) have been explored for a very long time, NF-fMRI (real-time neurofeedback scores computed from fMRI signals) appeared more recently and provides more robust results and more specific brain training.Using simultaneously fMRI and EEG for bi-modal neurofeedback sessions (NF-EEG-fMRI, realtime neurofeedback scores computed from fMRI and EEG) is very promising to devise brain rehabilitation protocols. However, fMRI is cumbersome and more exhausting for patients. The original contribution of this paper concerns the prediction of bi-modal NF scores from EEG recordings only, using a training phase where EEG signals as well as the NF-EEG and NF-fMRI scores are available. We propose a sparse regression model able to exploit EEG only to predict NF-fMRI or NF-EEG-fMRI in motor imagery tasks. We compare different NF-predictors steaming from the proposed model. We show that a proposed NF-predictor significantly improves the quality of NF session, over what EEG can provide alone, and correlates at 0.74 in median with the ground-truth.