Advancements in Real-Time Neurotechnology: Applications in Neurorehabilitation

rt-fMRI neurofeedback produced measurable motor improvements in a subset of patients with cerebellar ataxia using a noninvasive, outpatient-compatible protocol. The approach paired motor imagery with real-time decoding to engage motor circuits, and the feasibility data suggest promise for individualized, clinic-based motor rehabilitation.

In this feasibility study, 16 participants with cerebellar ataxia completed finger-tapping assessments before and after rt-fMRI neurofeedback combined with motor imagery to evaluate accuracy and precision. Investigators calculated tapping error (root mean square error) at 1 Hz and 4 Hz and trained whole-brain SVM classifiers on overt tapping to provide online feedback during imagery.

Although group-level comparisons did not reach statistical significance, 9 of 16 individuals showed meaningful post-training improvements in tapping accuracy, and imagery performance exceeded chance across runs. The degree of motor gain correlated with successful activation of motor-related regions during imagery, pointing to participant-specific benefit rather than a uniform group effect.

Frontal, basal ganglia, and cerebellar activations were most strongly associated with improved tapping precision; insular and parietal activity also covaried with individual gains. Because these activations tracked measurable reductions in tapping error, they identify mechanistic targets that could guide noninvasive interventions.

Key Takeaways:

• rt-fMRI neurofeedback paired with motor imagery is feasible and produced individual motor improvements in a subset of people with cerebellar ataxia.
• Changes in frontal, basal ganglia, cerebellar, and insular activity correlated with motor gains and may inform noninvasive neuromodulation targets.
• Translation to outpatient or wearable applications will require reductions in latency driven by advances in real-time machine learning and engineering.