Precision Neuromodulation with Real-Time Brain Decoding for Working Memory Enhancement.

Area of Science:

• Neuroscience
• Cognitive Science
• Biomedical Engineering

Background:

• Transcranial magnetic stimulation (TMS) offers non-invasive brain therapy but suffers from variable outcomes.
• Inter-individual differences in brain function are a likely cause of TMS variability.
• Personalized functional networks (PFNs) derived from fMRI offer a potential solution for targeted neuromodulation.

Purpose of the Study:

• To integrate personalized functional networks (PFNs) with a neural network decoder for real-time optimization of TMS.
• To enhance working memory (WM) task performance using closed-loop neuromodulation.
• To investigate the predictive capability of brain decoders for behavioral outcomes.

Main Methods:

• Functional magnetic resonance imaging (fMRI) was used to derive personalized functional networks (PFNs).
• A neural network-based decoder was employed to optimize transcranial magnetic stimulation (TMS) in real-time during a working memory task.
• Participants underwent randomized crossover sessions with optimal and suboptimal stimulation frequencies.

Main Results:

• Optimal TMS, guided by the neural network decoder, significantly improved working memory performance compared to suboptimal stimulation.
• No significant improvement was observed in a control task, indicating task-specific effects.
• The brain decoder's output accurately predicted behavioral performance during both TMS/fMRI and neuromodulation sessions.

Conclusions:

• Closed-loop TMS, guided by real-time brain decoding, enhances post-stimulation behavioral effects.
• Individualized functional connectivity networks are crucial for targeted neuromodulation.
• This approach represents a significant advancement in personalized brain stimulation for cognitive enhancement.