Towards tDCS Digital Twins using Deep Learning-based Direct Estimation of Personalized Electrical Field Maps from...

Area of Science:

• Neuroscience
• Computational Neuroscience
• Medical Imaging

Background:

• Transcranial Direct Current Stimulation (tDCS) shows promise for various neurological and psychiatric conditions, but inconsistent results hinder progress.
• Current methods for estimating tDCS current flow are often slow, error-prone, and require significant computational resources.
• Individualized current field estimation is crucial for optimizing tDCS efficacy and safety.

Purpose of the Study:

• To develop a proof-of-concept for generating individualized tDCS electrical field maps directly from T1-weighted MRI scans.
• To create a computationally efficient deep learning pipeline for personalized tDCS dose estimation.
• To improve the generalizability and calibration of electrical field modeling for tDCS.

Main Methods:

• A deep learning model was trained on a dataset of 442 unique T1-weighted MRIs.
• The model incorporates specialized loss regularizations to enhance performance across diverse individuals and electrode montages.
• The pipeline accepts unique MRI data and user-defined electrode montages for custom output generation.

Main Results:

• The developed pipeline generates personalized tDCS electrical field maps in minutes, a significant improvement over traditional physics-based methods (1-3 hours).
• The deep learning approach demonstrated improved generalizability and calibration for individual brain scans and electrode configurations.
• The system successfully processed a large dataset encompassing individuals across the adult lifespan.

Conclusions:

• This novel deep learning approach offers a rapid and accurate method for generating individualized tDCS electrical field maps.
• The streamlined process facilitates personalized current dose estimations, potentially enhancing the effectiveness of tDCS interventions.
• This work paves the way for more accessible and reliable application of tDCS in clinical and research settings.