Related Experiment Video
Updated: Dec 6, 2025
Neuronavigated Focalized Transcranial Direct Current Stimulation Administered During Functional Magnetic Resonance Imaging
Published on: November 15, 2024
Machine learning and individual variability in electric field characteristics predict tDCS treatment response
Alejandro Albizu1, Ruogu Fang2, Aprinda Indahlastari3
1Center for Cognitive Aging and Memory, McKnight Brain Institute, University of Florida, Gainesville, USA; Department of Neuroscience, College of Medicine, University of Florida, Gainesville, USA.
Machine learning accurately predicts brain stimulation response. This study shows individual differences in transcranial direct current stimulation (tDCS) current intensity and direction are key to improving cognitive function in older adults.
Area of Science:
- Neuroscience
- Cognitive Science
- Medical Imaging
Background:
- Transcranial direct current stimulation (tDCS) is explored for cognitive enhancement in older adults.
- Individual variability in brain electrical current delivery necessitates personalized approaches.
- Predicting tDCS response can optimize cognitive intervention outcomes.
Purpose of the Study:
- To use machine learning and MRI-derived models to predict working memory improvements from tDCS.
- To establish a proof of concept for precision cognitive intervention.
- To investigate individual differences in tDCS current influencing cognitive outcomes.
Main Methods:
- Fourteen healthy older adults received 2mA tDCS for 20 minutes during cognitive training.
- Working memory was assessed using an N-back task pre- and post-intervention.
- MRI-derived current models and Support Vector Machine (SVM) algorithms analyzed tDCS current components (intensity, direction).
Main Results:
- SVM models achieved 86% accuracy in classifying tDCS responders versus non-responders.
- Current intensity was the most significant factor in predicting working memory changes.
- Median current intensity and direction near electrodes positively correlated with intervention response (r=0.811, p<0.001 and r=0.774, p=0.001).
Conclusions:
- Pattern recognition of MRI-derived tDCS models accurately predicts individual treatment response (86% accuracy).
- Individual differences in current intensity and direction are critical for tDCS efficacy.
- Findings support precision dosing models for tDCS interventions and offer insights into response mechanisms.
