Background & Aims
To implement machine learning (ML) algorithms to predict the response to non-invasive brain stimulation and mirror therapy in patients with phantom limb pain (PLP).
Methods
This is a secondary analysis of a randomized controlled trial. This was a randomized, blinded, sham-controlled, 2 × 2 factorial trial. 112 participants with traumatic lower limb amputation were randomized into treatment groups. The interventions were active or covered MT for four weeks (20 sessions, 15 minutes each) combined with t weeks of either active or sham tDCS (10 sessions, 20 minutes each) applied to the contralateral primary motor cortex. The predicted outcome was an improvement of 50% in PLP levels after treatments. Clinical and motor cortex excitability variables were included as predictors. An ML approach of feature engineering, data pre-processing, and model optimization was used to create the most accurate predictive model. We performed a cross-validation process (80% data for training, 20% data for testing). We trained multiple ML classifiers: Support Vector Machine, Random Forest, Naive Bayes, and Multilayer Perceptron (MLP). We used the AUC of ROC for model comparison.
Results
The MLP was the most accurate classifier for all treatments (active treatments and sham treatments). In the test data, the AUC was higher for the prediction M1 tDCS response (0.71), followed by mirror therapy response (0.59). The important features to predict response to M1 tDCS were anxiety levels and intracortical inhibition. On the other hand, mirror therapy response was predicted by phantom limb sensation intensity and age.
Conclusions
Artificial Intelligence is a potential tool to optimize non-invasive brain stimulation (NIBS) treatments. The M1 cortical excitability biomarkers seem to predict tDCS response, and phantom limb sensations seem to predict MT response. There is a need for validation with new cohorts and larger sample sizes.
References
1. Pacheco-Barrios K, Meng X, Fregni F. Neuromodulation techniques in phantom limb pain: a systematic review and meta-analysis. Pain Medicine. 2020 Oct;21(10):2310-22.
2. Münger M, Pinto CB, Pacheco?Barrios K, Duarte D, Enes Gunduz M, Simis M, Battistella LR, Fregni F. Protective and risk factors for phantom limb pain and residual limb pain severity. Pain Practice. 2020 Jul;20(6):578-87.
3. Pacheco-Barrios K, Pinto CB, Velez FS, Duarte D, Gunduz ME, Simis M, Gianlorenco AL, Barouh JL, Crandell D, Guidetti M, Battistella L. Structural and functional motor cortex asymmetry in unilateral lower limb amputation with phantom limb pain. Clinical Neurophysiology. 2020 Oct 1;131(10):2375-82.
4. Gunduz ME, Pacheco-Barrios K, Bonin Pinto C, Duarte D, Vélez FG, Gianlorenco AC, Teixeira PE, Giannoni-Luza S, Crandell D, Battistella LR, Simis M. Effects of combined and alone transcranial motor cortex stimulation and mirror therapy in phantom limb pain: a randomized factorial trial. Neurorehabilitation and neural repair. 2021 Aug;35(8):704-16.
Presenting Author
Kevin Pacheco-Barrios
Poster Authors
Topics
- Specific Pain Conditions/Pain in Specific Populations: Neuropathic Pain - Central