TSG-DDT: Time-Series Generative Denoising Diffusion Transformers
Marco Zurdo-Tabernero, Pablo Enrique-Guillem, Ángel Canal-Alonso, Guillermo Hernández, Angélica González-Arrieta, Juan Manuel Corchado · Bernstein Conference 2024
Deep Brain Stimulation (DBS) involves implanting electrodes in the brain to modulate neural activity, treating conditions like Epilepsy [1]. Using synthetic micro-electrode array (MEA) signals as stimuli in a DBS system enables electrical impulses that mimic natural brain activity, improving precision and effectiveness [2]. This can reduce side effects and enhance therapeutic outcomes [3]. This study introduces Time-Series Generative Denoising Diffusion Transformers (TSG-DDT), an architecture for generating high-fidelity synthetic MEA data [4]. TSG-DDT combines denoising diffusion models [5] w