Development of a High-Precision Temporal Interference Stimulation System with Multi-Device Synchronization and Concurrent EEG

Temporal interference stimulation (TIS) has emerged as a promising noninvasive approach for modulating deep brain structures by exploiting the interference of high-frequency electric fields. However, the practical deployment of TIS in neuromodulation experiments remains constrained by limitations in stimulation precision, channel isolation, compatibility with concurrent electrophysiological (EEG) recording, and coordination with multimodal stimulation paradigms. We present a high-precision and versatile TIS system designed to address these challenges through a unified hardware architecture within a single validated platform. The system achieves current output accuracy within ± 1% and frequency control precision better than ± 0.1% across all channels. In TIS mode, it exhibits strong channel independence, with carrier leakage below 1% and inter-channel frequency isolation exceeding 98.9%. Beyond conventional continuous stimulation, the platform supports programmable burst paradigms, including intermittent and continuous theta-burst stimulation (iTBS/cTBS). To enable artifact-minimized concurrent EEG recording, the system integrates a hardware-based high-pass filtering strategy that substantially suppresses stimulation-induced artifacts during EEG acquisition. In a proof-of-concept validation, EEG power spectral density recorded during TIS with high-pass filtering showed high similarity to Sham stimulation (Pearson's r = 0.97), while stimulation artifacts were reduced by up to 30-fold compared to unfiltered conditions. In addition, the system provides microsecond-level trigger timing synchronization across devices, achieving inter-device timing errors below 200 μs, thereby enabling precise cortex-nucleus co-stimulation with external modalities. Numerical simulations and saline phantom experiments further confirmed accurate and reproducible control of the induced electric field, with strong spatial agreement between simulated and measured fields (Pearson's r > 0.97 for all components). These results show that the proposed system can deliver precise TIS, support concurrent EEG recording, and provide synchronized triggering for multimodal stimulation experiments.