Mechanistic investigation of Temporal Interference brain stimulation and its translational application

Abstract

Temporal Interference (TI) is a recently proposed non-invasive deep brain stimulation technique. It uses two kHz electric fields that differ by a small frequency difference, Δf, producing amplitude-modulated signal where the fields overlap. Early work showed that neurons responded to the Δf oscillation but filtered out the kHz fields, suggesting TI could achieve targeted stimulation away from the electrodes. The mechanism of action of TI is not completely understood and has so far been primarily investigated at the single-cell level. In this thesis, I extend the mechanistic investigation by characterising TI-evoked network-level neuronal activity in vivo. Using widefield calcium imaging, I mapped, for the first time, the direct spatiotemporal effects of TI relative to control, low frequency stimulation. My results highlight important differences between these conditions and point towards the potential contribution of network-level interactions to the effects observed. At the spatial level, responses measured were in agreement with the existing finite element method models. At the temporal level, TI, surprisingly, induced contralateral activation and a secondary harmonic response at 2Δf, suggesting possible recruitment of different cell types. I also found that pure kHz fields can induce transient onset responses, which can be further mitigated, but did not lead to a depolarisation block. Having established the network-level characteristics of TI, I investigated the therapeutic potential of different frequency protocols to target circuit dysfunction in the AppNL-G-F model of Alzheimer’s Disease. Two pilot studies performed suggest that TI may be able to affect relevant disease biomarkers. Firstly, stimulation in the delta range transiently reduced cortical excitability. Secondly, theta stimulation increased hippocampal neurogenesis markers and, potentially, microglial presence after eight days of repeated stimulation in freely moving mice. Overall, these results suggest that TI can effectively drive network-level neuronal dynamics and potentially correct aberrant activity patterns, highlighting its usefulness for translational applications.