Phase-targeted modulation of essential tremor with transcranial magnetic stimulation of motor cortex

Neural oscillations provide temporal frameworks for coordinating communication within and across distributed brain networks. In essential tremor (ET), pathological synchronization within the cerebello-thalamo-cortical circuit produces rhythmic activity that manifests as an involuntary action tremor. Although deep brain stimulation can effectively suppress tremor, its invasiveness and cost highlight the need for non-invasive interventions capable of selectively modulating pathological oscillations. Transcranial magnetic stimulation (TMS) offers a non-invasive means of engaging cortical circuits, yet conventional stimulation protocols are delivered independently of the ongoing neural dynamics. Such open-loop approaches ignore the temporal structure of tremor-related activity, potentially stimulating during both amplifying and suppressing phases of the oscillation. To address this, we compared two phase-targeted TMS paradigms: first-pulse phase-locked TMS (First-pulse-TMS), in which only the initial pulse of a stimulation train is aligned to the tremor phase, and cycle-by-cycle phase-locked TMS (Continuous-TMS), in which each pulse is continuously triggered based on real-time tremor phase. Ten patients with ET underwent stimulation guided by peripheral tremor recordings using an accelerometer, with tremor phase estimated in real time via the Oscilltrack algorithm. Sixty-four trains of TMS pulses were delivered at nine discrete phase bins of the tremor cycle, such that each phase bin was repeated approximately seven times. Continuous-TMS maintained accurate phase-locking across consecutive cycles (mean phase-locking value ~0.9), whereas First-pulse-TMS exhibited progressive drift over time and low phase consistency (mean phase-locking value <0.2). The circular concentration of stimulation phase was significantly greater for Continuous-TMS than First-pulse-TMS (Mann-Whitney U-test, p < 0.001), indicating a significant difference in overall phase-locking accuracy between the two protocols. Critically, Continuous-TMS, unlike First-pulse-TMS, induced bidirectional, phase-dependent modulation of tremor amplitude. Circular-linear modelling revealed a sinusoidal relationship between stimulation phase and changes in tremor amplitude, with tremor amplification and suppression occurring at opposite phases of the cycle. Covariates including baseline tremor amplitude and trial number were accounted for. In some people, tremor suppression outlasted the stimulation period, suggesting phase-locked TMS may be a potentially useful therapeutic tool. By enabling reliable, phase-specific stimulation of the tremor cycle, Continuous-TMS allows identification of the individual phase that produces maximal tremor suppression, supporting the development of personalized, phase-specific neuromodulation strategies. This proof-of-principle study demonstrates that temporally precise, closed-loop TMS can interact with pathological oscillations in real time, providing a mechanistic framework for probing oscillatory contributions to motor symptoms and a scalable therapeutic approach for ET and other oscillopathies.