Phase-specific stimulation of the human brain with real-time measurement instead of prediction

BackgroundThe responsiveness of the human brain to external input fluctuates. Timing the external perturbation with regard to the oscillatory brain state may improve the intended stimulation effects. However, current brain state-dependent interventions targeting phases of the oscillatory cycle need to apply prediction algorithms to compensate for latencies between measurement and stimulation, and are therefore imprecise. ObjectiveWe investigated the phase-specific precision of a novel non-predictive approach on the basis of integrated real-time measurement and brain stimulation. MethodsApplying a simulation, we estimated the circular standard deviation (SD) to hit 2, 4, 8, 16 or 32 equidistant phase bins of the oscillatory cycle with high precision. Furthermore, we used electroencephalography-triggered transcranial magnetic stimulation in healthy subjects to empirically determine the precision of hitting the targeted phase of the oscillatory cycle for 10 different frequencies from 4Hz to 40Hz using our approach. ResultsThe simulation revealed that SDs of less than 17.6{degrees}, 9.7{degrees}, 5.1{degrees}, 2.5{degrees}, and 1.3{degrees} were necessary to precisely hit 2, 4, 8, 16, and 32 distinct phase bins of the oscillatory cycle. By completing measurement, signal-processing and stimulation with a round-time of 1ms, our empirical approach achieved SDs of 0.4{degrees} at 4Hz to 4.3{degrees} at 40Hz. This facilitates selective targeting of 32 phases (at 4Hz), 16 phases (at 8, 12, 16, 20, 24Hz) and 8 phases (at 28, 32, 36, 40Hz), respectively. ConclusionIntegrated real-time measurement and stimulation circumvents the need for prediction and results in more precise phase-specific brain stimulation than with state-of-the-art procedures.