P 232. Leaving the beaten track of TMS waveform restrictions: Concepts and prototype for a ‘convertible’ stimulator that can generate almost every type of existing and future waveform

Abstract

Introduction Waveforms are an important issue in magnetic stimulation. They provide the key to an important characteristic of neurons: their distinctive nonlinear dynamics. Different waveforms seem to have different activation sites and may allow selective stimulation. This is reflected for example by their different corticospinal D-and I-wave patterns ( Di Lazzaro, 2004 ). However, available pulse types are very restricted. In essence, the only established waveforms are monophasic and biphasic pulses. In addition to selective stimulation and neuromodulation, higher efficiency ( Goetz, 2012 ) could enable highly focal small coils to overcome thermal issues and provide high repetition rates. Objectives Current devices can only provide very distinct pulses for technological reasons. Even in alternative, more flexible approaches, the underlying circuitry is limited to a certain class of stimuli. Among these stimuli, only a subclass does not change the device state of charge and can therefore be generated repetitively. Our objective was to conceive and design a technology that is able to generate almost any waveform. Accordingly, the selection of waveforms would no longer depend on device capabilities and technology, but merely follow the user’s needs. Methods For this aim, we had to abandon all traditional concepts and designs for pulse generation in magnetic stimulation as well as other fields of science and engineering. We developed novel topologies that can handle the special requirements of magnetic stimulation pulses, namely high voltage, high current and high speed. The winning approach uses a dynamical reconfiguration of small energy storages: Metaphorically speaking, it uses a high number of small-voltage batteries that can be combined in any parallel and series connection—dynamically changeable for every time step—so as to exactly control the coil voltage accurately to the curve outlined by the desired waveform. The key components of the technology are no longer expensive high-voltage semiconductors, but inexpensive mass-produced devices used in consumer electronics. We designed a control strategy for the distributed concept and implemented a prototype for characterization of the concept. Results The system can generate almost any waveform, both existing and potential future pulse shapes. All waveforms can be generated with a recovery of the pulse energy from the coil, which has been known from biphasic stimulators. We achieved this even for classical monophasic waveforms, which can—except for rare research devices ( Schmid et al., 1993 )—provide high pulse repetition rates, although being preferred. The high losses prevent that. Furthermore, the generation of waveforms is not done by any mechanical reconfiguration, but controlled dynamically. Therefore, the system can even fundamentally change the waveform from pulse to pulse in a theta-burst or double-pulse protocol. Conclusion The generation of arbitrary waveforms is possible and has several advantages also for classical waveforms. Among others, this includes the ability to use high repetition rates for almost every waveform and profit from cost-effective mass-produced components.