Deep brain stimulation (DBS) has emerged as a widely used treatment for various neurological disorders, in particular for movement disorders such as Parkinson’s disease (PD) (Deuschl et al., 2006). While improvements of PD motor symptoms are often observed within minutes to a few days (Castaño-Candamil et al., 2019), DBS effects in dystonia are far more delayed (Kupsch et al., 2006). Likewise, its efficacy for diseases primarily characterized by non-motor symptoms, such as depression, anxiety, and Alzheimer’s disease, is the subject of ongoing research (Kisely et al., 2018, Laxton et al., 2010). Currently, the major roadblock preventing optimized patient-specific stimulation protocols is the limited understanding of the mechanisms of action. This applies in particular when immediate responses are not observable. In this context, it is crucial to consider all potential uncertainties in DBS applications. An accurately manufactured electrode plays a key role, particularly in experiments using custom-made electrodes, where manufacturing uncertainties are often not considered.
To better understand these mechanisms and to refine DBS applications, rodent models are widely employed to study its use in treating various diseases and developing new therapeutic approaches. For example, the use of DBS for PD has been investigated in rats (Campos et al., 2020, Ruiz et al., 2022, Fauser et al., 2024), DBS for dystonia in hamsters (Lüttig et al., 2024), the regulation of inflammation in mice (Falvey et al., 2024). Further examples can be found in Zhang et al. (2024). Due to the size of the rodent’s brain, customized electrodes are usually used, introducing an additional source of uncertainty. However, a consistent evaluation of the impact of uncertainties of the electrode geometry is usually not conducted. While the electrode geometry can be straightforwardly integrated into numerical models and its impact on the field distribution in the brain can be elucidated (Gimsa et al., 2006), it has not yet become common practice to use experimentally validated geometries in the models neither in humans nor in rodents.
Computational models play a pivotal role in refining the understanding of the mechanism of interaction in DBS. By predicting stimulation outcomes, the models shall unravel the underlying mechanics of DBS. Advanced biophysical models and neuroimaging methods have unlocked patient-specific computational models (Butenko et al., 2020, Neudorfer et al., 2023). Yet, the models rely on many assumptions and are usually not validated against experimental data. Hence, their reliability and predictive power are limited. The main challenge towards validation is the limited accessibility of the stimulation system before and after implantation. Thus, relatively simple methods have to be used. For example, the impedance at 1 kHz is often considered a quality control measure for stimulation electrodes, despite neglecting the strong frequency dependence of the impedance, which can lead to changes in orders of magnitude (Evers et al., 2022).
In this work, we suggest a combination of microscopy and impedance spectroscopy to implement a validated geometry model of the stimulation electrodes. We investigated microelectrodes for DBS in rodents in an in vitro setting. Consequently, we established a volume conductor model (VCM) in a rat model to simulate the electric field and volume of tissue activated (VTA) during DBS. Our results suggest that using single frequency measurements as a quality control is not sufficient because it cannot capture deviations of the geometry from the manufacturer’s specifications. Indeed, the geometrical uncertainty significantly impacts the predicted electric field and VTA and needs to be considered.
A second aspect is model calibration, as the electrochemical properties of the stimulation electrodes and the tissue around the electrode can change during the stimulation (Gimsa et al., 2005, Lempka et al., 2009). In rats, for example, an encapsulation layer forms shortly after implantation and significantly affects the impedance (Badstübner et al., 2017). We present a concept to extract the properties of the encapsulation layer during stimulation by combining impedance measurements and the numerical model. Eventually, the goal is to realize a computational model that is informed continuously during the stimulation.
In sum, we present how to enhance the fidelity of the model by validating the stimulation electrode geometry in a controlled in vitro setting. Then, we outline how this concept can be translated to the in vivo condition. The approach is general and can be straightforwardly adapted to human DBS. As we base all our analysis exclusively on open-source software, it can be readily employed by other researchers.
Learn more about the ISX-3 impedance analyzer
Get to know the ISX-3 impedance analyzer
