The goals of this study were: 1) to develop quantitative techniques for predicting the DBS VTA when balancing specific amounts of current through adjacent electrode contacts, and 2) to theoretically evaluate the utility of using current steering to control the VTA. The results show that balancing current delivery across two contacts increases the size of the VTA, compared to monopolar stimulation. Further, current steering can be used to sculpt the VTA to achieve the desired overlap with target tissue structures. This additional functionality may expand opportunities to achieve therapeutically optimal stimulation in a given patient. The important engineering designation that enables “steering” is the concept of independent sources (voltage or current). It would also be possible to manipulate the shape of the VTA with voltage steering, given that independent voltage sources were available on the IPG. However, current-controlled stimulation provides assurance that the current delivered to the tissue is consistent and unaffected by changes in the impedance of the electrode-tissue interface.
Clinical DBS parameter selection is art that balances stimulation induced therapeutic benefit while trying to avoid any stimulation induced side effects [9
]. This process can be difficult and time consuming, and the outcomes are strongly dependent on the experience of the programming clinician [33
]. The results of this study show that it is theoretically possible to sculpt the VTA with current steering to maximize stimulation coverage of a given anatomical target (i.e. the STN). However, clinical identification of the theoretically optimal stimulation parameter setting may not be intuitively obvious without computational assistance. And, many of the symptoms of neurological disorders treated with DBS do not provide instantaneous clinical feedback on their control (or lack thereof) as stimulation parameter settings are changed. To address these issues, multiple groups have been working to develop 3D visualization systems that employ engineering optimization to predict theoretically optimal electrode placements and stimulation parameter settings [35
]. However, the prospect of employing patient-specific computer models to augment clinical DBS practices requires two important pieces of a priori
information: 1) precise knowledge of the DBS electrode location in the anatomy, and 2) clear definition of the therapeutic target volume of tissue that should be stimulated. Unfortunately, given existing imaging techniques and scientific understanding, substantial limitations surround both of these prerequisites.
Identifying the location of DBS electrodes in post-operative MRIs is complicated by the metallic artifact generated by the contacts, and while post-operative CTs can provide a more accurate estimate of the 3D electrode location relative to the skull, the CT image must be co-registered with an MRI to provide neuroanatomical detail. Because of artifact and/or registration errors, current imaging technology can only provide an estimate to within ~1mm of the true location of the electrode in the brain anatomy [13
]. Uncertainties of this magnitude have not really impacted clinical care because post-operative imaging data is not extensively used to guide the stimulation parameter selection process under existing practices. However, as DBS electrodes and IPGs become more advanced, detailed knowledge of the lead location in the anatomy will be especially important in defining the electrode contacts and stimulation paradigms that maximize device utility in specific patients.
To provide a clinically relevant example of how current steering might be used in DBS we examined stimulation of the STN. Extensive effort has been dedicated to identifying the anatomical location of therapeutic electrode contacts in the STN region [14
]. The results of these studies suggest that stimulation of the STN per se may not be the only anatomical structure in the region that is responsible for therapeutic benefit. However, aside from detailed modeling studies of two patients [13
], estimates of the electrical spread of stimulation, and its overlap with various anatomical entities in the STN region, have not been correlated with therapeutic outcomes. In turn, scientific definition of the target volume of tissue that should be stimulated for maximal therapeutic benefit remains an issue of debate.
For simplicity we used the atlas-defined anatomical borders of the STN to represent our target volume of stimulation and we placed the DBS electrode in the center of the STN. However, the numerous anatomical studies of DBS electrode implant locations have shown substantial variability across patients. This variability in electrode placement can be attributed to multiple factors (stereotactic frame accuracy, surgical philosophy on target coordinates, use of microelectrode recordings, ect.), but the end result is that not every electrode will be perfectly placed in the intended target region. In turn, techniques like current steering, that expand the stimulation coverage and/or flexibility of any given implanted electrode, could be an important asset in a clinical setting. However, when defining therapeutic stimulation parameter settings in a patient, it is unrealistic to clinically evaluate each of the thousands of different stimulation parameter settings available for a given electrode and IPG model. Therefore, new clinical stimulation parameter selection techniques will be necessary to maximize the utility of advanced DBS systems. Assuming that the above mentioned limitations in imaging and stimulation target definition can be resolved, coupling current steering DBS technology with computer-assisted patient-specific stimulation parameter selection tools may be an attractive option for the future.
Predicting the spread of stimulation during DBS requires the definition of a neural response outcome measure. The VTA predictions used in this study represent simplified estimates of the response of myelinated axons to DBS. We originally developed the general threshold prediction scheme used in this study to address monopolar stimulation [12
]. We have attempted to validate this technique by comparing our model predictions of axonal activation to clinical measurements of internal capsule activation (i.e. DBS induced EMG thresholds from corticospinal tract activation) [13
]. When adapting this methodology to study current steering it became readily apparent that both active electrode contacts needed to be considered in the prediction scheme. Several formulas were evaluated and the solution presented in the Methods provided the best overall fit to our data. However, alternative techniques may provide more accurate solutions in the future.
Our previous theoretical and experimental analyses suggest that interesting correlations can be defined between axonal activation and therapeutic DBS [13
]. However, several important underlying assumptions in our study should be noted. First, the tissue medium was undifferentiated with regard to the distribution of myelinated axons. Second, activation of more tissue was considered better than less. Third, there was no relative benefit to stimulation of different sub-regions of tissue in the target volume. In addition, it is distinctly possible that the underlying therapeutic effects of DBS are actually related to neural responses unconnected to axonal activation. Nonetheless, the basic mechanisms of DBS are related to modulating neural activity with electric fields, and current steering represents an additional tool to enhance control of the applied electric field.
Three general design techniques exist to control the electric field generated by DBS: 1) construction of the stimulus waveform/train [51
], 2) shape of the electrode contact [12
], and 3) configuration of active electrode contacts [21
]. This study focused on one aspect of the configuration of active electrode contacts (i.e. current steering between adjacent cathodes). However, numerous questions remain to be addressed on all of the above mentioned DBS design techniques. For example, in relation to current steering a logical next step is evaluation of the spacing between contacts. shows the voltage distribution generated by simultaneous activation of 2 DBS contacts with various spacing. These different voltage distributions will result in different VTA shapes and thereby provide additional opportunities to customize the shape of the VTA.
Figure 6 Effects of electrode contact spacing on voltage distribution. The effects of contact spacing are shown by comparing current injection through one contact (leftmost panel) to injection through two contacts. The current injected is indicated on each electrode (more ...)
In summary, the results of this study show that current steering can expand opportunities to sculpt DBS to fit the anatomical target of the stimulation. However, numerous technical hurdles remain before the full spectrum of DBS technology can be optimally utilized in clinical settings. In turn, future improvements in DBS systems will rely on parallel development of engineering hardware (e.g. advanced stereotactic targeting, IPG technology), computer software (e.g. 3D visualization, optimization algorithms), imaging technology (e.g. advanced imaging sequences, intra-operative imaging), and clinical training on maximizing utility of these new technologies.