Cortical Evoked Potential as a Biomarker for Deep Brain Stimulation

dc.contributor.advisor

Grill, Warren M

dc.contributor.author

Cassar, Isaac Russell

dc.date.accessioned

2021-09-14T15:08:26Z

dc.date.available

2023-09-13T08:17:16Z

dc.date.issued

2021

dc.department

Biomedical Engineering

dc.description.abstract

Deep brain stimulation (DBS) is a highly successful neuromodulation therapy for treating the motor symptoms of Parkinson’s disease (PD). However, DBS has been used for over 30 years with little change in clinically used stimulation parameters and technology, and consequently, there have been few improvements in therapeutic efficacy during that period. Fortunately, recent advances in DBS devices and techniques, including automated stimulation parameter selection, directional leads, closed-loop stimulation, and model-optimized temporal patterns of stimulation, have the potential to improve symptom reduction, decrease side effects, and extend device battery life. However, making use of many of these techniques requires a recordable electrophysiological signal, or biomarker, that correlates strongly with clinical outcomes. The goals of this dissertation were to develop tools that assist in the recording and application of biomarkers, to characterize a new potential biomarker, the cortical evoked potential (cEP), and correlate it with symptom reduction, and to understand mechanistically how the cEP relates to symptom reduction during DBS. First, we quantified the effects of a novel electrodeposited platinum-iridium coating (EPIC) on single unit recording performance. We implanted electrodes in rats and used electrophysiological and histological measurements to compare quantitatively the single unit recording performance of coated vs. uncoated electrodes over a 12-week period. The coated electrodes had lower impedance, reduced noise, increased signal-to-noise ratio, and an increased number of discernible units per electrode as compared to the uncoated electrodes. These results demonstrated that EPIC electrodes provided recording performance and longevity superior to uncoated electrodes, thus improving our ability to quantify potential biomarkers from single unit recordings. Second, we developed a modified genetic algorithm (GA) designed to optimize temporal patterns of stimulation. We developed five modifications to the standard GA repopulation step that adapted the GA to design patterns for neuromodulation applications. We evaluated each modification individually and all modifications collectively by comparing performance to a standard GA across three test functions and two biophysically-based models of neural stimulation. The modifications improved performance across the test functions and performed best when all were used collectively. Thus, we developed a powerful tool for optimizing temporal patterns of stimulation using model-based proxies of DBS biomarkers. Third, we characterized a new candidate biomarker for DBS, the cEP, and quantified its correlation with symptom reduction during DBS. We used the unilateral 6-hydroxydopamine (6-OHDA) lesioned rat model or parkinsonism, with stimulating electrodes implanted in the subthalamic nucleus (STN) and the electrocorticography (ECoG) recorded above motor cortex (M1). We recorded the cEP during a range of stimulation conditions and while performing behavioral assessments of hypokinetic symptoms. The cEP was strongly affected by stimulation condition, and the cEP magnitude declined and the cEP latency increased with higher stimulation frequencies. These effects occurred over multiple minutes and with multiple time-scales. Additionally, the cEP magnitude and latency were each strongly correlated with symptom reduction during DBS, with correlations that were stronger and more consistent than those of conventional spectral-based biomarkers. This study demonstrated the potential utility of the cEP as a biomarker for symptom reduction from DBS. Fourth, to understand better how the cEP may relate mechanistically to symptom reduction from DBS, we developed a computational model of antidromic cortical activation during STN DBS and assessed the ability of DBS to desynchronize pathological cortical beta band oscillations. We tuned and validated the model using experimental data from the 6-OHDA lesioned rat model, and we implemented a stochastic model of antidromic spike failure, which is the presumed cause of the observed changes in the cEP magnitude, to determine how changes in the cEP relate to cortical desynchronization. STN DBS desynchronized pathological oscillations at high stimulation frequencies via a mechanism analogous to the informational lesion theory. Specifically, the DBS-evoked spikes masked the intrinsic pathological spiking via a combination of refractoriness, spike collision, and synaptic depletion. Further, the model revealed that antidromic spike failure played a critical role in shaping the therapeutic frequency profile of this masking effect, enforcing a parabolic shape with maximum desynchronization at ~130 Hz. The results in this dissertation advance our understanding of the therapeutic mechanism of STN DBS, provide important tools for application of electrophysiological biomarkers in DBS, and characterize the utility of the cEP as a potential biomarker to improve therapeutic outcomes.

dc.identifier.uri

https://hdl.handle.net/10161/23744

dc.subject

Biomedical engineering

dc.subject

Neurosciences

dc.subject

Biomarker

dc.subject

Deep brain stimulation

dc.subject

Evoked compound action potential

dc.subject

Evoked potential

dc.title

Cortical Evoked Potential as a Biomarker for Deep Brain Stimulation

dc.type

Dissertation

duke.embargo.months

23.934246575342463

Files

Original bundle

Now showing 1 - 1 of 1
Loading...
Thumbnail Image
Name:
Cassar_duke_0066D_16337.pdf
Size:
7.2 MB
Format:
Adobe Portable Document Format

Collections