Quantifying Biomarkers for Brain Disease State Monitoring and Intervention

Abstract

The burden of neurological diseases in the country is growing and the need for improved therapies is critical. One approach to understanding and managing neurological diseases is through the electrical signals of the brain. Local field potentials (LFPs) are electrical signals recorded from the brain that represent the activity of populations of neurons surrounding the recording site. During injury and disease, the normal electrophysiological behavior of neurons is replaced by aberrant activity. Disease biomarkers are pathophysiological LFPs that are present during disease. Investigating neural biomarkers can shed light on disease progression as well as inform treatment options. In this dissertation, we improve current technology to allow for superior biomarker quantification in two applications: monitoring disease state and informing therapeutic intervention. In the first part of this dissertation, we developed new technology to enhance brain monitoring for ischemic stroke. After initial injury, there is a critical period of days and weeks during which the brain is at risk of further damage. Spreading depolarizations (SDs) are disease biomarkers that are implicated in worsened outcomes (i.e. larger volumes of dead tissues). However, the exact contribution of SDs to disease progression remains unclear: SDs might be detrimental or beneficial depending on the phase of the progression. Therefore, the ability to concurrently track SDs and brain disease state (i.e. tissue viability) is critical for advancing our knowledge of disease progression and ultimately drive therapy development. We have addressed technological challenges that, until now, prevented the simultaneous monitoring of SDs and stroke boundaries. We validated our novel recording technique using in vivo electrophysiology in rodents. Our findings show that high density micro-electrocorticography (µECoG) arrays are able to record reliably both sub-hertz activity (the frequency range of SDs) and underlying spontaneous brain activity (that reflects tissue health and viability). These advances will enable long-term tracking of SD contribution to stroke boundaries. In the second part of this dissertation, we developed novel paradigms to guide delivery of deep brain stimulation (DBS) in Parkinson’s disease (PD). DBS is a well-established treatment option for people with PD that manages motor symptoms. Current research efforts are focused on optimizing DBS by only delivering therapy as needed, based on a patient’s current clinical state. This approach is called closed-loop DBS and holds promise for providing optimal symptom relief while avoiding stimulation-induced side effects and reducing battery consumption. Even though biomarkers have already been extracted from the LFP of the basal ganglia during DBS for PD, clinical translation of these biomarkers as guides for closed-loop DBS faces other challenges. Most notably, the window between two stimulation pulses during which biomarkers local to the stimulation site are recorded is rather short (< 10 ms for therapeutic DBS frequencies of > 100 Hz). Stimulation artifacts often persist after the pulse is over (for up to multiple milliseconds) and obscure underlying biomarkers. Through intra-operative studies, we demonstrated that temporally non-regular patterns of DBS can provide longer windows for biomarker recording (~ 50 ms) while maintaining clinical efficacy. We also quantified the effect of mismatched source impedances on the stimulation artifact and proposed a novel pre-amplification circuit to minimize the stimulation artifact during DBS recordings. These advances will facilitate biomarker quantification during DBS therapy, ultimately aiding clinical translation of closed-loop paradigms.