abstract: The efficacy of deep brain stimulation (DBS) in Parkinson's disease (PD) has been convincingly demonstrated in studies that compare motor performance with and without stimulation, but characterization of performance at intermediate stimulation amplitudes has been limited. This study investigated the effects of changing DBS amplitude in order to assess dose-response characteristics, inter-subject variability, consistency of effect across outcome measures, and day-to-day variability. Eight subjects with PD and bilateral DBS systems were evaluated at their clinically determined stimulation (CDS) and at three reduced amplitude conditions: approximately 70%, 30%, and 0% of the CDS (MOD, LOW, and OFF, respectively). Overall symptom severity and performance on a battery of motor tasks - gait, postural control, single-joint flexion-extension, postural tremor, and tapping - were assessed at each condition using the motor section of the Unified Parkinson's Disease Rating Scale (UPDRS-III) and quantitative measures. Data were analyzed to determine whether subjects demonstrated a threshold response (one decrement in stimulation resulted in ≥ 70% of the maximum change) or a graded response to reduced stimulation. Day-to-day variability was assessed using the CDS data from the three testing sessions. Although the cohort as a whole demonstrated a graded response on several measures, there was high variability across subjects, with subsets exhibiting graded, threshold, or minimal responses. Some subjects experienced greater variability in their CDS performance across the three days than the change induced by reducing stimulation. For several tasks, a subset of subjects exhibited improved performance at one or more of the reduced conditions. Reducing stimulation did not affect all subjects equally, nor did it uniformly affect each subject's performance across tasks. These results indicate that altered recruitment of neural structures can differentially affect motor capabilities and demonstrate the need for clinical consideration of the effects on multiple symptoms across several days when selecting DBS parameters. / Dissertation/Thesis / Ph.D. Bioengineering 2013

Introduction: Exploring the brain for optimal locations for deep brain stimulation (DBS) therapy is a challenging task, which can be facilitated by analysis of DBS efficacy in a large number of patients with Parkinson’s disease (PD). The Unified Parkinson's Disease Rating Scale (UPDRS) scores indicate the DBS efficacy of the corresponding stimulation location in a particular patient. The spatial distribution of these clinical scores can be used to construct a functional model which closely models the expected efficacy of stimulation in the region. Designs and Methods: In this study, different interpolation techniques were investigated that can appropriately model the DBS efficacy for Parkinson’s disease patients. These techniques are linear triangulation based interpolation, ‘roving window’ interpolation and ‘Monopolar inverse weighted distance’ (MIDW) interpolation. The MIDW interpolation technique is developed on the basis of electric field geometry of the monopolar DBS stimulation electrodes, based on the DBS model of monopolar cathodic stimulation of brain tissues. Each of these models was evaluated for their predictability, interpolation accuracy, as well as other benefits and limitations. The bootstrapping based optimization method was proposed to minimize the observational and patient variability in the collected database. A simulation study was performed to validate that the statistically optimized interpolated models were capable to produce reliable efficacy contour plots and reduced false effect due to outliers. Some additional visualization and analysis tools including a graphic user interface (GUI) were also developed for better understanding of the scenario. Results: The interpolation performance of the MIDW interpolation, the linear triangulation method and Roving window method was evaluated as interpolation error as 0.0903, 0.1219 and0.3006 respectively. Degree of prediction for the above methods was found to be 0.0822, 0.2986 and 0.0367 respectively. The simulation study demonstrate that the mean improvement in outlier handling and increased reliability after bootstrapping based optimization (performed on Linear triangulation interpolation method) is 6.192% and 12.8775% respectively. The different interpolation techniques used to model monopolar and bipolar stimulation data is found to be useful to study the corresponding efficacy distribution. A user friendly GUI (PDRP_GUI) and other utility tools are developed. Conclusion: Our investigation demonstrated that the MIDW and linear triangulation methods provided better degree of prediction, whereas the MIDW interpolation with appropriate configuration provided better interpolation accuracy. The simulation study suggests that the bootstrapping-based optimization can be used as an efficient tool to reduce outlier effects and increase interpolated reliability of the functional model of DBS efficacy. Additionally, the differential interpolation techniques used for monopolar and bipolar stimulation modeling facilitate study of overall DBS efficacy using the entire dataset.

This study focuses on the development of an objective, automated method to extract clinically useful information from sustained vowel phonations in the context of Parkinson’s disease (PD). The aim is twofold: (a) differentiate PD subjects from healthy controls, and (b) replicate the Unified Parkinson’s Disease Rating Scale (UPDRS) metric which provides a clinical impression of PD symptom severity. This metric spans the range 0 to 176, where 0 denotes a healthy person and 176 total disability. Currently, UPDRS assessment requires the physical presence of the subject in the clinic, is subjective relying on the clinical rater’s expertise, and logistically costly for national health systems. Hence, the practical frequency of symptom tracking is typically confined to once every several months, hindering recruitment for large-scale clinical trials and under-representing the true time scale of PD fluctuations. We develop a comprehensive framework to analyze speech signals by: (1) extracting novel, distinctive signal features, (2) using robust feature selection techniques to obtain a parsimonious subset of those features, and (3a) differentiating PD subjects from healthy controls, or (3b) determining UPDRS using powerful statistical machine learning tools. Towards this aim, we also investigate 10 existing fundamental frequency (F_0) estimation algorithms to determine the most useful algorithm for this application, and propose a novel ensemble F_0 estimation algorithm which leads to a 10% improvement in accuracy over the best individual approach. Moreover, we propose novel feature selection schemes which are shown to be very competitive against widely-used schemes which are more complex. We demonstrate that we can successfully differentiate PD subjects from healthy controls with 98.5% overall accuracy, and also provide rapid, objective, and remote replication of UPDRS assessment with clinically useful accuracy (approximately 2 UPDRS points from the clinicians’ estimates), using only simple, self-administered, and non-invasive speech tests. The findings of this study strongly support the use of speech signal analysis as an objective basis for practical clinical decision support tools in the context of PD assessment.