This thesis establishes a comprehensive framework for oscillatory electrophysiological biomarkers through three progressive studies addressing peripheral diagnostics, central characterisation, and translational validation. Study 1 analysed EMG/accelerometer recordings from 396 subjects using machine learning with external testing across clinical centres to differentiate PD from essential tremor. Peripheral EMG power features achieved superior diagnostic discrimination (ROC-AUC: 1.00 internal dataset, 0.79 external test dataset), potentially reducing misdiagnosis. However, performance reduction in external test highlights the need for standardised protocols before clinical implementation. Building on these peripheral findings, Study 2 introduced a novel wavelet-based burst detection framework for subthalamic nucleus recordings from 7 advanced PD patients. This work established a dual band ω model that reframes PD pathophysiology: pathological low ω (13- 20 Hz) bursts with prolonged duration correlated with motor severity, while high ω (21-35 Hz) bursts negatively correlated with impairment, suggesting compensatory mechanisms. Limited sample size necessitates validation in larger cohorts. Study 3 then translated these findings using structural equation modelling to establish causal relationships between oscillatory biomarkers, dopaminergic function, and motor impairment in both animal models and human patients. High ω parameters independently predicted motor dysfunction (S = 0.67, p < 0.01), while low ω activity mediated by striatal dopaminergic loss predicted both motor impairment and neurodegeneration (S = 0.65, p < 0.01). This work establishes neural oscillations as clinically viable biomarkers bridging mechanistic understanding with practical applications. The findings reveal high ω’s potential protective role versus low ω’s pathological correlation with dopaminergic degeneration, providing foundation for accessible diagnostic tools, objective disease monitoring, and personalised therapeutic strategies including frequency-specific deep brain stimulation. These biomarkers offer immediate utility in resourcelimited settings and precision medicine approaches, with future longitudinal studies needed to validate clinical implementation protocols.
Tanmoy Sil (Thu,) studied this question.