karimi, fatemeh2024-02-022024-02-022024-02-022024-01-12http://hdl.handle.net/10012/20330Freezing of gait (FOG) is a complex and debilitating gait disturbance in Parkinson's disease (PD) that significantly impacts mobility and quality of life in PD patients. FOG affects approximately 86.5% of individuals with advanced PD and up to 37.8% in the early stages of the disease. Unfortunately, dopaminergic medication provides limited benefits for FOG, and the lack of a clear understanding of its underlying neurophysiological mechanisms has resulted in limited efficacy of alternative treatment options based on emerging technologies such as brain-computer interface (BCI), deep brain stimulation (DBS), transcranial magnetic stimulation (TMS), and transcranial direct current stimulation (tDCS). Electroencephalogram (EEG), as a portable and noninvasive cortical recording technique, records oscillations representing the collective neural activity underlying the neural communication in the brain. EEG contains various components exhibiting a variety of possible characteristics across spatial and temporal scales. Integrative EEG analysis in PD patients with FOG holds the potential to establish new and efficient FOG rehabilitation pathways, while also introducing biomarkers for diagnosing FOG, monitoring its progression, predicting treatment response, and evaluating the effectiveness of therapeutic interventions. In this study, we adopted a comprehensive approach. First, we analyzed power and phase-related EEG features in PD patients with different levels of FOG as well as PD patients without FOG and age-matched healthy controls (HC) during a simple lower limb task. We then extended the findings to more realistic and complex walking tasks. For this purpose, a total of 41 individuals, including patients with freezing (N = 14, 11 males), patients without freezing (N = 14, 13 males), and HC (N = 13, 10 males), were recruited. All participants performed two sets of tasks: ankle dorsiflexion (AD) while seated and walking tasks. PD patients with FOG were further classified into mild and severe cases to investigate EEG features associated with freezing severity. Initially, we examined the morphological features of a specific EEG signature called Movement Related Cortical Potential (MRCP) during AD task. Additionally, we investigated power activities of distinct frequency bands (theta, alpha, and beta) during the same task across different participant groups and channels. In the second stage, we explored phase-related features of MRCP and other frequency bands in superficial and deeper networks by applying a spatial filter called surface Laplacian (SL). Lastly, considering the observed alterations in power and phase of different frequency bands during the simple lower limb task (AD), we investigated phase amplitude coupling (PAC) between various frequency bands during normal walking condition (NW) and FOG episodes (FE) in PD patients with FOG and compared the results with PD patients without FOG and HC during normal walking. The initial results of this thesis, focused on the cortical dynamics during AD, revealed significant differences between patients with severe FOG and both HC and patients without FOG. Moreover, patients with mild and severe FOG exhibited distinct cortical activity patterns. In patients with FOG, the initial component of MRCP was significantly reduced compared to HC (P = 0.002), and its magnitude was influenced by the severity of FOG. Notably, patients with FOG demonstrated a remarkable absence of desynchronization in the beta frequency band, particularly in the low-beta range over primary motor cortex (M1), prior to movement initiation, which was also correlated with the severity of FOG condition. Low-beta and high-beta activities represented unique characteristics for each group. While HC exhibited beta event-related desynchronization over M1 before movement, patients with FOG showed partial replacement of this pattern by theta band synchronization. Patients with severe FOG also showed some degree of theta band synchronization over the contralateral SMA. Regarding the phase-related features of FOG during AD task, frontoparietal theta phase synchrony emerged as a distinctive characteristic in the superficial layers of PD patients with FOG. In deeper networks, interhemispheric frontoparietal alpha phase synchrony was significantly dominant in PD patients with FOG, in contrast to beta phase synchrony observed in PD patients without FOG. Furthermore, alpha phase synchrony was more widely distributed in PD patients with severe FOG, particularly exhibiting higher levels of frontoparietal alpha phase synchrony. In addition to FOG-related abnormalities found in the phase-locking value (PLV) analysis during AD task, PAC analysis was performed on frequency bands with PLV abnormalities during this task. PAC analysis revealed abnormal coupling between theta and low-beta frequency bands in PD patients with severe FOG at the superficial layers over frontal areas. In deeper networks, both theta and alpha frequency bands showed significant PAC over parietal areas in PD patients with severe FOG. Alpha and low-beta bands also exhibited PAC over frontal areas in PD groups with FOG. PAC analysis comparing normal walking (NW) to freezing episodes (FE) also identified significant differences between these conditions in PD patients both with and without FOG, as well as in HC. Our results demonstrated that PAC between theta and low-beta frequencies in PD patients with FOG during FE exhibited higher statistical significance compared to PAC in PD patients with FOG during NW, PD patients without FOG during NW, and HC during NW, specifically over the (pre-) SMA and parietal areas (p<0.01). Additionally, the findings showed elevated PAC between alpha and low-beta frequency bands in the parietal area during FE (p<0.01). There was also a higher PAC between theta and alpha frequency bands in PD patients with FOG compared to the other two groups (p<0.01), regardless of the experimental condition. This thesis makes several contributions to current literature bridging both neurophysiological and engineering domains. Our findings highlight the critical role and potential of phase-related EEG signal features in postulating a unified mechanism for FOG. These results also suggest PLV and PAC during AD task as prospective EEG-based biomarkers for diagnosing and monitoring FOG. Additionally, these results provide novel perspectives for developing non-pharmacological strategies for FOG intervention and rehabilitation.enDoctoral Thesis