Identification and characterization of short-term motor patterns in rest tremor of individuals with Parkinson’s disease

• Ano de defesa: 2023
• Autor(a) principal: Rabelo, Amanda Gomes
• Tipo de documento: Tese
• Tipo de acesso: Acesso aberto
• Idioma: eng
• Instituição de defesa: Universidade Federal de Uberlândia; Brasil; Programa de Pós-graduação em Engenharia Elétrica
• Palavras-chave em Português: Empirical Mode Decomposition; k-means; gyroscope; Parkinson’s disease; rest tremor; short-term motor patterns (STMPs); t-SNE; giroscópio; doença de Parkinson; tremor de repouso; padrões motores de curto prazo (STMP); CNPQ::ENGENHARIAS::ENGENHARIA BIOMEDICA::BIOENGENHARIA::PROCESSAMENTO DE SINAIS BIOLOGICOS; Engenharia elétrica; Parkinson, Doença de; Distúrbios psicomotores; Giroscópios
• Link de acesso: https://repositorio.ufu.br/handle/123456789/39125; http://doi.org/10.14393/ufu.te.2023.515

Tremor serves as a significant biomarker for various diseases, including Parkinson's Disease, and plays a crucial role in monitoring disease progression, assessing treatment efficacy, and aiding in the diagnosis of movement disorders. Despite considerable progress in tremor research over the past thirty-eight years, challenges still remain in understanding the nature of tremors and within-individual fluctuations. A deeper understanding of tremors can lead to personalized treatment approaches and optimize pharmacogenomics studies for the pathology. The objective of this research is to identify and characterize the Short-Term Motor Patterns (STMPs) present in the rest tremor signal using inertial sensors. STMPs manifest in the signal in less than 1 second and exhibit self-similar structures across multiple time scales. They have a hidden dynamic with underlying structures contributing to the abnormal movement observed in tremors. The study involved healthy individuals (N = 12, mean age 60.1 ± 5.9 years) and individuals with Parkinson's Disease (N = 14, mean age 65 ± 11.54 years). Signals were collected using a triaxial gyroscope placed on the dorsal side of the hand during a resting condition. The data were pre-processed, and seven features were extracted from each 1-second window with 50% overlap. The STMPs were identified using the k-means clustering technique applied to the data in the two-dimensional space generated by t-Distributed Stochastic Neighbor Embedding (t-SNE). The frequency, transition probability, and duration of the STMPs were assessed for each group. All STMP features were averaged across the groups. Three distinct STMPs (STMP1, STMP2, and STMP3) were identified in the tremor signals (p < 0.05). STMP1 was predominant in the healthy control (HC) subjects, STMP2 was present in both the healthy and Parkinson's disease group, and STMP3 was observed in the Parkinson's disease group. Only the coefficient of variation and complexity not showed significant differences between the groups. Regarding signal dynamics, signals from individuals with Parkinson's disease tended to exhibit lower STMP transition probabilities and longer durations of STMP than the healthy control subjects. These findings can assist professionals in characterizing and evaluating the severity of tremors and assessing treatment efficacy.