以連續走路影像量化帕金森氏症患者之步態起始與對稱性

Abstract

帕金森氏症是一種中樞神經系統退化性的疾病，更進一步會造成患者在外觀行為上的改變，而步態即為帕金森氏症之主要的特徵。時至今日，步態之起步與對稱性的偏差評估分析在神經醫學疾病仍為一重要議題。對於帕金森氏症患者而言，由於神經系統退化的影響不僅會造成行為調節的困難對於起步更是一項挑戰。因此，我們希望能發展一套客觀且量化之步態分析方法用以輔助醫師來評估大量帕金森氏症患者在步態起步與對稱性的嚴重程度。 在本研究裡，我們所發展的影像式系統首先運用重心追蹤演算法來定出腳踝與膝蓋之位置並獲得膝部角度訊號，最後提取在第一步與穩定狀態下的量化步態參數與其對應的對稱性指標。結果顯示了在第一步與穩定狀態下，帕金森氏症患者於用藥前不僅與年紀相近的正常人在走路速度與跨步長度具有統計上的差異，同時在帕金森氏症患者於用藥前後在這兩項參數上也具備明顯差異(t-test)。此外，正常人相較於帕金森氏症患者用藥前後在跨步長度與跨步速度之對稱性指標仍具有統計上的差異。我們更進一步去探討所得到之步態參數與統一帕金森氏症分級量表(Unified Parkinson Disease Rating Scale)之間的關連性。在量表的評估項目中: 步態與姿態，兩腳靈敏度和全身動作遲緩這三項評估項目上，與跨步長度得到顯著性的關連性係數為-0.5978, -0.5909和-0.6362與走路速度得到顯著性的關連性係數為-0.6033, -0.5725和-0.6566。這同時更能進一去解釋帕金森氏症患者在姿態反應與動作遲緩等症狀的變化。本研究發展的系統証明了和某些臨床症狀上具有關聯性，並為診斷帕金森氏症步態提供了一個客觀參考指標。

Parkinson’s disease (PD) is a neurodegenerative disease of central nervous system which causes motor fluctuations and gait is considered as the major hallmark of PD. Nowadays, the assessment of specific gait disturbances of gait initiation and symmetry is an important issue in clinical neurological diseases. Neuro-degeneration in PD patients increases the deficiencies in regulating the locomotion and challenges initiation of gait. Therefore, we want to develop an objective and quantitative gait analysis method for assisting clinicians with assessing massive PD patients in the severity of gait initiation and symmetry. The proposed image-based system utilizes a Centroid Tracking algorithm to localize the positions of knee and ankle, extracts knee angle signal and obtains quantitative gait parameters and associated symmetry indexes in first step duration and stable condition. In first step duration and stable condition, the results demonstrate that the statistical differences of swing length and velocity between age-matched normal subjects and PD patients before drug treatment and between PD patients before and after drug treatment are confirmed by t-test (p<0.05). Besides, the symmetry indexes of swing length and swing speed also have statistical differences between normal subjects and PD patients before and after drug treatment. The further relationship between gait parameters and the Unified Parkinson Disease Rating Scales (UPDRS) is investigated. The UPDRS sub-scores: Gait & Posture, Leg Agility and Body Bradykinesia have significant correlation coefficients of -0.5978, -0.5909 and -0.6362 with swing length and -0.6033, -0.5725 and -0.6566 with velocity. These results can further explain the variations of postural reaction and bradykinesia. The developed system is proved to be correlated with some clinical symptoms and provides a quantitative method in the assessment of parkinsonian gait.