手部的運動是人類日常生活中最複雜的運動行為，而控制該部位運作的大腦運動區，牽扯複雜的神經系統，所以成為科學家研究的重點。為了探討大腦運動皮質區的激活反應，研究學者藉由腦造影技術來分析作用區域及生理反應，而近紅外光譜(Near-infrared spectroscopy，NIRS)為近期學界熱烈討論的腦造影技術，因為它具有靈敏的時間解析度及高空間解析度，且能在開放空間中作動態實驗。 本篇研究提出以近紅外光譜作為握力實驗的大腦掃描方式，套用高階統計學的演算法－獨立成份分析法(Independent Component Analysis，ICA)，將訊號分解成許多獨立成份(Independent component，IC)，我們先以傳統的平均方法，建立訊號樣版，接著以訊號樣版作為選擇獨立成份的依據，選擇適當的獨立成份，進行訊號重建以萃取出與運動相關的紅外光血流訊號。本研究量測執行左手或右手握拳運動的紅外光訊號，並比較握力量級的振幅變化。本篇研究的實驗結果證實，獨立成份分析法可有效的分離出手部運動訊號，提升訊號訊雜比。 Hand movements is the most complicated motor behaviors in our daily life. The motor cortex, controlling the human motor functions, is a complex system whose working mechanism is an important research issue for scientists. To investigate the functions of motor cortex during movements, researchers analyze hemodynamic and neural activities using brain imaging techniques. The near-infrared spectroscopy (NIRS) has drawn great attention due to its high temporal and spectral resolutions, and can be used to perform experiments in open-space environments. In this study, NIRS has been adopted to study motor cortex function in human brain during hand-grip task. The independent component analysis (ICA) was chosen for extracting oxygen and deoxygen hemodynamic responses. For each subject, a temporal template was first constructed for each subject from traditional averaging process. Each independent component (IC) was correlated with the temporal template, and those ICs with high correlation values were chosen as task-related components for constructing noise-suppressed hemodynamic response. This thesis studied the NIRS signals during performing right and left grip tasks with different force levels. Our study results have shown the ICA is an effective tool to improve the signal-to-noise ratio (SNR) of motor-related NIRS signals.