駕駛在不同瞌睡程度下大腦訊號傳遞網絡的變化

Abstract

在車禍事故的原因當中，駕駛者瞌睡普遍被認為是主要因素。過去我們研究團隊層探討駕駛者瞌睡的腦電波現象，設計利用腦電波偵測瞌睡之演算法，並且開發成為駕使者的無線可攜式瞌睡偵測設備。縱使我們已經有良好的瞌睡偵測指標，然而過去的方法中並不能解決從清醒到瞌睡漸進程度上的偵測。本研究探討的是，受測者從清醒到瞌睡過程中腦內訊號傳遞網絡是如何改變，以及真實生活中動態刺激造成的訊號傳遞。一共有六位受測者參與模擬夜間高速公路動態及非動態的虛擬實境開車實驗，並且利用此環境給與受測者事件相關開車偏移任務。受測者的腦電波會經過獨立訊號分析、Granger因果關係分析、時域頻譜轉換等方法進行分析比較。結果顯示，從駕駛者清醒到瞌睡的過程中，訊號傳遞的目的地會從腦中前面的區域移動至後面的區域。此外在動態模擬下駕駛腦內掌管視覺和運動區域會比非動態的情況還要活躍。未來可利用這樣的結果，在適當的腦區位置偵測駕駛目前的瞌睡狀態，並且於進入瞌睡前就給與警示，使駕駛者的行車表現能保持良好水平。

Driver drowsiness was generally regarded as a main reason of causing car accidents. Our team had investigated EEG signals in drowsy state of driver, designed the algorithm that detecting drowsiness by EEG signals, and developed the wireless and portable application of detecting drowsiness for drivers. Although we already had the good indicators of detecting drowsiness by EEG signals, detecting the levels of drowsiness remained unsolved nowadays. The aim of this study is to explore the changes of brain signal transferring network from alertness to drowsiness and those signal flows generated by kinetic stimulation in real life. Six subjects participated in virtual-reality (VR)-based highway driving experiments on motion and motionless platform, and the event-related lane-departure task was used in the VR environment to simulate the long-term highway driving. The task-related EEG was analyzed using independent component analysis, Granger causality, and time-frequency. Results demonstrated the destination of signal flow was from anterior brain region shifting to posterior region respective alert to drowsy state. Furthermore, the EEG transferring dynamics were more active in occipital area and motor area on motion platform. In the future, the results can be used for detecting drowsiness in proper brain region, and warn the driver before drowsiness to make the performance of driver keep at a good level.