運用穩態視覺誘發電位檢測於適應性分析之偏頭痛狀態分類

Abstract

偏頭痛是個被我們低估的神經系統疾病，它能無任何徵兆地在任何時間發作。而為什麼偏頭痛是周期性的以及為什麼會產生偏頭痛都還是未知數。如果能分辨出患者處於哪個時期，對於我們對抗即將發生的頭痛有很大的幫助。但是這需要大量腦波資料來進行研究，以目前腦波實驗的方法來看，想收集到足夠的患者腦波資料以及足夠的偏頭痛各時期的資料是很困難的。這是因為傳統腦波實驗在實驗前的準備及實驗後的處理都需要花費不少的時間，而且要收集各時期的資料勢必需要受測者配合長期的實驗以收齊腦波資料，這對於受測者來說是相當麻煩的。我們引進無線腦波儀來幫助我們改善這個問題以增加受測者配合實驗的意願。 本篇論文目的在建立偏頭痛狀態分類器，將收集到的臨床病患腦波資料做訓練並驗證分類器的準確率。有了臨床腦波資料的基礎後，我們將無線腦波儀所收集到的資料也進行訓練及分類，驗證是否無線腦波儀所收集的資料也能使用，並提供一個初步的無線腦波儀應用結果，增加無線腦波儀廣泛運用於偏頭痛研究的可行性。

Migraine is an underestimated disorder. Migraine attacks can happen without any aura and at anytime. Why migraine attacks happen periodically, and the mechanism behind migraine, are uncertain. If we can identify which migraine stage patients are in, it can help us control or relieve the coming attacks. But the research on migraine stages require huge amount of EEG data from patients. And it's difficult to collect enough data via current experiment method. That's because the preparation of EEG experiment will take a long time, subjects usually have to spend 2 to 3 hours for single experiment. If we want to collect enough EEG data of each migraine stage, subjects have to participate the long-term EEG experiment. That would bother subject's daily life and strongly lower subject's willing to record EEG data. We Utilize the mobile EEG device to reduce the preparation of EEG experiment and improve subject's participation. In this study, we build migraine state classifier. Then we exam the classifier by clinical EEG data, and use the result as the foot stone to applied classifier on mobile EEG data. Exam the performance of mobile EEG data and provide a preliminary result for application of mobile EEG device and increase the possibility of generally using mobile EEG device on migraine researches.