| 摘要: | 腦機介面為癱瘓患者與外在環境溝通的新興技術，近幾年所發展之各式腦機介面中，穩態視覺誘發電位腦機介面因具有使用者只需簡易訓練、較佳的資訊轉換率與準確率以及低成本等特性，成為發展的重點，然而為了達到穩態視覺誘發電位較高之訊雜比，基於其振幅與頻率的特性採用視覺誘發電位閃光頻率常低於20赫茲。但是若閃光閃爍頻率低於臨界閃光頻率，反而易造成使用者視覺上的不適；此外腦電波訊號為生理狀態(例如：情緒、專注力…等)神經活動的電生理反應，所以每個人所測得腦電波是有差異的，因此需要發展一個有效的分類方法改善相關問題。本篇論文將針對上述兩個問題提出改善方法：(1)設計高工作週期的視覺誘發電位閃光頻率去減少使用者視覺上的不適，在實驗設計中，使用13.16赫茲閃光頻率的發光二極體，其每一週期（T）為76毫秒（包含了亮狀態TON與暗狀態TOFF，且T = TON + TOFF），工作週期定義為TON/T，並測試不同工作週期(從10.5% 到89.5%)之穩態視覺誘發電位與採用問卷方式去調查使用者對閃光的舒適度，由六位受測者實機控制滑鼠得到高資訊轉換率（25.08位元/分）的實驗結果可歸納89.5% 工作週期的閃光頻率對使用者具有較高舒適性亦可適用於相位編碼的穩態視覺誘發電位腦機介面。(2)設計以支持向量機為基礎的分類方法進行相位編碼穩態視覺誘發電位腦機介面之目標偵測去改善使用者之間腦電波訊號差異所造成誤判問題，首先在分類器的訓練階段，每一使用者以離線紀錄所獲得穩態視覺誘發電位的振幅與相位去訓練支持向量機分類器，隨後應用於實機應用測試上，其方法首先以Kolmogorov-Smirnov(K-S)測試穩態視覺誘發電位的相位分布是否為有效資料，若判定為無效則再加入下一個狀態的相位，直到確定為有效資料，然後將有效資料所包含的相位與其對應的振幅當作分類器的輸入用以目標偵測，實機測試所得結果的準確性高達89.88 ± 4.76 %，反應時間為1.13 ± 0.02秒與較佳的資訊轉換率（50.91 ± 8.70位元/分），可顯著改善目標偵測的準確性。
Brain computer interface (BCI) is an emerging technology for paralyzed patients to communicate with external environments. Among current BCIs, the steady-state visual evoked potential (SSVEP)-based BCI has drawn great attention due to its characteristics of easy preparation, high information transfer rate (ITR), high accuracy, and low cost. Due to the amplitude-frequency characteristic of SSVEP, the flickering frequency of an SSVEP-based BCI is typically lower than 20 Hz to achieve high signal-to-noise ratio (SNR). However, a visual flicker with a flashing frequency below the critical flicker-fusion frequency often makes subjects feel flicker jerky, and causes visual discomfort. In addition, electroencephalogram (EEG) signals are electrophysiological responses reflecting the underlying neural activities which are dependent upon subject’s physiological states (e.g., emotion, attention, etc.) and usually variant among different individuals. The development of classification approaches to account for each individual’s difference in SSVEP is needed but was seldom reported. To overcome the above two problems, hence, the dissertation is divided into two studies. In the first study, we present a novel technique using high duty-cycle visual flicker to decrease user’s visual discomfort. The proposed design uses light emitting diodes (LEDs) flashing at 13.16 Hz, driven by flickering sequences consisting of repetitive stimulus cycles with a duration of T (T = 76 ms). Each stimulus cycle included an ON state with a duration TON and an OFF state with a duration TOFF (T = TON + TOFF), and the duty cycle, defined as TON/T, varied from 10.5% to 89.5%. This study also includes a questionnaire survey, and analyzes the SSVEPs induced by different duty-cycle flickers. An 89.5% duty-cycle flicker, reported as a comfortable flicker, was adopted in a phase-tagged SSVEP system. Six subjects were asked to sequentially input a sequence of cursor commands with 25.08 bits/min ITR. In the second study, a multiclass support vector machine (SVM)-based classification approach is proposed for gaze-target detections in a phase-tagged SSVEP-based BCI. In the training steps, the amplitude and phase features of SSVEP from off-line recordings were used to train a multiclass SVM for each subject. In the on-line application study, effective epochs which contained sufficient SSVEP information of gaze targets were first determined using Kolmogorov-Smirnov (K-S) test, and the amplitude and phase features of effective epochs were subsequently inputted to the multiclass SVM to recognize user’s gaze targets. The on-line performance using the proposed approach has achieved high accuracy (89.88 ± 4.76 %), fast responding time (effective epoch length = 1.13 ± 0.02 s), and the ITR was 50.91 ± 8.70 bits/min. The multiclass SVM-based classification approach has been successfully implemented to improve the classification accuracy in a phase-tagged SSVEP-based BCI. The present study has shown the multiclass SVM can be effectively adapted to each subject’s SSVEPs to discriminate SSVEP phase information from gazing at different gazed targets. |
