| 摘要: | 近年來本國隨著對語言治療的重視日益增加，因此診斷出有語言障礙的患者也逐年成長，導致了語言治療師的人力缺乏，而且在傳統語言治療中，溝通障礙患者需要語言治療師一對一去進行治療，故語言治療師同一時間內很難同時對多位病人進行治療，因為人力不足導致於其治療成效不佳。基於以上原因，本研究開發一套針對訓練音高和音量的嚴肅遊戲，藉由遊戲的方式進行復健，希望提高受測者的動機和改善其構音障礙和音量不足的問題。此系統利用筆記型電腦上的麥克風收錄人聲訊號，由電腦計算出此段語音的基頻和音量，再結合遊戲動畫，利用遊戲動畫給予受測者練習發音的視覺回饋。為了評估此系統對語音復健的有效性，本研究與衛福部部立桃園醫院復健科的語言治療團隊合作，徵求了五位患有音聲障礙的實驗組參與實驗，同時也安排六位正常組進行實驗以比對差異，內容根據患者實際狀況決定，內容包含音高訓練和音量訓練，總共三次練習，一個禮拜一天，總計進行六次的發音練習實驗，每次耗時約30分鐘，除了做本研究開發的語音復健系統(A系統)的實驗之外，也同時做本實驗室所先前所開發的音聲障礙復健系統(B系統)的實驗。在實驗的最後，我們會提供一個問卷給受測者做填寫，上面有11個問題，以評估實驗過程中的體驗。完成實驗後，比對受測者第一次與最後一次的實驗數據，以評估其音聲障礙是否有被改善，同時也比較A系統與B系統的結果差異。為了比較實驗組與正常組在練習前後音高和音量是否有所不同，我們使用T檢定 ( T test ) 算出練習前和練習後的p值。在A系統方面，實驗組最後一次練習後的平均音域比第一次練習後的平均音域增加了103Hz，p值為0.025 ，有顯著差異；而平均音量則增加了12bB，p值為0.023 ，有顯著差異；而正常組部分，練習前後對其音域和音量皆無顯著差異。在B系統方面，實驗組最後一次練習後的平均音域比第一次練習後的平均音域增加了68Hz，p值為0.005 ，有顯著差異；而平均音量則增加了16bB，p值為0.01 ，有顯著差異；而正常組部分，練習前後對其音域和音量皆無顯著差異。在音量訓練上B系統的訓練成果優於A系統的訓練成果，造成此種結果的可能性有三個，分別是遊戲型態、個案數量不足和人為誤差。在音高訓練上A系統的訓練成果優於B系統的訓練成果，推論可能是因為程式中遊戲動畫觸發器設定不同所造成。經由上述實驗結果可知，實驗組完成實驗後，其音量和音域皆有明顯的成長，且與其他學者的研究結果相近，證明了本系統可以用來幫助患者改善發音，並且在問卷回饋中，受測者都對此系統保持正面看法。;As the attention on the language therapy increased in Taiwan in recent years, the growth of patients discovered with language disorder leads to lack of language therapists. For traditional treatment of speech disorder, speech therapists need to assess and treat patients with communication disorder individually. Such personalized therapy process is inefficient for massive patients waiting for their next rehabilitation appointments. In addition, this treatment performance is ineffective since lack of language therapists. Therefore, we proposed a serious game for rehabilitation on voice intensity and pitch. We hoped patients would not only improve the dysarthria and decay of volume but also be motivated for speech rehabilitation by playing the game. The system recorded a short speech signal from human and calculated its volume (in decibel, dB) and fundamental frequency (in Hertz, Hz) immediately. These two parameters were used to control the animated character and provide visual feedbacks for patients. In order to assess the effectiveness of this game system, we collaborated with a speech therapist of the Taoyuan General Hospital. Eleven subjects including five patients with voice disorder and six normal subjects were involved in the clinical study to see if there were any differences in between. Experiments were carried out for three weeks (one day per week and six times per day with a total time of 30 minutes). Each experiment included 4 pitch and 2 volume games.
Besides the experiment of our system (A system), we did the experiment of the game system (B system) for voice disorder from previous study of our laboratory. At the end of the experiments, subjects were presented with a questionnaire which included 11 items aiming at evaluating various aspects of the training experience. After finishing all the experiments, the speech performance of patients was assessed by comparing the results of the first and last experiments. Next, we would like to see the difference between the results of A system and B system. To examine the difference between normal subjects and patients, we analyzed the results of the first and last experiments with the T test and calculated the p value. The results of A system were found that the average frequency range of patients was increased by 103Hz and the average of volume was increased by 12dB. The p value of frequency range is 0.025 and the p value of the volume is 0.023, hence there were significant differences. The results of B system were found that the average frequency range of patients was increased by 68Hz and the average of volume was increased by 16dB. The p value of frequency range is 0.005 and the p value of the volume is 0.01, hence there were significant differences. The results of A system and B system showed the average frequency range and volume of normal subjects for normal subjects were not significant changed in the last experiment. In voice intensity treatment, the effectiveness of B system was better than the effectiveness of A system. We surmised the probable causes including
personal error, the lack of sample size and the difference of the game type between two systems. In pitch treatment, the effectiveness of A system was better than the effectiveness of B system. We inferred that the probable reason was the difference of the animation trigger between two systems. The superior frequency range and volume of patients after practices were similar to the results of other research in the literature. It has proved the benefits of this game system for speech therapy. In addition, the feedback of questionnaire showed the subjects were all confident in the system that we presented. |
