|Abstract: ||過去人們曾經認為隨著年齡變化工作記憶(Working memory)的容量仍然是恆定的，但是越來越多的心理生理研究顯示，工作記憶的容量是可以經由適當且長期的訓練造成改變，例如工作記憶的容量提升會反映在任務表現與智力測驗的增加上。但仍有部分的學者提出相反的意見，認為工作記憶的容量提升並沒有反映在智力測驗分數的增減上，其中的原因可能是，認知訓練的成效造成神經生理結構上的改變，但最主要的效果並不是反映在智力測驗上。因此本研究希望由事件相關同步化/非同步化(ERS/ERD)方法，從神經生理的電訊號改變中，找到一個客觀量化的指標，以用來評估工作記憶訓練的成效。|
腦電圖資料使用莫萊小波轉換做後處理，分析頻帶為Low theta 4-6Hz、High theta 6-8Hz、Low alpha 8-10Hz、High alpha 10-12Hz，接著使用事件相關同步化/非同步化技術，找出兩組表現相異之處。除此之外，反應時間與回答正確率也在實驗結束後做蒐集，加入統計分析中。
對於行為數據的統計結果顯示，反應時間與回答正確率中，兩組並不存在組間差異，這結果與先前的文獻一致，行為數據的變化未必能反映出工作記憶的成效。經由迴歸分析，我們發現記憶負荷與行為數據存在這非線性的相關，但這結果與訓練無關。腦電圖的結果顯示了工作記憶訓練的成效(組間差異)主要表現在後側腦區的ERD/ERS (Low theta, 0~500 ms)。另外，考慮到不對稱的因素，我們發現訓練後在中央腦區(Low theta, 500~1500ms)和後側腦區 (Low theta 、Low alpha , 250~1000ms)單邊優勢有顯著的組間差異。
;It was once believed that working memory (WM) capacity is constant through life span. Recently, psycho-physiologic evidence has emergence that, working memory capacity is adaptive after adequate training as reflected in better task performance or greater standardized intelligence scores (IQ scores). However, unified conclusions has not yet reached because experimental results from behavioral studies were inconsistent: the IQ scores didn′t increase after training in some studies while some did. The reason may rest upon the fact that, the neurophysiologic plasticity after training is not necessarily related to the IQ scores. In this study, we aimed to find a estimate of working memory training effect by electroencephalography (EEG) and the event-related synchronization/ desynchronization (ERS/ERD) method.
Seventeen healthy university students were recruited and randomly divided into two groups (Control:8 and Experimental:9). The experimental group was required to play a spatial span task-liked puzzle game to train the WM system while the control group played only a movement-related game. The training intensity was 30 minutes a day, 5 days a week for 3 weeks for both groups. EEG data (Sampling rate = 2000 Hz, down sampling to 200Hz) from all subjects were recorded during spatial n-back tasks(n=0,1,2,3) at different training phases (pre- and post-training). EEG data were transformed into time-frequency signal by Morlet wavelet transform and divided into four frequency bands (Low theta 4-6Hz、High theta 6-8Hz、Low alpha 8-10Hz、High alpha 10-12Hz) for further analysis. The ERD/ERS of these four frequency bands were computed for each subject and ANOVA test were performed to find the difference between groups. In addition, the behavior data of reaction time (RT) and response correct rate also entered the statistical test.
The statistical test on behavioral data shows that there is no significant difference between two groups. This result is in line with some previous studies that the behavioral result may not be able to reflect the WM training effect. By using the regression analysis, we found that there exists a nonlinear relation between memory load and the behavioral data, irrespective to training.
The EEG results show that the WM training effect (i.e. the difference between two groups) was most manifest in posterior ERD/ERS (low theta, 0~500 ms). Furthermore, when taking into account the lateralization factor, we found a significant between-group difference of unilateral dominance over central (low theta, 500~1500ms) and posterior regions (low theta、low alpha 250~1000ms) after training. Collectively, our findings suggest that working memory training would induce the neurophysiological changes, though it is not necessarily reflected in the behavior data. In the future work, we will be looking for the mechanisms underlying these training changes.