行政院國家科學委員會專題研究計畫成果報告:記憶系統中之自動與控制過程 -- 以電腦化的神經心理檢查與事件相關電位為研究方法

Abstract

傳統上將認知功能時的覺知程度二分成自動化或控制性模式。熟悉的常規性工作，交給 自動化模式去處理。複雜的工作則會啟動控制性模式，比如在處理上超出日常常規性工 作之複雜度者，其次舊資料需用新方式加以處理時，或者必須加以創新處理的新資訊。 控制性模式的量測，以處理”隨機非結構”表徵之資訊時的反應加以量測。自動化模式的 量測則針對俱”結構性序列”表徵之訊息的處理反應加以量測。事件相關電位則以32 頻道 之NeuroScan 腦波機予以記錄。自動與控制則以240 毫秒正波時之平均值以雙楔內插法 運算得之，並計算其組間T 值腦電波圖譜。刺激用之幾何圖形呈現在15 吋之液晶銀幕上 並以60 公分之距離來觀看。每圖呈現一秒並有300 毫秒之空白間格。對於第一組(8 個 圖形)呈現之圖形，受試者須全神貫注並加以牢記。接著受試者須在第一及第二組(另8 個圖形)隨機交錯出現之圖形進行辨認及”是”、”否”出現過之回答。第三階段辨認測試時 又加入16 個從未出現過之圖形，此時並同時記錄事件相關電位腦波，每組圖形會有總共 64 腦波之記錄。 實驗結果顯示，處理俱”結構性序列”表徵之訊息的反應時間(平均時長: 17.1 ± 3.1 秒)較 處理”隨機非結構”表徵之資訊時的反應時間(平均時長: 15.5 ± 3.5 秒)為長。兩者之相差 顯著(P=0.0002)。反應正確率也以處理”隨機非結構”表徵之資訊時為高(”隨機非結構” : 6.6 ± 4.3; ”結構性序列” : 7.4±3.8)，兩者相差亦顯著(P=0.00569)。 事件相關電位腦波圖譜的研究，支持較熟悉之自動化制性模式其大腦部位在較後 區，而不熟悉之新訊息的控制性模式則需動員大腦前區之資源加以應對處理。總之，本 研究計畫運用電腦化的神經心理檢查與事件相關電位的研究方法探討了記憶系統中之自 動與控制的過程並見證了其表現與其在大腦內的活性變化。

The conventional concept about the level of awareness during cognitive processing can be divided in a dichotomy way into automatic and controlled processes. Routine tasks that we are rather familiar with go through automatic process. Complex tasks invoke controlled process when the level of complexity of a task requires more than routine processing, when old information must be considered in new ways or when the information is to be processed de novo. The measurement of the controlled process of the memory system was performed by using tasks with “random-disorganized” features, which invoked more awareness in handling the more complex situation. The measurement of the automatic process of the memory system was performed by using tasks with “structured-organized” features, which utilized less awareness in handling the less complex situation. ERP were recorded with a 32- channel ERP machine. Brain mappings were prepared at the mean latencies of P240 with the grand average of ERP using double spline interpolation. Between-group t-Maps were also constructed. Geometry shapes were presented to the subjects in a 15-inches LCD screen at about 60-cm distance. The shapes of the first group were presented for one second with an inter-break of about 300 ms. Subjects were instructed to pay full attention and to memorize them by heart. After the presentation the subjects were asked to recognize and to do a yes-no response on the shapes of the first group from a randomized presentation of shapes from the first (8 forms) and the 2nd group (8 forms). ERP were recorded during the 3rd stage of presentation and recognition test. Another 16 novel shapes that have never been presented joined in a random presentation sequence with the 1st and 2nd groups. In total 64 trials ERP were collected from each group for further analysis. It took longer to complete the “structured-organized” task (mean duration: 17.1 ± 3.1 sec) than to complete the “randomized-disorganized” task (mean duration: 15.5 ± 3.5 sec) with a statistic significant difference (p = 0.00002). The mean scores (correct – omission – commission) were lower (p = 0.00569) with “structured-organized” task (6.6 ± 4.3) than with “randomized- disorganized” task (7.4±3.8). In ERP study, the three conditions had similar waveforms especially the early components while their brain topographic mappings varied in the spatial distribution. The well-memorized targets induced activities over the bilateral parietal cortexes more on the right side. Those implicitly learned targets or de novo targets induced some frontal activities appeared in addition to parietal activities. In conclusion, “controlled process” recruits more brain resources but performs better in terms of speed and correct rates. ERP showed results compatible with that brain regions involved in this kind of automatic process are more posterior in location while the controlled process involves brain areas mostly within the frontal lobe.