幾何錯覺的分類：一個心理解剖學的觀點

Abstract

幾何錯視的發生，起因於物理上沒有扭曲的圖形透過視覺系統的處理，產生扭曲的視覺效果。幾何錯視的發生階段究竟為何？是在視覺處理過程中的視覺位置登錄系統（V1）階段，還是在更高層次辨識物體形狀的階段發生的？本研究的研究方法延續柯伊芳（2002）的基礎稍作修正，由於視覺處理過程中，視覺訊號經由下游的V1再傳到偵測物體運動現象的運動系統（MT），根據此順序，實驗設計將有錯視誘發框架的圖形和沒有框架的圖形（本研究選用Müller-Lyer向內向外兩種、Poggendorff、Caféwall、Ebbinghaus、Zöllner、Hering錯視），在可以產生運動的時間差，作有位移的交替呈現，探討運動系統所接收到的圖形是否受到錯視誘發框架圖形影響，而產生形變的似動運動，若有形變的似動運動產生，則此錯視是因為V1的圖形位置訊息扭曲所造成，運動系統就會接收到已經扭曲的圖形來作交替呈現，也就代表此錯視是發生在V1階段；反之，若無形變似動運動產生，則此錯視在圖形位置訊息登錄時沒有被扭曲，沒有被扭曲的圖形在運動系統和另一個沒有被扭曲的圖形作交替呈現，也就代表此錯視的發生階段晚於MT。利用此假設的推測，我們可以知道不同的幾何錯視的發生階段，並且若有發生階段的不同，也可藉此方式來分類幾何錯視。本研究的結果顯示我們無法用運動系統作為分類的依據，七種幾何錯視都發生在運動系統計算之前，可能也就是V1的階段。

We aimed to investigate the anatomical/functional level of shape distortion induced by optical illusions in the current study. Specifically, we tried to determine whether the illusion exerts its effect at the level of position encoding (V1) or at a later stage where the more abstract and global form information was derived. The extraction of both form and motion involved the computation of position information. If the optical illusion exerted its effect at level of V1 where the most accurate position was available, the induced shape change should affect both form and motion perception. If, on the other hand, the shape distortion occurred at a later stage, the motion system could be immune to optical illusions. We used Hering, Müller-Lyer, Poggendorff, Caféwall, Ebbinghaus and Zöllner illusion to create perceptually distorted figures, and tested if these figures could drive the motion system to produce transformational apparent motion. All of results showed that the motion system was influenced by the illusions, indicating that these illusions probably occurred at the level of V1.