透過您的圖書館登入
IP:3.145.89.60
  • 學位論文

視網膜中的預測行為與編碼

Prediction and Coding for Temporal Patterns in the Retina

指導教授 : 嚴震東
共同指導教授 : 陳志強(C.K. Chan)
若您是本文的作者,可授權文章由華藝線上圖書館中協助推廣。

摘要


編碼隨時間變化的外界刺激並產生預期性的反應是神經系統重要的特性。視網 膜這樣的初級視覺系統,便能在一串週期性刺激停止後產生遺缺刺激反應,此反 應同步化、產生於準確的時間點、且能預測刺激的動態。然而,遺缺刺激反應的 生物物理機制以及在非周期性刺激下的預測行為仍不清楚。此研究中,我們重複 視網膜中遺缺刺激反應的實驗,並將實驗結果與一種預期性動力學的模型比較, 最後以信息理論推廣在複雜動態刺激下的預測行為。我們以多通道電極陣列量測 牛蛙視網膜在不同動態光刺激下的反應,顯示視網膜在刺激週期為100-250毫秒時 能產生具有預測性的遺缺刺激反應。藉由週期性刺激後給予的探測刺激,發現視 網膜產生遺缺性刺激的時間尺度長達3秒,此時間尺度的適應行為可能與突觸的 鈣離子動態有關。而在包含多重週期的刺激停止後,所引起的反應相對去同步化 且有高歧異度。以隨機過程刺激視網膜,我們計算隨機光刺激週期與視網膜反應 在不同時間點的互信息、量化預期信息並推廣視網膜在連續刺激下的預測行為。 視網膜的預期行為與刺激的統計性質有關、且能偵測隨機刺激中的隱藏變量以進 行預測,而這些時間序列的統計特性能對應暫態反應(即原先的遺缺刺激反應實 驗)中所量測的時間尺度。透過數值方法模擬,我們進一步探討視網膜預測時間 序列的行為與可能機制。

並列摘要


Encoding time-dependent inputs and generating predictive activities are fun- damental properties in the nervous systems. Omitted stimulus response (OSR), a synchronized activity preserved after a periodic entrainment terminates, has been observed in primary visual systems such as the retina. OSR sensitively detects change and precisely predicts the upcoming stimulus patterns. However, the un- derlying biophysical mechanisms for OSR and responses to more general temporal patterns are still unknown. In this study, we repeated experiments for OSR, com- pared the behavior with an adaptive model that simulates OSR, and investigated predictive performance under stochastic stimuli. Experiments were operated with multiple electrode array recording for the bullfrog retina under programmed light stimuli. We show that OSR occurs in a dynamic range, when the period of stimuli is 100-250 ms. A probe provided after the periodic entrainment reveals the time scale of adaptation, showing preserved tendency to produce OSR after time delay up to 3 seconds, which might relate to the time scale of synaptic calcium dynamics. Under complex temporal patterns with multiple periods, the post-stimulus response is less synchronized and heterogeneous. By calculating the mutual information be- tween the input stochastic intervals and the retinal activity at different time shifts, we could quantify predictive information and characterize the predictive behaviors under stationary responses. It is shown that the predictive behavior depends on the statistics of the stimuli and the retina could detect the hidden variable in the stochastic process to make prediction. The time scales identified in the transient OSR phenomenon are observed in the predictive behavior under stationary responses and numerical methods were applied to implement possible mechanisms for temporal prediction.

參考文獻


[1] Tom Baden, Philipp Berens, Katrin Franke, Miroslav Román Rosón, Matthias Bethge, and Thomas Euler. The functional diversity of retinal ganglion cells in the mouse. Nature, 529(7586).
[2] H. B. Barlow. Summation and inhibition in the frog’s retina. The Journal of Physiology, 119(1):69–88, January 1953.
[4] Michael J. Berry, Iman H. Brivanlou, Thomas A. Jordan, and Markus Meister. Anticipation of moving stimuli by the retina. Nature, 398(6725):334–338, 1999.
[9] William Bialek, Ilya Nemenman, and Naftali Tishby. Predictability, complexity, and learning. Neural computation, 13(11):2409–2463, 2001.
[12] Stewart A. Bloomfield and Béla Völgyi. The diverse functional roles and regu- lation of neuronal gap junctions in the retina. Nature Reviews. Neuroscience, 10(7):495–506, July 2009.

延伸閱讀