計算神經行為學 ─ 發展虛擬蟲系統以及其體內能量恆定機制

探究生物如何決定執行某特定行為的原因，可由神經系統、生物行為與環境間關係的研究開始，在電腦技術蓬勃發展的當下，更可以透過電腦模擬的輔助，來達成目標，此即為計算神經行為學的基本應用。因此，本實驗室致力於開發能夠達成該目標的模擬器─Hanitu，這是一套神經網路與環境的模擬系統，此系統模擬一個二維的世界，其中有數個食物與毒物來源以及任意數量的虛擬蟲。使用者須為虛擬蟲設計神經迴路，目標為賦予虛擬蟲有能力以嗅覺偵測環境中的食物或毒物，產生動作反應前往食物的位置並且獲取食物，避開毒物，以維持自身生命值，達到持續存活在環境中的目的。以先前開發的版本為基礎，本專案以模組化的設計方式(modular design)重新建構Hanitu系統，分成圖形化介面(GUI)、Hanitu(環境模擬)與Flysim(神經元模擬器)三大主軸，此架構有助於未來的功能擴充與版本維護。為使虛擬蟲之模擬能更進一步貼近真實生物，後續版本中做了部分更動：更新嗅覺系統與新增體內能量恆定系統，前者實現了單一嗅覺神經元(olfactory receptor neuron)可以接受多種氣味分子刺激的機制；後者則加入了生物體內的能量恆定相關賀爾蒙─Ghrelin的模擬，並於虛擬蟲體內增設了接受Ghrelin刺激之NPY(neuropeptide Y)神經元，虛擬蟲將因自身體力值的增減而影響其攝食行為，另外，本專案也呈現了不同虛擬蟲神經迴路於加入NPY神經元後，展現之不同行為模式，說明生物體內能量高低與行為改變之關係。

關鍵字

計算神經科學；神經行為學；飢餓肽；神經迴路；神經模擬；決策

並列摘要

Neuroethology is the study of animal behavior with an emphasis on the underlying neural mechanisms and interactions between subjects and environment. A great approach to address questions in neuroethology that cannot be addressed by experiment and observation has arisen due to the booming development of computational neuroscience. From this, our lab has developed a software tool, known as the Hanitu system. The system is a simulation environment that simulates the behavior and neural activity of user-designed virtual worms in a two-dimensional virtual world. Hanitu was developed around a central idea, which given a challenging environment, how should the nervous system be designed for a virtual worm to forage for foods, avoid toxicants, and reach the ultimate goal – to survive. This thesis is focused on rebuilding the system based on modular design strategy for easy upgrading and maintenance. There are three core components of the Hanitu system as follows: the graphical user interface (GUI), Hanitu (the virtual world), and Flysim (the neural network simulator). There are two main functions added to the latest version of the Hanitu system: a cross-sensory mechanism and an energy homeostasis mechanism to simulate the more realistic behavior of animals. The former implements a phenomenon for a single olfactory receptor neuron that is active from multiple odorants .The latter that simulates the secretion of an internal energy related hormone, ghrelin, in animals and adds the neuropeptide Y (NPY) neuron which is stimulated by Ghrelin into the neural circuit of virtual worm. The results of the simulation illustrate that different behavior arise from the energy homeostasis mechanism.

並列關鍵字

computational neuroscience ； neuroethology ； ghrelin ； neural circuit ； neural simulation ； decision making

參考文獻

Brette, R., Rudolph, M., Carnevale, T., Hines, M., Beeman, D., Bower, J.M., Diesmann, M., Morrison, A., Goodman, P.H., Jr, F.C.H., Zirpe, M., Natschläger, T., Pecevski, D., Ermentrout, B., Djurfeldt, M., Lansner, A., Rochel, O., Vieville, T., Muller, E., Davison, A.P., Boustani, S.E., Destexhe, A., 2007. Simulation of networks of spiking neurons: A review of tools and strategies. J. Comput. Neurosci. 23, 349–398. doi:10.1007/s10827-007-0038-6

Davis, M.J., 1997. Design as a Catalyst for Learning. Association for Supervision and Curriculum Development.

Day, M., Carr, D.B., Ulrich, S., Ilijic, E., Tkatch, T., Surmeier, D.J., 2005. Dendritic Excitability of Mouse Frontal Cortex Pyramidal Neurons Is Shaped by the Interaction among HCN, Kir2, and Kleak Channels. J. Neurosci. 25, 8776–8787. doi:10.1523/JNEUROSCI.2650-05.2005

Franklin, S., 1997. Artificial Minds. MIT Press.

Gantz, I., Fong, T.M., 2003. The melanocortin system. Am. J. Physiol. - Endocrinol. Metab. 284, E468–E474. doi:10.1152/ajpendo.00434.2002

國際替代計量

計算神經行為學 ─ 發展虛擬蟲系統以及其體內能量恆定機制

全文下載

主題瀏覽