Large-scale neural models of cognitive function

我们已经开发出使用生物学上合理的刺突神经元来模拟认知行为的方法，从而产生了目前世界上最大的功能性大脑模型Spaun，该模型最近发表在《科学》杂志上（Eliasmith et al., 2012）。该模型是最近出版的《如何构建大脑》（Eliasmith, 2013）一书中详细描述的工作成果，该书概述了构建这种模型的一般方法和架构，称为语义指针架构（SPA）。我们现在遇到的最重要的挑战，也是本提案的重点，是扩大理论和模拟的规模，以解决更苛刻环境下的认知任务。

We have developed methods for simulating cognitive behaviour using biologically plausible spiking neurons that have resulted in what is currently the world's largest functional brain model, Spaun, which was recently published in the journal Science (Eliasmith et al., 2012). This model is the culmination of work that is described fully in the recent book “How to build a brain” (Eliasmith, 2013), which outlines a general method and architecture for constructing such models called the Semantic Pointer Architecture (SPA). The over-riding challenge that we now encounter, and the focus of this proposal, is scaling up both the theory and the simulations to tackle cognitive tasks in more demanding circumstances. Our central objective is to build biologically detailed models of cognition. We have identified three main challenges that we need to address to retain our position as leaders in the field. First, no group has yet employed highly realistic single neuron models (e.g. compartmental, conductance based models) in large-scale cognitive models while capturing psychological functions. Doing so provides the promise of better characterizing the relationship between biological and psychological properties allowing us to test new medical interventions (e.g. drugs, deep brain stimulation, etc.). Second, biologically based cognitive models are, by their very nature, large in scale and computationally demanding. Consequently, we intend to develop simulation tools and specialized hardware infrastructure that allow for significant speed and scale improvements, rivaling the best in the world. Third, understanding and exploiting the critical role that adaptation plays in a large, complex, functioning system is difficult. Challenges stem from ensuring that the system remains stable during adaptation, generalizing learned behaviours across tasks at the cognitive level, and understanding the relations between different kinds of adaptation evident across different domains (e.g. perception, motor control, and cognition). Addressing these three broad challenges will allow us to construct the next generation of `whole-brain' models. This research will further test a new theoretical framework for integrating perceptual, motor, and cognitive behaviour in biologically plausible models. The proposed research will significantly improve the adaptability and functionality of SPA models, and have the important practical consequence of more directly connecting models to medical interventions that can help with the diagnosis, prevention, and repair of damage to neural systems. Finally, understanding the unique solutions that nature has found for difficult real-world problems should pave the way for innovative engineering solutions in intelligent systems, pattern recognition, and decision-making applications.