A fundamental goal of system neuroscience is to understand how the brain processes sensory information. In this project we go beyond the classical open-loop approach and study the visual system in a closed-loop manner by using advanced optimization algori

系统神经科学的一个基本目标是了解大脑如何处理感觉信息。然而，经过几十年的研究，我们仍然没有能够完全捕捉初级视觉皮层神经元反应选择性的模型。目前使用线性系统理论的开环方法在表征非线性神经元方面存在不足，而非线性神经元是初级视觉皮层中普遍存在的神经元类型。此外，关于刺激如何在单个神经元和/或神经元群体的活动中被编码的讨论也存在争议。非线性神经元的表征和对单个神经元与群体活动之间关系的更好理解将极大地促进我们对初级视觉皮层和一般皮层功能的了解。突破经典的开环方法，以闭环的方式根据神经元的反应来改变刺激是解决这些问题的关键。进化计算领域的进步产生了强大的优化算法，例如利用群体智能的算法。这些算法可以处理高维参数空间和具有多模态参数分布和自适应的非线性和噪声目标函数。因此，它们非常适合优化基于神经元反应的视觉刺激。在这个项目中，我们将以闭环的方式研究单个神经元和神经元群。我们将使用优化算法来构建基于单个神经元反馈的最优刺激集合。这将使我们能够研究单个神经元的特性及其与嵌入群体的关系，并提高我们对感觉处理原理的认识。

A fundamental goal of system neuroscience is to understand how the brain processes sensory information. Yet, after decades of research we still do not have models that can completely capture the response selectivity of neurons in primary visual cortex.Current open-loop methods that make use linear-system theory fall short in characterizing non-linear neurons, the prevailing neuron type in the primary visual cortex. Furthermore, it is controversial discussed how stimuli are encoded in the activity of single neurons and/or populations of neurons. Characterization of non-linear neurons and a better understanding of the relationship between single neuron and population activity would greatly advance our knowledge about the function of the primary visual cortex and cortex in general.Going beyond the classical open-loop approach by modifying the stimulus based on the neuronal responses in a closed-loop way is the key to these questions. Advances in the field of evolutionary computation have produced powerful optimization algorithms, e.g. algorithms that make use of swarm intelligence. These algorithms can cope with high dimensional parameter spaces and non-linear and noisy objective functions that show multimodal parameter distributions and adaptation. Thus they are very suited to optimize visual stimuli based on neuronal responses.In this project we will study single neurons and populations of neurons in a closed-loop fashion. We will use the optimization algorithms to construct optimal stimulus ensembles based on feedback of single neurons. This will allow us to investigate single neuron properties and their relationship to the embedding population and advance our knowledge about the principles of sensory processing.