The development of mathematical models to characterize perceptual and cognitive processes dates back almost to the inception of the field of psychology. Since the 1990s, human functional neuroimaging has provided for rapid empirical and theoretical advances across a variety of domains in cognitive neuroscience. In more recent work, formal modeling and neuroimaging approaches are being successfully combined, often producing models with a level of specificity and rigor that would not have been possible by studying behavior alone. In this review, we highlight examples of recent studies that utilize this combined approach to provide novel insights into the mechanisms underlying human cognition. The studies described here span domains of perception, attention, memory, categorization, and cognitive control, employing a variety of analytic and model-inspired approaches. Across these diverse studies, a common theme is that individually tailored, creative solutions are often needed to establish compelling links between multi-parameter models and complex sets of neural data. We conclude that future developments in model-based cognitive neuroscience will have great potential to advance our theoretical understanding and ability to model both low-level and high-level cognitive processes.

中文翻译：

---

将理论模型与功能性神经影像学相结合

描述感知和认知过程的数学模型的发展几乎可以追溯到心理学领域的诞生。自 1990 年代以来，人类功能性神经影像学为认知神经科学的各个领域提供了快速的经验和理论进步。在最近的工作中，正式建模和神经影像学方法正在成功结合，通常产生的模型具有一定程度的特异性和严谨性，这是单独研究行为无法实现的。在这篇综述中，我们重点介绍了近期研究的例子，这些研究利用这种组合方法为人类认知的潜在机制提供了新的见解。这里描述的研究跨越感知、注意力、记忆、分类和认知控制等领域，采用各种分析和模型启发的方法。在这些不同的研究中，一个共同的主题是通常需要单独定制的、创造性的解决方案来在多参数模型和复杂的神经数据集之间建立引人注目的联系。我们得出的结论是，基于模型的认知神经科学的未来发展将具有巨大的潜力，可以提高我们对低级和高级认知过程进行建模的理论理解和能力。