Selesnick S A, Piccinini Gualtiero
Department of Mathematics and Computer Science, University of Missouri - St. Louis, St. Louis, Missouri, 63121, USA.
Department of Philosophy, Associate Director, Center for Neurodynamics, University of Missouri - St. Louis, St. Louis, Missouri, 63121, USA.
J Biol Phys. 2018 Dec;44(4):501-538. doi: 10.1007/s10867-018-9504-9. Epub 2018 Jun 8.
In earlier work, we laid out the foundation for explaining the quantum-like behavior of neural systems in the basic kinematic case of clusters of neuron-like units. Here we extend this approach to networks and begin developing a dynamical theory for them. Our approach provides a novel mathematical foundation for neural dynamics and computation which abstracts away from lower-level biophysical details in favor of information-processing features of neural activity. The theory makes predictions concerning such pathologies as schizophrenia, dementias, and epilepsy, for which some evidence has accrued. It also suggests a model of memory retrieval mechanisms. As further proof of principle, we analyze certain energy-like eigenstates of the 13 three-neuron motif classes according to our theory and argue that their quantum-like superpositional nature has a bearing on their observed structural integrity.
在早期的工作中，我们奠定了基础，以解释在类神经元单元簇的基本运动学情况下神经系统的类量子行为。在此，我们将这种方法扩展到网络，并开始为它们发展一种动力学理论。我们的方法为神经动力学和计算提供了一个新颖的数学基础，该基础从较低层次的生物物理细节中抽象出来，转而关注神经活动的信息处理特征。该理论对诸如精神分裂症、痴呆症和癫痫等病症做出了预测，并且已经积累了一些相关证据。它还提出了一种记忆检索机制的模型。作为进一步的原理证明，我们根据我们的理论分析了13种类三神经元基序类的某些类能量本征态，并认为它们的类量子叠加性质与观察到的结构完整性有关。
