论文标题
信息理论自由能作为情感潜力：情感价与复杂性和新颖性的函数
Information-Theoretic Free Energy as Emotion Potential: Emotional Valence as a Function of Complexity and Novelty
论文作者
论文摘要
这项研究扩展了我们先前提出的情感维度的数学模型（Yanagisawa等人，2019年，前计算Neurosci），以考虑感知的复杂性和新颖性，作为唤醒潜力的来源。假定Berlyne的唤醒电势（或逆U形曲线，所谓的Wundt曲线）的享乐功能。我们将唤醒电位建模为在感知刺激（或认可）之后要在大脑中处理的信息内容，我们将其称为“感觉惊人”。我们在数学上证明了感觉惊人代表自由能，这等同于信息增益（或来自新颖性的信息）和感知的复杂性（或复杂性信息）的总结，这是构成唤醒电位的分解变量。我们以视觉刺激（蝴蝶的轮廓形状）支持了以下假设，即感知新颖性和复杂性塑造了逆U形美感功能。我们讨论了自由能作为数学原理解释情感启动者的潜力。
This study extends the mathematical model of emotion dimensions that we previously proposed (Yanagisawa, et al. 2019, Front Comput Neurosci) to consider perceived complexity as well as novelty, as a source of arousal potential. Berlyne's hedonic function of arousal potential (or the inverse U-shaped curve, the so-called Wundt curve) is assumed. We modeled the arousal potential as information contents to be processed in the brain after sensory stimuli are perceived (or recognized), which we termed sensory surprisal. We mathematically demonstrated that sensory surprisal represents free energy, and it is equivalent to a summation of information gain (or information from novelty) and perceived complexity (or information from complexity), which are the collative variables forming the arousal potential. We demonstrated empirical evidence with visual stimuli (profile shapes of butterfly) supporting the hypothesis that the summation of perceived novelty and complexity shapes the inverse U-shaped beauty function. We discussed the potential of free energy as a mathematical principle explaining emotion initiators.
