互信息介绍及MATLAB代码实现

互信息(Mutual information): 一种量化两个随机变量的非线性依赖性的信息理论度量方法

First, we evaluate the similarity between each connectome harmonic and each of the 7 resting state networks using mutual information, an information theoretical measure quantifying the non-linear dependence of two random variables, computed as
MI⁡(ψj,f{ RSN})=∑vi∈Vp(ψj(vi),f{ RSN}(vi))log⁡p(ψj(vi),fRSN(vi))p(ψj(vi))⋅p(fRSN(vi)) \operatorname{MI}\left(\psi_{j}, \mathrm{f}_{\{\mathrm{RSN}\}}\right)=\sum_{v_{i} \in \mathcal{V}} p\left(\psi_{j}\left(v_{i}\right), \mathrm{f}_{\{\mathrm{RSN}\}}\left(v_{i}\right)\right) \log \frac{p\left(\psi_{j}\left(v_{i}\right), f_{\mathrm{RSN}}\left