FEM modeling for tDCS

• 对于 tDCS 电流在大脑中的建模，主要是靠解决下面这个等式来的：
  $$M \cdot U = I$$
  
  • $M$ 是tDCS的forward matrix
  • $I$ 是injected current vector
  • $U$ 是在已知以上输入电流情况下，要求解的potential vector
  • $M$ combines volume conduction properties (FEM stiffness matrix) and electrical boundary conditions， including the complete electrode model to simulate current injection
    $$\nabla \cdot (\sigma \nabla u) = 0, \; (x \in \Omega)$$
    $$u + z_l \sigma \frac{\partial u}{\partial n} = U_l, \; (on \; \partial \Omega_{e_l}, x\in e_l)$$
    $$\int_{e_l} \sigma \frac{\partial u}{\partial n} = I_l (l = 1,2,\cdots, L)$$
    $$\sigma \frac{\partial u}{\partial n} = 0, \; (x\in \partial \Omega \backslash \cup_{l=1}^L e_l )$$
  • In more detail, $M$ is composed of the regular FEM stiffness ($A \in N \times N$, with $N$ representing the number of FEM nodes) and additional electrical boundary conditions specified as submatrices $A_2, B, C$ ($B \in N\times L, C \in L\times L$, with $L$ representing the number of electrodes)
  • $M$ is composed such as:
    $$M = \begin{pmatrix}A \quad -B \\ -B^T \quad C \end{pmatrix}$$
    $$\begin{pmatrix} A \quad -B \\ -B^T \quad C\end{pmatrix} \begin{pmatrix}U _n \\U_e\end{pmatrix} = I = \begin{pmatrix} I_n \\ I_e\end{pmatrix}$$
    $$A(i,j) = A_1(i,j) + A_2(i,j) = \int_\Omega \sigma \nabla \phi_i \cdot \nabla \phi_j d\Omega + \sum^L_{l=1} \int_{e_l} \frac1{z_l}\phi_i \phi_j d\Omega_{e_l}$$
    $$B(i,l) = \frac1{z_l} \int_{e_l}\phi_i d\Omega_{e_l}$$
    $$C(i,l) = \frac1{z_l} \int_{e_l}d\Omega_{e_l}$$
  • $M \in \mathcal{R}^{N+L\times N+L}, A_1, A_2 \in \mathcal R^{N\times N}, B \in \mathcal R^{N\times L}, C \in \mathcal R^{L\times L}$
  • $U_n$ represents potentials at nodes, $U_e$ represents potentials at electrodes, $I_e$ is currents at electrodes, $\phi$ denotes linear basis functions, $z_l$ resistive electrode impedance.
  • -

Reference:
https://www.sci.utah.edu/images/software/SCIRun/ElectricalBrainStimulationTutorial.pdf