Algorithm 1: The Pipeline of the Image Generation Stage
1: input EEG $\mathbf{x}\in {\mathbb{R}}^{1\times 128\times 440}$
2: do: Feature extraction and position embedding $\to {\mathbf{x}}_0\in {\mathbb{R}}^{77\times 768},{\mathbf{x}}_0\backsim q\left({\mathbf{x}}_0\right)$
3: do: FrozenCLIPEmbedder $\to \mathcal{E}\left({\mathbf{x}}_0\right)$
4: for $\mathrm{t}=1\cdots \mathrm{T}$, do:$\mathrm{q}\left({\mathbf{x}}_1:{\mathbf{x}}_T|\mathcal{E}({\mathbf{x}}_0\right) :=\prod _{t=1}^{T}q\left({\mathbf{x}}_t|{\mathbf{x}}_{t-1}\right) \to {\mathbf{x}}_T\backsim \mathcal{N}\left(\mathrm{0,1}\right)$
5: for $\mathrm{t}=\mathrm{T}\cdots 1$, $z\backsim \mathcal{N}\left(\mathrm{0,1}\right)$do:
6: ${\mathbf{x}}_{t-1}:=\frac{1}{\sqrt{{\alpha }_t}}\left({\mathbf{x}}_t-\frac{1-{\alpha }_t}{\sqrt{1-\overline{{\alpha }_t}}}{ϵ}_{\theta }\left({\mathbf{x}}_t,t\right)\right)\to \mathcal{E}({\mathbf{x}}_0)\in {\mathbb{R}}^{h\times w\times 3}$
7: do: AutoEncoderKL
8: $\mathrm{r}\mathrm{e}\mathrm{t}\mathrm{u}\mathrm{r}\mathrm{n} {\mathbf{x}}_0\in {\mathbb{R}}^{H\times W\times 3}$