Algorithm 1: TabTransformer-Based Missing Value Imputation for EEG Amplitude Data.

Input: Dataset D with elements $d_{ij}$, where $d_{ij}=\mathrm{null}$ indicates a missing value     Output: Imputed Dataset $D^{\prime }$     Step 1: Data Organization and Preprocessing         Initialize: $D^{\prime }\leftarrow D$         Mark missing values: $\forall i,j\mathrm{where}{d}_{ij}=\mathrm{null},\mathrm{set}{m}_{ij}=1$         Assign unique IDs: $ID\left[i\right]\leftarrow i$ for each row i in D         Separate features: $X_{\mathrm{Data}},Y_{\mathrm{Data}}\leftarrow \mathrm{split}(D,\mathrm{features})$     Step 2: Model Training and Prediction     Step 3: Reorganize         $D^{\prime }\leftarrow \mathrm{sort}(D^{\prime },ID)$     Step 4: Validation and Comparison       Initialize LSTM model $\mathcal{L}$       $R\leftarrow \mathrm{list}\mathrm{of}\mathrm{other}\mathrm{imputation}\mathrm{results}$       $R.\mathrm{append}\left(D^{\prime }\right)$