neural/linear/stats.py [293:346]:
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
    axes[0, 2].locator_params(axis='x', nbins=20)
    axes[0, 2].locator_params(axis='y', nbins=10)
    axes[0, 2].grid()
    axes[0, 2].axvline(start, ls="--")
    axes[0, 2].text(x=start, y=0, s="init")

    # Scalar Correlation score
    # axes[1, 2].bar([0, 1, 2], [0, r_scalar, 0])
    # axes[1, 2].set_title("Correlation Score")

    # Distributional Correlation score
    # epoched
    scores = r_average_times.T.flatten()
    pca_labels = np.concatenate(
        [[idx] * r_average_times.shape[0] for idx in range(r_average_times.shape[1])])
    df = pd.DataFrame({"scores": scores, "pca_labels": pca_labels})
    sns.boxplot(x="pca_labels", y="scores", data=df, ax=axes[1, 2])
    axes[1, 2].set_title("Overall Correlation")
    # evoked
    scores = r_average_times_evoked.mean(0)
    pca_labels = np.arange(r_average_times.shape[-1])
    axes[1, 2].plot(pca_labels, scores, label="corr of the trial-mean")
    axes[1, 2].legend(bbox_to_anchor=(0, 1), loc='upper left', ncol=1)

    # Dynamic MSE score
    axes[0, 3].plot(mse_dynamic_epochs.mean(-1), label="epoch-wise mse")
    # axes[0, 2].plot(r_dynamic_evoked, label="evoked-wise correlation")
    axes[0, 3].plot(mse_average_epochs.mean(-1), label="baseline mse")
    axes[0, 3].legend()
    axes[0, 3].set_title("Relative MSE along time")
    axes[0, 3].set_ylim(0, 1)
    axes[0, 3].locator_params(axis='x', nbins=20)
    axes[0, 3].locator_params(axis='y', nbins=10)
    axes[0, 3].grid()
    axes[0, 3].axvline(start, ls="--")
    axes[0, 3].text(x=start, y=0, s="init")

    # Distributional MSE score
    # epoched
    scores = mse_average_times.T.flatten()
    pca_labels = np.concatenate(
        [[idx] * mse_average_times.shape[0] for idx in range(mse_average_times.shape[1])])
    df = pd.DataFrame({"scores": scores, "pca_labels": pca_labels})
    sns.boxplot(x="pca_labels", y="scores", data=df, ax=axes[1, 3])
    axes[1, 3].set_title("Overall Relative MSE")
    # evoked
    scores = mse_average_times_evoked.mean(0)
    pca_labels = np.arange(mse_average_times.shape[-1])
    axes[1, 3].plot(pca_labels, scores, label="rel. MSE of the trial-mean")
    axes[1, 3].legend(bbox_to_anchor=(0, 1), loc='upper left', ncol=1)

    plt.tight_layout()
    plt.savefig(path)
    plt.close()
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -



neural/visuals.py [126:179]:
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
    axes[0, 2].locator_params(axis='x', nbins=20)
    axes[0, 2].locator_params(axis='y', nbins=10)
    axes[0, 2].grid()
    axes[0, 2].axvline(start, ls="--")
    axes[0, 2].text(x=start, y=0, s="init")

    # Scalar Correlation score
    # axes[1, 2].bar([0, 1, 2], [0, r_scalar, 0])
    # axes[1, 2].set_title("Correlation Score")

    # Distributional Correlation score
    # epoched
    scores = r_average_times.T.flatten()
    pca_labels = np.concatenate(
        [[idx] * r_average_times.shape[0] for idx in range(r_average_times.shape[1])])
    df = pd.DataFrame({"scores": scores, "pca_labels": pca_labels})
    sns.boxplot(x="pca_labels", y="scores", data=df, ax=axes[1, 2])
    axes[1, 2].set_title("Overall Correlation")
    # evoked
    scores = r_average_times_evoked.mean(0)
    pca_labels = np.arange(r_average_times.shape[-1])
    axes[1, 2].plot(pca_labels, scores, label="corr of the trial-mean")
    axes[1, 2].legend(bbox_to_anchor=(0, 1), loc='upper left', ncol=1)

    # Dynamic MSE score
    axes[0, 3].plot(mse_dynamic_epochs.mean(-1), label="epoch-wise mse")
    # axes[0, 2].plot(r_dynamic_evoked, label="evoked-wise correlation")
    axes[0, 3].plot(mse_average_epochs.mean(-1), label="baseline mse")
    axes[0, 3].legend()
    axes[0, 3].set_title("Relative MSE along time")
    axes[0, 3].set_ylim(0, 1)
    axes[0, 3].locator_params(axis='x', nbins=20)
    axes[0, 3].locator_params(axis='y', nbins=10)
    axes[0, 3].grid()
    axes[0, 3].axvline(start, ls="--")
    axes[0, 3].text(x=start, y=0, s="init")

    # Distributional MSE score
    # epoched
    scores = mse_average_times.T.flatten()
    pca_labels = np.concatenate(
        [[idx] * mse_average_times.shape[0] for idx in range(mse_average_times.shape[1])])
    df = pd.DataFrame({"scores": scores, "pca_labels": pca_labels})
    sns.boxplot(x="pca_labels", y="scores", data=df, ax=axes[1, 3])
    axes[1, 3].set_title("Overall Relative MSE")
    # evoked
    scores = mse_average_times_evoked.mean(0)
    pca_labels = np.arange(mse_average_times.shape[-1])
    axes[1, 3].plot(pca_labels, scores, label="rel. MSE of the trial-mean")
    axes[1, 3].legend(bbox_to_anchor=(0, 1), loc='upper left', ncol=1)

    plt.tight_layout()
    plt.savefig(path)
    plt.close()
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -