# 文档对话上下文

## 第 1 轮

### 你的问题

请只核查这篇论文的 Figure 3，不要总结其他章节。请逐项说明：1）图中 causes 的数据是怎样生成的；2）它展示了哪种模型与估计方法的对照；3）相邻正文和图注允许得出什么范围的结论。每项给出英文关键句和原文出处，并明确这个计算示例能否当作直接的人体实证。没有依据的内容保留为“未确认”。

### Doc2X 回答

1）图中 causes 的数据生成方式：Causes were sampled from the unit normal density to generate a response, which was used to recover the causes, given the parameters.（原因从单位正态密度中采样以生成响应，该响应用于在给定参数下恢复原因。）这表明数据是合成的计算示例。 [1]

2）展示的模型与估计方法对照：Figure 3 shows a comparative evaluation of ReML and DEM using the same data. The estimates are similar but not identical. This is because DEM hyperparameterises the covariance as a linear mixture of precisions, whereas the ReML scheme used a linear mixture of covariance components.（图3展示了使用相同数据的ReML和DEM的比较评估。估计值相似但不完全相同。这是因为DEM将协方差超参数化为精度的线性混合，而ReML方案使用了协方差分量的线性混合。）此外，图注指出 Example of estimation under a mixed-effects or hierarchical linear model... These models are also known as parametric empirical Bayes (PEB) models.（混合效应或分层线性模型下的估计示例……这些模型也称为参数经验贝叶斯模型。） [2][3][4][5][6]

3）相邻正文和图注允许得出的结论范围：Slight differences in the hyperparameter estimates (upper right), due to a different hyperparameterisation, have little effect on the conditional means of the unknown causes (lower right), which are almost indistinguishable.（由于不同的超参数化，超参数估计中的微小差异对未知原因的条件均值影响很小，后者几乎无法区分。）这意味着在该特定合成数据示例中，两种方法在恢复潜在原因方面表现高度一致。 [7]

关于是否可作为直接人体实证：未确认。文中明确说明这是基于“unit normal density”采样的合成数据示例（worked-example），用于演示算法性能，并未提及该具体图表涉及真实人体生理数据或临床验证。 [1]

### 原文依据

**[1] Hierarchical-Models-in-the-Brain.pdf**

> Causes were sampled from the unit normal density to generate a response, which was used to recover the causes, given the parameters.

**[2] Hierarchical-Models-in-the-Brain.pdf**

> Figure 3 shows a comparative evaluation of ReML and DEM using the same data.

**[3] Hierarchical-Models-in-the-Brain.pdf**

> The estimates are similar but not identical.

**[4] Hierarchical-Models-in-the-Brain.pdf**

> This is because DEM hyperparameterises the covariance as a linear mixture of precisions, whereas the ReML scheme used a linear mixture of covariance components.

**[5] Hierarchical-Models-in-the-Brain.pdf**

> Example of estimation under a mixed-effects or hierarchical linear model.

**[6] Hierarchical-Models-in-the-Brain.pdf**

> These models are also known as parametric empirical Bayes (PEB) models (left).

**[7] Hierarchical-Models-in-the-Brain.pdf**

> Slight differences in the hyperparameter estimates (upper right), due to a different hyperparameterisation, have little effect on the conditional means of the unknown causes (lower right), which are almost indistinguishable.