Recently, Hu and Iannetti [1] discussed pitfalls of using machine learning (ML) for predicting pain from neuroimaging data. We agree that it is important to establish dos and don’ts for using ML in pain neuroimaging to avoid spurious conclusions. However, we believe that several opinions they offer are based on flawed premises; here, we provide a brief counterpoint.

中文翻译：

---

疼痛预测中的问题——收益大于痛苦

最近，Hu 和 Iannetti [1] 讨论了使用机器学习 (ML) 从神经影像数据预测疼痛的陷阱。我们同意，重要的是要确定在疼痛神经影像学中使用 ML 的注意事项，以避免得出虚假结论。然而，我们认为他们提供的一些意见是基于有缺陷的前提；在这里，我们提供一个简短的对立面。