The human brain distinguishes speech sounds by mapping acoustic signals into a latent perceptual space. This space can be estimated via multidimensional scaling (MDS), preserving the similarity structure in lower dimensions. However, individual and group-level heterogeneity, especially between native and non-native listeners, remains poorly understood. Prior approaches often ignore such variability or cannot capture shared structure, limiting principled comparisons. Moreover, the literature often focuses on latent distances rather than the underlying features themselves. To address these issues, we develop a Bayesian mixed MDS method that accounts for both subject- and group-level heterogeneity, allows for the recovery of unique, identifiable latent features, facilitating their biological interpretability, while also determining the effective dimensionality of the latent space in an automated, data-adaptive manner. Simulations and an auditory neuroscience application demonstrate how these features reconstruct observed distances and vary with individual and language background, revealing novel insights.

翻译：人脑通过将声学信号映射至潜在感知空间来区分语音信号。该空间可通过多维缩放（MDS）进行估计，从而在低维空间中保留相似性结构。然而，个体与群体层面的异质性（尤其是母语与非母语听众之间的差异）仍未被充分理解。现有方法常忽略此类变异或无法捕捉共享结构，限制了原理性比较。此外，现有文献多关注潜在距离而非潜在特征本身。为解决上述问题，我们提出了一种贝叶斯混合MDS方法，该方法能同时解释受试者与群体层面的异质性，实现唯一可辨识潜在特征的恢复以增强其生物学可解释性，并通过自适应数据驱动方式自动确定潜在空间的有效维度。仿真实验与听觉神经科学应用表明，这些特征如何重构观测距离、如何随个体差异与语言背景变化，揭示了全新的认知洞见。