Figure 1.

Procedures to implement a decoding model using linear support vector machines (SVMs) to analyze frequency-following responses (FFRs) to linguistically relevant pitch patterns (Mandarin tones): T1, high-level; T2, low-rising; T4, high-falling. This SVM analysis approach was reported in Xie et al. (2018). Leave-one-fold-out: The linear SVM classifier (model) is estimated with three of the four folds to classify FFRs into one of the three tone categories and is validated to see how well it can generalize to FFR data in the held-out fold. Decoding accuracy reflects the percentage that the SVM model correctly identified the tone categories across the four FFR folds.