“…To further assess the behavioral relevance of α -FFR modulations to SIN listening, we investigated vowel decoding of FFRs at low vs. high α states. Similar ML approaches have been applied to speech-evoked FFRs to decode stimulus classes, e.g., Mandarin lexical tones (Llanos, Xie and Chandrasekaran, 2017; Xie et al, 2019) and speech tokens (Xie, Girshick, Dollár, Tu and He, 2017; Cheng, Xu, Gold and Smith, 2021). In these studies, decoding performance in correctly classifying FFRs is used as an objective measure of speech discrimination.…”