Table 4.

LME model of prediction accuracy and amplitude and latency of the TRF peaks in function of speech rate for the linguistic model^a

		Rate			Rate2
		β	CI (95%)	p	β	CI (95%)	p
Prediction accuracy		−4.59 × 10– 5	±3.31 × 10– 5	0.0081	Does not improve AIC
Phoneme surprisal	TRF amplitude	6.81 × 10– 3	±2.82 × 10– 3	<0.001	Does not improve AIC
	TRF latency	5.47 × 10– 3	±4.20 × 10– 3	0.015	Does not improve AIC
Cohort entropy	TRF amplitude	4.21 × 10– 3	±3.07 × 10– 3	0.008	Does not improve AIC
	TRF latency	3.59 × 10– 3	±5.11 × 10– 3	NS	Does not improve AIC
Word surprisal	TRF amplitude	6.46 × 10– 3	±3.09 × 10– 3	<0.001	Does not improve AIC
	TRF latency	−3.33 × 10– 3	±5.38 × 10– 3	NS	Does not improve AIC
Word frequency	TRF amplitude	6.44 × 10– 3	±3.00 × 10– 3	<0.001	Does not improve AIC
	TRF latency	−1.09 × 10– 4	±3.06 × 10– 3	NS	Does not improve AIC

^aEvery line indicates a different model.