Skip to main content
. 2018 May 14;9:1877. doi: 10.1038/s41467-018-04235-3

Fig. 5.

Fig. 5

Predictive encoding of planned inflections prior to speech onset. a Example of a target word speech spectrogram. Neural data were aligned to speech onset (i.e., production of the target word) at time zero. b Time-course of neural modulation for the morpho-syntactic manipulations. The earliest divergence in activity across sites is displayed in gray, whereas peak divergence is displayed in black. The mean time points are marked by the vertical lines, whereas their standard deviations (across sites) are signified by horizontal lines. Note that although peak modulation based on syntactic inflection occurred shortly prior to speech onset, the earliest modulation based on the specific manipulation (e.g., past–present or single–plural) occurred up to 2 s prior to production. c Trial-by-trial decoding predictions over time for a representative subject. Below is the correct prediction performance for the syntactic vs. control trial conditions across all trials. Above are the representative principal components, again comparing the syntactic (red) and control (blue) trials