Fig 1.
An example of a pupil size recording from our experiment, showing time points of where a rapid dilation was detected according to the ICA software (red triangles).
Fig 2.
Self-paced reading times for the grammatical gender violation experiment.
Critical regions for the experiment is located at word number 0.
Fig 3.
Raw ICA for the grammatical gender violation experiment.
Critical region onset for the experiment is located at word number 0; the grey bar marks the hypothesised critical region of 500ms, peaking at 1s. The difference between conditions in this region is significant at p < 0.05 on both eyes, see also Table 1.
Table 1.
Linear mixed effects models for both eyes in the self-paced reading experiment with grammatical gender match / mismatch manipulation.
Fig 4.
Self-paced reading times for the semantic anomaly experiment.
The critical region is located at word number 0.
Fig 5.
Raw ICA for the semantic fit experiment.
Critical region onset for the experiment is located at word number 0; the grey bar marks the region of 500ms, peaking at 1s. The difference between conditions is significant at p < 0.001 for the left eye and p < 0.01 for the right eye, see also Table 2.
Table 2.
Linear mixed effects models for semantic anomaly in SPR.
Fig 6.
Self-paced reading times for the RC Experiment.
Critical region hat/haben is located at word number 0. Extremely slow reading times in the ORC case are reflective of the high difficulty in this condition.
Fig 7.
Raw ICA for the relative clause experiment.
Critical region onset for the experiment is located at word number 0; the grey bar marks the hypothesized critical region of 500ms, peaking at 1s. For details, see Table 3.
Table 3.
Linear mixed effects models for ICA in SPR relative clause experiment.
Results for data from first half of experiment.
Table 4.
Difficulty settings used for the ConTRe steering task in our experiments.
Fig 8.
Cross-correlations for the ICA with the reference bar and the steering bar, respectively.
Cross-correlation values for the reference bar (top plots) and steering bar (bottom plots) with the ICA of the left eye (left plots) and right eye (right plots). Dashed lines indicate 95%CIs. The periodicity in the cross-correlation plot is due to the periodicity of the steering task. We can see that the correlation of the ICA is higher for steering (i.e. the driver’s own reaction) than for the reference bar (i.e., the stimulus).
Fig 9.
Raw ICA for the dual-task grammatical gender violation experiment.
Critical region onset for the experiment is located at word number 0; the grey bar marks the region of 500ms, peaking at 1s. The difference between conditions was statistically significant, see also Table 5.
Table 5.
Linear mixed effects models for left and right eye in dual task gender match/mismatch.
Fig 10.
Raw ICA for the dual-task semantic violation experiment.
Critical region onset for the experiment is located at word number 0; the grey bar marks the critical region of 500ms, peaking at 1s. Differences between conditions are not statistically significant for this region. However, there is an earlier significant effect for the 500ms time window peaking at 700ms after word onset (indicated here by dashed vertical bars); see Table 6.
Table 6.
Linear mixed effects models for left and right eye in dual task semantic match/mismatch experiment for a 500ms time interval peaking at 700ms post onset of the semantically anomalous word.
Fig 11.
Raw ICA for the dual-task relative clause experiment.
Critical region onset for the experiment is located at word number 0; the gray bar marks the hypothesized critical region of 500ms, peaking at 1s. The differences between the conditions are statistically significant at p < 0.05 for both eyes; see also Table 7.
Table 7.
Linear mixed effects models for left and right eye in dual task relative clause experiment.
Fig 12.
Raw ICA for the connective region of the visual world experiment.
Critical region onset for the experiment is located at word number 0. The grey bar marks the hypothesized critical region of 500ms, peaking at 1s. See also Tables 8 and 9.
Table 8.
Linear mixed effects models for left and right eye in dual task gender match/mismatch.
Table 9.
Linear mixed effects models for left and right eye in dual task gender match/mismatch; first half of experiment.