Figure 1.
Schematic illustrating experimental design and phase coherence analysis (for a single MEG channel).
The colors represent single-trial responses to each of the six audiovisual streams. The coherence analyses are performed on each of the 157 MEG channels separately. (a) The cross-trial phase coherence is calculated on all 15 trials of the same stimulus condition (same color) and compared to a mixture of trials (see Methods) to get the phase-based and power-based movie discrimination ability (see Figure 2ab). (b) Cross-movie phase coherence is calculated by combining response trials across two movie stimuli (two different colors in each column), where one dimension is matched in auditory (SameAud), visual (SameVis), or neither modality input (NoSame). See more equation details in Methods.
Figure 2.
Phase-based and Power-based movie discrimination ability.
Phase (a) and power (b) discrimination ability as a function of frequency (2–50 Hz) for 20 auditory (solid circles) and 20 visual channels (open circles) selected from separate auditory and visual localizer pretests for each participant. The gray box denotes the delta-theta range (2–7 Hz) selected for further analyses. The phase discrimination score in this range is significantly above 0. Error bars indicate the standard error across the 36 calculated samples (six stimulus conditions, six subjects).
Figure 3.
Low-frequency band phase pattern reflects within- and across-modality tracking.
Cross-movie response coherence (how similar are the responses elicited by two movies) in delta-theta phase pattern (a) and power pattern (b) for the 20 auditory and 20 visual channels selected from independent localizer pretests (see Figure 1b and Methods for analysis illustration). SameVis: movie clip pair sharing the same visual but different auditory input; SameAud: movie pair sharing same auditory but different visual input; NoSame: movie pair differing in both auditory and visual inputs. For example, for movie clip V1A1, the SameVis, SameAud, and NoSame movies correspond to V1A3, V2A1, and V3A2, respectively. Error bars indicate the standard error across six subjects.
Figure 4.
Low-frequency cross-movie phase coherence distribution map.
Delta-theta cross-movie phase coherence distribution map for each of the six subjects, indicating within-modality tracking. In this flat map of the MEG recordings, left is left, right is right, and red indicates larger cross-movie phase coherence. Left: distribution map for larger cross-movie delta-theta phase response coherence of SameVis movie pair versus SameAud movie pair. The comparison implicates occipital (visual) cortex. Right: distribution map for larger cross-movie delta-theta phase coherence of SameAud movie pair versus SameVis movie pair. This analysis shows auditory activation.
Figure 5.
Inter-trial low-frequency phase coherence depends on audiovisual temporal.
Cross-trial coherence in delta-theta phase pattern (a) and power pattern (b) for Matched (black bar) and Mixed (grey bar) stimulus conditions, in 20 auditory and 20 visual channels (see Figure 1a and Methods for analysis illustration). Cross-area (auditory and visual) coherence in delta-theta phase pattern (c) and power pattern (d) for Matched (black bar) and Mixed (grey bar) stimulus conditions. Error bars indicate the standard error across 10 subjects.
Figure 6.
Low-frequency phase-pattern-based classification performance.
Grand average of delta-theta-phase-based classification histograms for each of the six audiovisual stream conditions (3 matched and 3 mixed conditions) for auditory (a) and visual areas (b). Note that the sum of the clustered bar sums to 1. Error bars indicate the standard error across six subjects. (c) Generalization and statistical analysis of classification performance (ab).
Figure 7.
Low-frequency phase coherence and “optimal phase.”
(a) Grand average of phase histograms (x-axis) as a function of inter-trial delta-theta phase coherence (y-axis, 0∼1) across six subjects in auditory channels (upper) and visual channels (lower). Note that the sum of each row is 1. (b) Deviation score from uniform distribution as a function of inter-trial delta-theta phase coherence (x-axis, 0–1). Error bars indicate the standard error across subjects. (c) “Optimal” phase (0 and ±) fraction for matched (black bar) versus mixed (grey bar) conditions in auditory (upper) and visual (lower) channels. Error bars indicate the standard error across six subjects.