The phi complex as a neuromarker of human social coordination

Abstract

Many social interactions rely upon mutual information exchange: one member of a pair changes in response to the other while at the same time producing actions that alter the behavior of the other. However, little is known about how such social processes are integrated in the brain. Here, we used a specially designed dual-electroencephalogram system and the conceptual framework of coordination dynamics to identify neural signatures of effective, real-time coordination between people and its breakdown or absence. High-resolution spectral analysis of electrical brain activity before and during visually mediated social coordination revealed a marked depression in occipital alpha and rolandic mu rhythms during social interaction that was independent of whether behavior was coordinated or not. In contrast, a pair of oscillatory components (phi(1) and phi(2)) located above right centro-parietal cortex distinguished effective from ineffective coordination: increase of phi(1) favored independent behavior and increase of phi(2) favored coordinated behavior. The topography of the phi complex is consistent with neuroanatomical sources within the human mirror neuron system. A plausible mechanism is that the phi complex reflects the influence of the other on a person's ongoing behavior, with phi(1) expressing the inhibition of the human mirror neuron system and phi(2) its enhancement.

Fig. 1.

The experimental setting and behavioral findings. Visual contact between subjects is manipulated by turning on (A) and off (B) a liquid crystal screen. (C) Distribution of trials for all eight pairs of subjects, arranged in increasing degree of coordination. An index (γ_cv) based on circular variance was used to assess the strength of coordination. Orange bars represent unsynchronized behavior (γ_cv < 0.03); green bars are from synchronized trials (γ_cv > 0.73). Dark and light green signify in-phase and anti-phase synchronization, respectively. Yellow bars represent transient behavior. The overall relative phase distributions for all trials corresponding to the vision-off and vision-on segments are shown in D and E, respectively. By using similar color coding as in C, further decompositions into synchronized (F), transient (G), and unsynchronized (H) trials are displayed.

Fig. 2.

Identification of spectral components in the brain activity of participants. (A) The dual-EEG of pairs was recorded with two caps each containing 60 channels. The head schematic of the subject on the right shows the 60 electrodes color-coded to reflect their spatial location. A similar identification was conducted on the second participant's EEG (data not shown). Circled areas indicate regions of peak rhythmic activity: mu (electrodes colored brown situated above Rolandic fissure); phi (burgundy above right centro-parietal area); and alpha (blue above the occipital pole). (B) Spectral plots used to identify mu, phi, and alpha components during visual contact. Because behavior in this example was completely unsynchronized, spectra show only phi₁, mu, and alpha, but no noticeable phi₂. (C) Topographical distributions of the mu and alpha rhythms before (Upper) and during (Lower) vision. The power of rolandic mu and posterior alpha was significantly depressed during visual contact, independently of the coordination achieved by each pair.

Fig. 3.

Time-frequency spectrum exhibiting vanishing bursts in the 10-Hz range during visual contact (from t = 20 to t = 40 s). The topographical maps at the top of the figure show the total energy in the range 9–11 Hz and capture both mu and alpha processes before and after vision and their desynchronization during vision. In this particular subject, no phi complex was detected. However, a wide spectral maximum at 11–14 Hz over occipital areas and an ample but narrow spectral maximum centered at 9.2 Hz over rolandic areas were observed.

Fig. 4.

Phi₁ and Phi₂ rhythms distinguish synchronized and unsynchronized (intrinsic) behavior. (A) Electrodes used to identify the phi complex are located 6 cm from the midline. (B) Plots of power differences between matching left and right electrodes of A. Because of their symmetry, most spectral components cancel out and only the asymmetrical components are retained. (C) Box-and-whiskers plot of power changes in Phi₁ showing selective increase in the right hemisphere and a corresponding decrease in the left hemisphere during unsynchronized behavior. (D) For Phi₂, power selectively increases in the right hemisphere only during synchronized behavior. Note the absence of overlap between the active phi distributions in the right hemisphere and their controls in the left hemisphere. (E) Representative examples of corresponding maps of power change showing that the topography of Phi₁ (unsynchronized behavior) and Phi₂ (synchronized behaviors) are similar. L, left hemisphere; R, right hemisphere.

Fig. 5.

Relation between Phi₂ and social coordination. (A) Time-frequency spectrum for electrode CP4 from a single trial. Phi₂ is low before and after vision but increases during vision. (B) Corresponding relative phase between finger movements. Synchronized in-phase behavior is observed most of the time during visual contact. The momentary loss of coordination around t = 31 s (highlighted by the arrow) is associated with the disappearance of Phi₂, seen by the gap from time t = 30 to t = 35 s in the time-frequency spectra.