Human Brain and Behavior Laboratory Center for Complex Systems & Brain Sciences

1. Human Brain and Behavior Laboratory Center for Complex Systems & Brain Sciences Florida Atlantic University

2. Social coordination: a group of people perceives the behavior of one another (informational coupling) AND adjust their behaviors (coordination) AT THE SAME TIME Behavior resulting from a person’s individual tendencies: intrinsic behavior Psychological denomination: intentional action Neural system: intrinsic motor preparation system Effect of perceiving another’s behavior: imitation tendency Psychological denomination: social embodiment (Niedenthal et al., 2005) Neural system: mirror neuron system (MNS, Rizzolatti & Craighero 2004)

3. Both intrinsic premotor system and MNS are functionally connected to the motor cortex (Baldissera et al., 2001; Lau et al., 2004) -> what system regulates their respective influence on the behavior? How are conflicts resolved? Here we study social coordination in a task requiring a pair of subjects to watch each other while performing self-paced rhythmic movements. We seek neural signatures associated with coordinated action and with independent behaviors.

4. SUBJECTS: Sixteen subjects (10 males, 6 females, aged between 22 and 41 years, (mean 29 years). Constituted 8 pairs: 4 gender-mixed; 3 male-male; 1 female-female. All right-handed on the basis of self-report. Had normal or corrected-to-normal vision and reported no history of neurological disease. TASK: perform regular continuous finger movements at a comfortable pace during one minute trials. Subjects are instructed to adopt the most comfortable pace, at any time during the trial t=0-20s t=20-40s t=40-60s DUAL-EEG RECORDING: Dual-EEG recorded using two 60-channel EEG caps with Ag-AgCl electrodes (Falk Minow Services, Germany) arranged according to the 10 percent system (midline and rows 1 to 8), with 2 distinct referential montages. Signals directed to a single amplifier (Synamp2, Neuroscan, Texas) analog filtered (Butterworth, bandpass from 0.05 Hz ( -12 dB/octave) to 200 Hz (-24 dB/octave), amplified (gain= 2010) and digitized at 1000 Hz with a 24 bits ADC in the range ±950 microV (vertical resolution of 0.11nV). Recording performed with the respective grounds located at FPz sites and the references at the corresponding linked mastoids. Impedances maintained below 10 kOhms. TRIALS: 36 trials lasting 1 minute. Vision of the other controlled through fast-switching (1.2ms) LCD screen, turning transparent at t=20s and back to opaque at t=40s

5. IDENTIFY COORDINATED AND UNCOORDINATED BEHAVIORS: In contrast with action observation paradigms, our social coordination paradigm provides the opportunity to obtain an explicit behavioral criterion of brain changes elicited by the perception of another’s behavior. In contrast with imitation studies, social coordination allows one to quantify behavioral biases in social interaction, and to do this in real time. For rhythmical movements, coordination is identifiable with minimal ambiguity: it is manifest when the collective variable (relative phase between the two oscillatory behaviors) becomes stable (Kelso 1995). Trials were classified by 3 judges using frequency and relative phase information. Trials with complete uncoordination, partial locking and complete coordination were observed. The absence of overlap between the distributions of synchronization index for coordinated and uncoordinated trials was verified numerically using a synchronization index gcv

6. SEGMENT THE EEG FOR EACH SUBJECT: EEG spectra were investigated over windows of 16.5 seconds (for each 20-second periods: transients were discarded at the onset (3s) and at the end (0.5s) due to non stationarity). Resulting spectral resolution: 0.06 Hz. Single trials were tapered with a Tukey window (10%), and Discrete Fourier Transforms (DFT) were used to estimate amplitude spectra. For display purposes, spectra were smoothed with a 5-point Bartlett filter. Spectra from different channels were color-coded in a continuous colorimetric space (Cie L*a*b), to identify spatial spectral patterns.

7. Spectral peaks isolated for each subject, frequency, amplitude and boundaries extracted. Three rhythms were consistently identified in the 7.5-13H: rolandic mu, occipital alpha, and a new (right) centroparietal component we called phi.

9. PHI COMPLEX: Plots of power difference were used to identify phi. We computed the difference in the spectral amplitudes between the interhemispheric pairs of electrodes in rows 3 and 4 of the 10%-montage (e.g., difference C3-C4 over the central sulcus). Symmetrical components cancel out and only asymmetrical components (including phi) are retained. In most subjects, a dissociation between two phi subcomponents appeared. The boundaries of the components of the phi complex were then extracted and their changes in power, (ERD/ERS: event-related desynchronization and synchronization) were calculated for trials showing coordinated and uncoordinated behavior respectively. Active phi components (components increasing their amplitude during visual contact in 'coordinated' or 'uncoordinated' behaviors) were identified and their properties examined further.

11. Phi1 and Phi2 rhythms distinguish coordinated and uncoordinated behaviors 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. 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. Representative examples of corresponding maps of power change showing that the topography of Phi1 (unsynchronized behavior) and Phi2 (synchronized behaviors) are similar. Abbreviations: L: left hemisphere. R: right hemisphere.

12. Time-frequency spectrum for electrode CP4 from a single trial in a subject showing little alpha and mu activity. The bursts correspond to a prominent phi2. Phi2 is low before and after vision, but increases during vision. Bottom: corresponding relative phase between finger movements of this subject and his partner. Synchronized inphase 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 sec in the time-frequency spectra.

13. Apparent lateralization: simulated data - f and m power aggregated This simulation assumes an experimental manipulation in which mu is suppressed (during perception of the other's behavior) and phi is not affected. It shows that an apparent lateralization of the ERD can result when the components are not separated. (a) During movement observation (right bars), the mu power in each electrode (C3 above the left motor cortex, C4 above the right motor cortex) is suppressed as compared to baseline (left bars). This suppression is bilateral. (b) During movement observation, the power in the phi band is higher in the right hemisphere than in the left, but movement observation does not differ from baseline. (c) When mu and phi are combined, the initial hemispheric difference between the left and right phi contributes more to the remaining power and the initial right asymmetry is amplified. This shows that apparent hemispheric lateralization of the ERD is possible without any source exhibiting differential changes across hemisphere.

15. DISCUSSION: For effective social coordination to occur, at least one of the participants has to be affected by vision of what the other is doing. In neural terms, the mirror-neuron system must effectively influence the motor cortex of at least one participant. In contrast, when no phase-locking tendency is observed individual behaviors ('intrinsic dynamics') predominate, potentially by enhancing activity in the premotor system or by inhibiting the mirror-neuron system. Our results suggest that enhanced functional connectivity between the intrinsic premotor system and the motor cortex is reflected by phi1, whereas enhanced functional connectivity between the mirror-neuron system and motor cortex is reflected by phi2. CONCLUSION: Our data suggest that the mirror-neuron system effects appropriate behavioral changes by recruiting an oscillatory complex that is spatially and spectrally distinct from rolandic mu. Mu and phi seem to play distinctly different roles. Whereas mu contributes to somatosensory awareness when the acting partner is perceived, the phi complex plays the role of a gating mechanism selectively parsing social from individual behavior.

16. References: Niedenthal et al., (2005). Personality and Social Psychology Review, 9(3): 184-211. Rizzolatti & Craighero (2004). Annual Review of Neurosciences, 27: 169-192. Lau et al., (2004). Science, 303: 1208-1210. Baldissera et al., (2001). European Journal of Neurosciences, 13:190-194. Kelso (1995).MIT press, Cambridge, MA.