Neural and behavioural manifold dynamics align across interacting individuals

The coordination of actions with others is a crucial aspect of social behavior, demanding that the nervous system constantly adjust motor output to match a partner's evolving movements. However, the underlying neural population principles that enable such coordination are still largely unknown. To investigate this, researchers conducted two dual-EEG studies involving 44 dyads, comprising 88 participants, who either synchronized their finger movements instructedly or engaged in spontaneous face-to-face interactions.

By employing kinematics-informed deep contrastive learning combined with dynamical-systems modeling, the scientists identified low-dimensional neural manifolds. These manifolds exhibited both geometric and temporal alignment across interacting partners, mirroring their coordinated behavior and revealing interpersonal alignment that traditional synchrony measures had missed.

Crucially, these manifolds displayed a flexible attractor-like organization, consistent with a synergistic, dynamical approach to motor control. Notably, the attractor properties of these neural manifolds were co-regulated across partners. The researchers propose that interpersonal coordination emerges through the intermittent updating of internally organized dynamics by observing a partner's movements.

More broadly, the study demonstrates that movement-informed latent-space modeling can uncover low-dimensional population dynamics linking neural activity, neuromuscular control, and interpersonal coordination.