The microstructure of intra- and interpersonal coordination

Abstract

Movements are naturally composed of submovements, i.e. recurrent speed pulses (2-3 Hz), possibly reflecting intermittent feedback-based motor adjustments. In visuomotor (unimanual) synchronization tasks, partners alternate submovements over time, indicating mutual coregulation. However, it is unclear whether submovement coordination is organized differently between and within individuals. Indeed, different types of information may be variably exploited for intrapersonal and interpersonal coordination. Participants performed a series of bimanual tasks alone or in pairs, with or without visual feedback (solo task only). We analysed the relative timing of submovements between their own hands or between their own hands and those of their partner. Distinct coordinative structures emerged at the submovement level depending on the relevance of visual feedback. Specifically, the relative timing of submovements (between partners/effectors) shifts from alternation to simultaneity and a mixture of both when coordination is achieved using vision (interpersonal), proprioception/efference-copy only (intrapersonal, without vision) or all information sources (intrapersonal, with vision), respectively. These results suggest that submovement coordination represents a behavioural proxy for the adaptive weighting of different sources of information within action-perception loops. In sum, the microstructure of movement reveals common principles governing the dynamics of sensorimotor control to achieve both intra- and interpersonal coordination.

Keywords: interpersonal coordination; intrapersonal coordination; motor control; movement intermittency; sensorimotor loops; submovements.

Figure 1.

Experimental setup and conditions. (a) Dyadic task. Pairs of participants were seated at a table facing each other, with a panel placed in the middle to prevent them from seeing each other's faces. They were instructed to keep the ulnar side of their left and right forearms on the table and clench their hands into fists, pointing only their left and right index fingers toward their partner's corresponding index fingers (upper left panel). (b) Participants were asked to perform slow (0.25 Hz) rhythmic flexion-extension movements of both their index fingers together, as much as possible synchronized in-phase or anti-phase with their partner (upper right panel). (c) Solo task. Single participants were asked to sit at a table in a posture similar to that described for the dyadic task, but with their left and right index fingers pointing toward each other (lower left panel). (d) Participants were asked to perform rhythmic movements of flexion-extension, synchronizing their fingers as much as possible either in-phase or anti-phase, while keeping (in separate blocks) their eyes open (vision) or closed (no-vision; lower right panel).

Figure 2.

Schematic of submovement-level analysis. (a) Low-pass filtered velocity traces (cut-off: 4 Hz, two-pass, third-order Butterworth) from one example participant's right hand (in black) and the respective partner's left hand (in grey) during the dyadic task, in-phase condition. The green area highlights data segments corresponding to an individual movement of both hands (see Data segmentation). (b) Schematic of the submovement-locked analysis pipeline. The average velocity profile time-locked to movement onset – obtained by averaging (trial-wise) all movements' kinematic profiles – is subtracted (point-by-point) from each individual movement segment (belonging to the same trial). Submovements are then identified as local peaks in the right hand data (triangles). The velocity data for both the right and left hands are then segmented by taking time windows centred around the submovements identified in the right hand and ranging from −0.6 to +0.6 s. The red area highlights an example time window centred around a submovement (i-th) detected in the right hand (red triangle). Data segments from the right and left hands time-locked to the i-th submovement in the right hand are highlighted as thick black and grey lines, respectively. This procedure is repeated for all detected submovements and for all movements. Finally, to quantify submovement-level coordination, all submovement-locked data segments are averaged (separately for each hand). For the right hand, the result of this analysis corresponds to the average velocity profile of submovements (expected result is shown by the black line in the lower panel); for the left hand, the result indicates velocity as a function of time from submovements produced by the right hand (results change as a function of the experimental condition).

Figure 3.

Movement- and submovement-level kinematic features. (a,b) Mean and variability (s.d.) of movement pace for the dyadic and solo task calculated as the time interval between the onset of successive movements. The dashed lines indicate the normative time interval between individual movements, i.e. 2 s for an instructed pace of 0.25 Hz. Dyadic task. In-phase: mean = 2.02 ± 0.31, s.d. = 0.6 ± 0.22; anti-phase: mean = 2.12 ± 0.31, s.d. = 0.68 ± 0.20 (mean ± s.d. seconds). Solo task. In-phase/vision: mean = 2.35 ± 0.45, s.d. = 0.52 ± 0.16; in-phase/no-vision: mean = 2.28 ± 0.24, s.d. = 0.46 ± 0.16; anti-phase/vision: mean = 2.22 ± 0.31 s.d. = 0.46 ± 0.18; anti-phase/no-vision: mean = 2.29 ± 0.29, s.d. = 0.48 ± 0.15 (mean ± s.d. seconds). (c,d) Mean and variability (s.d.) of movement coordination performance for the dyadic and solo task calculated as the absolute difference in movement onset time between the two partners’ (spatially corresponding) right and left fingers (dyadic task) and between one's own right and left fingers (solo task). Dyadic task. In-phase: mean = 142.18 ± 29.31, s.d. = 107.79 ± 24.42; anti-phase: mean = 174.47 ± 23.90, s.d. = 129.90 ± 19.44 (mean ± s.d. milliseconds). Solo task. In-phase/vision: mean = 48.88 ± 17.47, s.d. = 46.34 ± 20.47; in-phase/no-vision: mean = 67.31 ± 33.95, s.d. = 67.74 ± 38.49; anti-phase/vision: mean = 58.48 ± 22.84, s.d. = 56.71 ± 25.47; anti-phase/no-vision: mean = 79.07 ± 30.86, s.d. = 76.35 ± 32.19 (mean ± s.d. milliseconds). (e,f) Mean and variability (s.d.) of the inter-submovement time for the dyadic and solo task, i.e. the time interval between successive submovements. Dyadic task. In-phase: mean = 345.76 ± 18.37, s.d. = 123.78 ± 10.91; anti-phase: mean = 353.95 ± 17.42, s.d. = 132.57 ± 9.44 (mean ± s.d. milliseconds). Solo task. In-phase/vision: mean = 358.10 ± 21.28, s.d. = 129.10 ± 15.73; in-phase/no-vision: mean = 361.45 ± 16.90, s.d. = 138.85 ± 15.97; anti-phase/vision: mean = 365.16 ± 19.05, s.d. = 139.06 ± 14.68; anti-phase/no-vision: mean = 372.12 ± 23.60, s.d. = 145.65 ± 17.65 (mean ± s.d. milliseconds). All reported p-values are obtained from paired-sample t-tests between in-phase and anti-phase coordination modes (dyadic task) and from two-way ANOVAs for repeated measures with coordination mode (in-/anti-phase) and feedback (vision, no-vision) as within-subject factors (solo task). *p ≤ 0.05; **p ≤ 0.01; ***p ≤ 0.001. The boxplots depict the interquartile range (IQR) and the median value (horizontal line inside the box), while the whiskers indicate the range (min-max).

Figure 4.

Submovement coordination between partners and between effectors. (a) Upper panels. Velocity of the left-hand finger time-locked to submovements generated by the respective partner's right-hand finger (blue: in-phase, red: anti-phase). The black lines show the average velocity profile of submovements (i.e. submovement-locked averages, right hand data only; mean ± s.e.m.). Lower panels. Submovement probability (expressed as percentage deviation from mean probability) for the left-hand finger as a function of time relative to submovements generated by the respective partner's right-hand finger (mean ± s.e.m.). The black line and grey shaded area indicate the mean ± 3*s.d. for the surrogate data. Thick horizontal black lines indicate the time points that survived the nonparametric two-tailed statistical test against surrogate data (Bonferroni-corrected for multiple comparisons across time). (b) Upper panels. Velocity of the participants' left-hand finger time-locked to submovements generated by their own right-hand finger (blue: in-phase; red: anti-phase). The black lines show the average velocity profile of submovements (i.e. submovement-locked averages, right hand data only; mean ± s.e.m.). Lower panels. Submovement probability (expressed as percentage deviation from mean probability) for participants’ left finger as a function of time relative to submovements generated by their right finger (mean ± s.e.m.). The black line and grey shaded area indicate the mean ± 3* s.d. for the surrogate data. Thick horizontal black lines indicate the time points that survived the nonparametric two-tailed statistical test against surrogate data (Bonferroni-corrected for multiple comparisons across time). The coloured boxes drawn on the hand pictures highlight the relevant relationship that is being considered for the analysis of submovement coordination, i.e. whether the relationship between the submovements produced by the hands of the two interacting partners (interpersonal) or by the two hands of the same participant (intrapersonal) is being analysed.

Figure 5.

Submovement coordination with and without visual feedback. (a) Upper panels. Velocity of the participants' left-hand finger time-locked to submovements generated by their right-hand finger (blue: in-phase; red: anti-phase) when visual feedback is available (vision). The black lines show the average velocity profile of submovements (i.e. submovement-locked averages, right hand data only; mean ± s.e.m.). Lower panels. Submovement probability (expressed as percentage deviation from mean probability) for participants’ left finger as a function of time relative to submovements generated by their right finger when visual feedback is available (mean ± s.e.m.). The black line and grey shaded area indicate the mean ± 3* s.d. for the surrogate data. Thick horizontal black lines indicate the time points that survived the nonparametric two-tailed statistical test against surrogate data (Bonferroni-corrected for multiple comparisons across time). (b) Upper and lower panels. The same as shown in (a) but in the absence of visual feedback (no-vision; cyan: in-phase, magenta: anti-phase). Note that the sample size varies across the experimental conditions: the in-phase/vision condition was performed by 33 subjects, while 13 subjects (out of 33) also completed all other conditions.