I shall skip the dyad aspect because I must have misunderstood the original context of your post (read it as commented by a colleague).

‘longitudinal’ is typically used by developmental scientists (psychologists, neuroscientists etc) and denotes often interest is change and underlying processes. Since human development is multicausal and shows equifinality its not easy to reduce to simple causal effects. We like complexity!

Dunno about whole communities but we can definitely see the ‘longitudinal dyad’ as showing trajectories of behaviour different from either interaction partner (infant-caregiver dyad). We analysed these separate trajectories here- perhaps scalable to bigger groups? doi.org/10.1111/desc...

Full ref:
Duda-Goławska, J., Laudańska, Z., et al. Vocalisations are coupled with movement of all limbs throughout infancy. Sci Rep 15, 44899 (2025). doi.org/10.1038/s415...

8/8 Implications: strong cross-domain motor-vocal links in vocal production during naturalistic social interactions;
New targets for studying motor constraints on early speech dev, esp in NDDs where motor problems are common and varied.

7/ finally, our results could also reflect 3) insufficient differentiation of motor actions in infancy when they are trying to produce responses to caregiver’s speech.
Note converging findings of @jeremyborjon.com from older infants/toddlers: increased temporal alignment of head mov with vocs

6/ so far we so not know the functional role of these movements: 1) a “spillage” from general mote activation (e.g. motor overflow described by Soska etal or our team Koziol et al 2023). 2) Conversely they could support alignment of breathing with vocal production (adult work by Pouw &Fuchs)

5/ the arms move systematically around the onset across the entire studied period (4-12 mo) while leg movements are also present but decrease significantly at 9 and 12 mo media.springernature.com/full/springe...

4/ speech-like vocalisations (~ 2000 per timepoint, ~25-35 per infant) were all manually annotated. We measured acceleration of arms and legs around the onset of each vocalisation. We found systematic brief increases of acc just before the onset followed by decelerations within 1 second.

3/ we wanted to capture spontaneous vocal activity during typical dyadic infant-caregiver interactions. Limb movements were measured in real time with wearable accelerometers. The study was longitudinal so we have repeated measures at 4, 6, 9 and 12 months of age.

2/ Speech production is embodied but what does it actually mean early in development? We investigated whether individual instances of speech-like vocalisations (as opposed to cries or laughs) also involve brief arm and leg movements.