In working with larger numbers of speed/power athletes per session I’ve had continually positive results, regarding speed work, by initiating with only two cues ‘head down, and arm drive’. While the skill of sprinting is a multi-faceted animal, I’ve found those two factors alone to produce very positive peripheral mechanical influences elsewhere where they are needed.
While I think the optimal biomechanical model during early acceleration has the sprinter creating a relatively straight line from cervical spine to ankle in ground support, I’ve found that a degree of cervical flexion poses no negative influence and almost always contributes towards improved drive mechanics across the board.
Asafa is a great example as he typically doesn’t level his line of sight/chin (line his cervical spine with the thoracic) until he’s around the 40m mark.