What design patter to use? if any

Want to know your opinion on this use case, simplified, I have a task, with several subprocess (function calls, objects and such), these subprocess calculate deviations and parameters for the main subprocess (depicted as subprocess 4),

Currently each subprocess has a direct access to subprocess 4 but that seems not optimal, some subprocesses calculate multiple parameters so a simple return is not possible, and also most of them read from subprocess 4 (depicted as bidirectional arrow)

what could be a good design pattern for this mess?

here an image of the process for more clarity:

https://imgur.com/a/boi0cGE