There should be only three categories of review outcome.
Accept, Reject and Minor Revisions.
Part of the Editorial decision making will have to be whether the experiments demanded by the reviewers are reasonable as "minor" or not. I suggest a lean towards accepting only the most minimal demands for additional experimentation as "minor revisions" and otherwise to choose to reject.
And no more of this back and forth with Editors about what additional work might make it acceptable for the journal as a new submission either.
We are handing over too much power to direct and control the science to other people. It rightfully belongs within your lab and within your circle of key peers.
If J Neuro could take a stand against Supplemental Materials, they and other journals can take a stand on this.
I estimate that the greatest advantage will be the sharp decline in reviewers demanding extra work just because they can.
The second advantage will be with Editors themselves having to select from what is submitted to them, instead of trying to create new papers by holding acceptances at bay until the authors throw down another year of person-work.