Towards a consensus on developmental regression