Both, to be honest. There still needs to be a continuity of treatment throughout the CJ system, from incarceration through re-entry.
These scans, if accurate, will simply be another tool in the toolbox. Various assessments (old-school interview style) regarding decision-making skills are already being used. This would add another measurement tool to the mix.
Decision-making skills play a significant role, but there are plenty of other factors that help to reduce recidivism rates, such as anti-social belief systems, mental health, criminal companions, etc.
It's important people realize this only measures likelihood of re-offending. It's not a yes/no. No "pre-crime" scenarios could be used. Even if someone measured out at 99% likelihood to re-offend, it doesn't guarantee they will.