Behavioral treatment of automatically reinforced SIB: 1982 – 2015

Rooker et al. (2018) read every paper from 1982 to 2015 that used behavioral treatment for self-injury maintained by automatic reinforcement. They looked at 33 years of data to see which tactics worked best.

The team focused on kids and adults with intellectual or developmental disabilities whose SIB happened for internal sensory pay-offs, not attention or escape.

The clear winner was noncontingent reinforcement built from competing-stimulus assessments. When therapists first tested items that matched the sensory channel, then delivered those items on a schedule, SIB dropped more than when they used items from a simple preference assessment.

Preference-assessment-based NCR still helped, but the effect was smaller and less steady.

Verriden et al. (2025) ran the exact experiment the review called for. They compared CSA-selected items against preference-assessment items inside the same kids. CSA items cut challenging behavior more for three of four children, backing up the 2018 claim.

Phillips et al. (2017) seems to disagree at first glance. Their case series showed that automatic-reinforcement cases often needed extra procedures beyond NCR. The difference is scope: Phillips looked at real-world clinical cases where staff added extinction or DRO when NCR alone stalled, while Rooker averaged across all published studies where extra pieces were already blended in.

Lancioni et al. (2009) had already labeled NCR plus extinction as 'well-established.' Rooker refines that label by showing the assessment step that makes NCR work best for automatic SIB.

Stop skipping the CSA. Before you write an NCR plan for hand-biting or head-hitting, spend 10 minutes testing sensory matched items—vibration, music, textured toys—then use the top one on a fixed-time schedule. You will likely see faster, larger drops in SIB than if you grab the child's 'preferred items' list. One extra assessment, less injury, fewer sessions wasted on weak items.