Analysis of unexpected disruptive effects of contingent food reinforcement on automatically maintained <scp>self‐injury</scp>

Rooker et al. (2022) tested a simple idea: give kids lots of food for pressing a switch and see if their worst self-injury drops. Nine children with developmental delay joined. Each child had severe hand-biting or head-hitting that past treatments could not touch. The team ran short sessions in a quiet room. Press the switch, get a bite of favorite food. No other demands. No extinction. Just dense reinforcement.

Every child’s self-injury fell, often to zero, within the first few ten-minute sessions. The behavior stayed low as long as the food kept coming. When the team stopped the food, injury ticked back up. Bring food back, injury dropped again. The pattern showed the effect was real and tied to the reinforcement, not coincidence.

Johnston et al. (2017) warned that rich reinforcement can cause a big rebound later. Their lab study with adults showed the same density trade-off: stronger suppression now, sharper return when reinforcers stop. Rooker’s data match that warning; SIB came back when food ended.

Steege et al. (1989) did something similar decades earlier. They used microswitches to deliver food to severely handicapped children and also saw near-zero SIB for months. Rooker extends that old finding by showing the trick still works for today’s toughest cases.

Reid et al. (1999) paired mild punishment with leisure-item reinforcement. They needed reprimands plus toys to cut SIB. Rooker removed punishment entirely and still won, showing dense food alone can do the job when the behavior is automatically maintained.

If you face self-injury that never budges, try a high-rate food schedule before adding punishment or restraint. Pick a simple response the child can already do, deliver tiny bites every few seconds, and watch the data. Plan to thin the schedule slowly to avoid the rebound Smith flagged. This low-intrusion option could spare clients from more restrictive plans.