Maternal nutrition and four-alternative choice.
Fetal under-nutrition does not slow rats’ later choice shifts across four concurrent VI schedules, and a contingency-discriminability model outperforms generalized matching.
01Research in Context
What this study did
Michael and colleagues raised two groups of rat moms. One group ate normal chow. The other group got half-rations during pregnancy. After birth, all pups ate the same diet. The team then tested the grown pups on a four-lever box. Each lever paid off on its own variable-interval schedule. The rats could switch any time. The scientists slowly changed the pay rates. They watched how fast the rats followed the new odds.
The goal was to see if prenatal hunger altered how quickly the animals re-allocated their choices.
What they found
Both rat groups adjusted their lever pressing at the same speed. Under-nutrition in the womb did not slow learning. A contingency-discriminability model fit the data better than the classic generalized matching equation. The rats still followed the pay rates, but the newer model captured tiny biases the old one missed.
How this fits with other research
Bron et al. (2003) ran a near-copy study with possums. They also compared the two models. Their data fit both, but they kept the simpler generalized matching law for clarity. Michael et al. (2007) show that when you add a fourth option and a nutrition twist, the contingency-discriminability model wins. The newer tool is worth the extra parameters.
Katz et al. (2003) showed pigeons can track payoff ratios that flip every five seconds. Michael’s rats saw slower, session-long shifts yet still re-allocated. Together, the papers set bounds: behavior can follow both lightning-fast and gradual schedule changes.
Spanoudis et al. (2011) found pigeons over-correct when contingencies first change, then settle. Michael’s rats showed no over-shoot; they slid straight to the new ratio. Species, task, or signal clarity may explain the difference.
Why it matters
If you run concurrent-schedule interventions, consider the contingency-discriminability model when you see stubborn bias. It may describe your client’s choices better than pure matching. The rat work also warns us that prenatal environment did not dent flexibility; postnatal learning can still proceed at full speed. Keep teaching new contingencies even when biological risk factors are present.
Want CEUs on This Topic?
The ABA Clubhouse has 60+ free CEUs — live every Wednesday. Ethics, supervision & clinical topics.
Join Free →Plot your client’s response ratios against reinforcement ratios; if simple matching under-predicts, try fitting the contingency-discriminability model instead.
02At a glance
03Original abstract
Two groups of 10 male rats were trained to nose poke for food pellets at four alternatives that provided differing rates of pellet delivery on aperiodic schedules. After a fixed number of pellets had been delivered, 5, 10 or 20 in different conditions of the experiment, a 10-s blackout occurred, and the locations of the differing rates of pellet delivery were randomized for the next component. Two groups of rats were used: The AD group consisted of 10 rats born to dams that had normal (ad libitum) nutrition during pregnancy, whereas the 10 rats in the UN group were from dams exposed to reduced food availability during pregnancy. All pups received normal nutrition after birth. Choice between the nose-poke alternatives quickly adapted when the rates of pellet delivery were changed in both groups, but there were no consistent differences in the speed of adaptation between the two groups. The generalized matching relation failed to describe the allocation of responses among alternatives, but the contingency-discriminability model provided a precise description of performance.
Journal of the experimental analysis of behavior, 2007 · doi:10.1901/jeab.2007.12-06