Subject mice were standard Black-6 lab mice ordered from Charles River Laboratories. Prior to the experiment period they had been trained to operate a food dispenser and allowed to reach a BMI of 20% above the accepted healthy level.
At the start of the experiment the mice were given four weeks to acclimate to their habitats. Each habitat had two rooms. The "home room" was kept at an ideal temperature, exposed to natural lighting patterns, supplied with adequate nesting material, and contained a "bottomless" water dispenser. It also contained Food Dispenser "A", calibrated to release a pelleted ration of 5 grams of food each day when prompted by the rats (previously trained to operate the dispenser). The home room also contained a freely available exercise wheel equipped to monitor use. If the mice did not voluntarily produce 1000 wheel rotations each day, they were placed on the wheel and not allowed to exit until they had completed the daily regimen.
Each habitat contained a second room, "the fridge", accessible through a short tube. The temperature of the fridge was kept at 40 degrees F and it was lit 24 hours a day at extreme levels. The fridge contained Food Dispenser "B". Unlike the dispenser in the home room, Dispenser B would supply an unlimited amount of food in response to prompting. The fridge did not contain a water dispenser.
At the end of the 4 week acclimation period, the tube connecting the two rooms was electrified. A rat passing through the tube at a normal rate of speed received a shock calibrated to be 5% less powerful than an amount that would cause lack of muscle control. At the end of week 1, Food Dispenser A was reset to dispense 4.9 grams of pelleted rations. This ration reduction continued at a rate of 1/10th of a gram per week for 40 weeks. Thus, for the final week of the experiment the food ration was 1 gram per day.
The following observations were made:
During the initial 4 weeks the number of mice that accessed the fridge room and obtained "extra" food was 70%.
When the shocks were initiated, that number dropped to 5%.
After 3 weeks, when food rations in the home room were 4.7 grams (94% of the original home room diet), 50% of the mice chose to endure at least 2 shocks (one each way) daily to obtain "extra" food.
After 20 weeks, when rations in the home room had dropped to 3 grams, 80% of the mice chose to endure at least 2 daily shocks. Another 10% of the mice had begun to reside in the fridge, thereby avoiding the daily shocks. These mice all died of dehydration by week 21.
After 30 weeks the experiment was stopped. All but 5% of the mice had relocated to the fridge, avoided the daily shocks, and died of dehydration.
Of the surviving 5%, 90% had never visited the fridge after being shocked 10 or fewer times. The remaining 10% (who were observed to be suffering the greatest from the reduction of calories) began to visit the fridge only in a cycle regulated by their BMI. Typically choosing to obtain extra food when their BMI dropped below 95% of normal, and continuing to do so until they reached a BMI of 105% normal. As the Dispenser A produced fewer calories each week the cycle accelerated.
Looking at the survivors only, 95% of the subjects lost weight on a regimen of exercise and calorie restrictions, including some who "cheated" on their diet. Simply stated: diet and exercise works. The authors therefore wholeheartedly endorse the continued recommendation that obese individuals should improve their health through diet and exercise and should avoid medication and other drastic interventions whenever possible.