Shufflebrain

I repeated and confirmed these observations: cerebral hemispheres regenerated in about six weeks, and I found no restoration of lower parts of the brain, even after several months. Six weeks was plenty of leeway for shufflebrain, for I knew from student days that two pieces of brain tissue could easily weave a functional bridge within three weeks. My results would be harvested well before regeneration became a factor. Still, I wanted a more positive way of eliminating regeneration, just in case long-term experiments became necessary. Now a grafted organ can prevent regeneration of a lost part. Reattach an amputated limb, for instance, and, if it heals properly, no new parts regenerate. I had performed many such experiments. Perhaps I could prevent regeneration if I grafted the spinal cord in place of the amputated cerebrum? I experimented. Indeed, the spinal-cord segment quickly attached to the stump, and no new cerebrum developed. In fact, I found that any portion of the central nervous system would quickly weave fibers into the stump and substitute for and prevent regeneration of any other part of the brain. The transplants retained their original anatomy.

My first formal plunge into shufflebrain experiments involved a repetition of Lashley's basic operation. I mapped the brain into regions. Then, in an extensive series of operations, I replaced given map regions with pieces of spinal cord, systematically moving down the brain, region by region, among the various subgroups of experimental animals.

The animals invariably fed the moment they recovered from postoperative stupor, no matter which region I removed. Massive destruction of the brain reduced feeding but did not stop it. Clearly, there was no exclusive repository of feeding programs in any single part of the salamander's brain.

My next series of experiments involved interchanging right and left hemispheres. Feeding survived. Next, I tried rotating the cerebral hemispheres 180 degrees around their long axes so that up faced down and down faced up. I did this with each hemisphere separately and then with both simultaneously. Feeding survived. Next I turned the cerebral hemispheres around so that the front faced to the rear. Feeding survived.

I then moved down to the diencephalon, where the optic nerves enter the brain. I knew from my own observations and from the vast literature on the subject that disturbing the diencephalon might affect vision. But I was sure that touch and sonar from the lateral line system would enable the animal to sense the worm. My first experiments in this series involved removing the diencephalon and fusing the cerebral hemispheres directly onto the midbrain. As I had expected, surgery blinded my animals. They were also much less active than before. But when a worm came close, the salamander would attack and eat it, diencephalon or no!

In the next series of experiments with the diencephalon, I either rotated the part 180 degrees or reversed it front to back. These animals did recover vision. But often, when they struck (or seemed to strike) at a worm, they'd move in the wrong direction. At least this was how I interpreted their strange behavior toward worms, on the basis of eye rotation experiments performed by Roger Sperry in the 1940s. However, when I dimmed the lights and forced the salamanders to use sonar and touch, instead of vision, they were able to attack the worms. (In other words it wasn't the whack in the head--diencephalon to be more precise--that had induced the weird visual response.)

My next experiments involved the exchange of brain parts. For example, I switched the diencephalon with the cerebrum. I moved the midbrain up front and either the cerebrum or the diencephalon to the rear. I performed every operation I could think up. But nothing eradicated feeding. Let me describe one series of rather drastic operations in more detail.

I decided to see what effect an extra medulla would have on feeding. The best approach, I thought, would be to remove the entire brain of one animal, down to and including its medulla, and fuse it onto the medulla of another animal--the host. The available space inside my prospective host's cranium, however, wasn't sufficient to accommodate this amount of brain--not if I used a sibling as a donor (which I most frequently did). Therefore, I decided to use a large tiger salamander larva (Amblystoma tigrinum) as the host and a little marble salamander larva (Amblystoma opacum) as the donor. The fit was perfect, and the operations went very well. A week to ten days later, all five of the animals in the group were awake and feeding again. And no facet of their behavior even hinted at the bizarre contents of their crania.