Neuroscientist and PNAS study co-author David Levinthal is also a board-certified internist who specializes in gastroenterology. We asked Levinthal to comment on the investigation, which identified the neural networks that connect the cerebral cortex to the adrenal medulla, which is responsible for the body’s rapid response in stressful situations. (The adrenal medulla is the inner part of the adrenal gland.)

What did you find most surprising in the course of this study?

DL: When we initially set out to map the cerebral cortical representation of the adrenal gland, our lab already had preliminary evidence (from rats) showing that motor cortical regions play a role in the control of sympathetic nervous system. That preliminary data itself was a surprise, as it suggested that internal organ representation within the cortex was at least partially embedded within a classically regarded "somatotopic" map of muscle control, falling within the territory corresponding to trunk and leg muscles. Yet, the prevailing view about the sources of cerebral cortical influences on stress responses in nonhuman primates and humans was that the brain's influence on autonomic control arose predominantly from non-motor areas of the cortex, including the cingulate cortex and the insula. As there could have been important differences between rats and primates regarding cortical regulation of the sympathetic nervous system, we still expected to find that the dominant representation of the adrenal gland was in the cingulate cortex and the insula in monkeys, with perhaps some representation in the motor areas.

The biggest surprise about the findings we reported was that the dominant representation of the adrenal gland was found not in the cingulate or insula, but within the motor cortical areas on the hemisphere opposite to the adrenal gland that we injected with rabies virus. In this regard, the cortical representation of the adrenal gland "looks like a muscle." It was also surprising to find that adrenal representation was located not only within a few cortical motor areas (primates have six premotor areas and a primary motor cortex), but within all of them, and moreover within regions of these motor areas best corresponding to axial or trunk muscle representation. All of these areas of cortex send projections to the spinal cord (via the "corticospinal tract"). The findings call for a re-evaluation of the function of these motor cortical areas and the corticospinal tract, which was believed to convey only motor commands. Our evidence would suggest that both motor and autonomic commands are channeled from the brain to the body using this pathway. It also suggests that movements, particularly those using core muscles, can have a direct impact on internal organ function and stress responses. This aspect of our anatomic findings could provide an explanation for the health and mental impacts of yoga, Pilates, or tai chi.

Were there difficulties that you had to overcome in this study?

DL: Perhaps the greatest difficulty was the sheer amount of time, effort, and patience needed to do the detailed anatomical analyses necessary to make our surprising claim. I think about the quote often attributed to Carl Sagan -- "extraordinary claims require extraordinary evidence." Furthermore, great science must be reproducible. Thus, we needed to be really sure the data about the adrenal representation was correct. Over a period of a little more than six years in which we conducted multiple independent studies, spending hundreds of hours at the microscope to plot the location of hundreds of thousands of rabies-infected cortical neurons, and probably a hundred hours or more creating unfolded cortical maps from those plots, we felt confident about the findings because we obtained the same basic result each time. Richard Dum spent by far the most time devoted directly to this effort, and it is hard to articulate how patient and determined Richard had to be to see this work through. Great science required that effort and attention to detail!