Letter to the Editor
Scientific American
Re: News Scan article "On the Other Hand" June 2009
From: Charles Matthews M.D.
Director, the North Carolina Comprehensive Headache Clinic
Raleigh N.C.
Dear Sir:
In the June 2009 edition there is a short report on the interesting observation that two patients who underwent double-hand transplants switched handedness. In both cases, researchers noted that both patients, who had lost both hands three to four years before transplantation, were initially right handed (and most likely dominant left cortex).
Crossings in nervous systems are ubiquitous in vertebrates, and yet their function is entirely unexplained. Any collection of neurons which makes contact with another collection of neurons undergoes a crossing (or "decussation") of the majority of their fibers; examples include the motor system, which crosses at the medullary decussation; the somatosensory system, which crosses at the thalamus; the cerebellum, which crosses twice, once in connecting to brainstem nuclei and again when brainstem nuclei connect to cortex; and in the visual system, where retinal neurons cross at the optic chiasm. In addition to these right-to-left crossings, such fiber tracks maintain a spatial organization which is associated with an inverted representation in the cortex. For example, the famous "homunculi" of both the motor and sensory cortical strips, which run vertically, are organized such that they are "upside down", with the head at the bottom of the strip, and the feet at the top. So, connections developing in the nervous system between two populations of neurons, in which spatial relationships are preserved, invert right to left, and top to bottom.
Does switching handedness after bilateral hand transplant have any relationship to these other inversions and crossings that occur ubiquitously in vertebrates? I believe they do, and the reason may be because fiber tracks making connections between two populations of neurons with preserved spatial organization undergo a form of pupillary constriction of the fibers which produces an inversion of the spatial information. In short, the pupillary functions of the eye are but one example of "pupil-like" information transformations throughout the nervous system.To explain how this might work, let us review how the pupil works in the eye.
The traditional picture of how the pupil and retina work is to draw a picture of an object, and trace light rays scattered off this object through the pupil, which projects an inverted image on the retina.
With the pupil maximally dilated, scattering of light rays are not restricted.
Constriction of the pupil maps objects in the visual field to an inverted representation on the retina. As the pupil constricts from the maximally dilated position, spatial relations become more distinct, as scattering is progressively reduced.
The "pupil effect" of inversion and amplification of spatial information at the expense of loss of more complete information from scattering is not intrinsically related to light rays as such, but rather arises between any set A of (spatially related) points are mapped onto set B of (spatially related) points and then subsequently subjected to a pupil-like restriction. The same inversion effect would be expected at the attachment site of a hand, in which neural connections are first made widely (analogous to a fully dilated pupil), and then are subsequently restricted, bringing spatial (somatotopic) information across the attachment site.
The pupillary restriction effect could come about through a simple diffusion mechanism. When severed nerves regrow, an enlargement of the nerve is produced at the growth site which is called a neuroma. A simple diffusion model of nerve growth factors at the neuroma site would account for the pupillary effect as fibers on the outside of the neuroma are farther away from the center of mass; fibers on the outside, for example, are exposed to growth factors only on the interior side. A progressive dropout of the external fibers leads to the pupil effect.
The pupillary restriction effect could come about through a simple diffusion mechanism. When severed nerves regrow, an enlargement of the nerve is produced at the growth site which is called a neuroma. A simple diffusion model of nerve growth factors at the neuroma site would account for the pupillary effect as fibers on the outside of the neuroma are farther away from the center of mass; fibers on the outside, for example, are exposed to growth factors only on the interior side. A progressive dropout of the external fibers leads to the pupil effect.
In addition to the development of a fixed pupil effect during neurogenesis, such pupillary effects could occur dynamically, mediated by the autonomic nervous system. In essence, blood flow changes to myelinated brain tracks could be part of the same mechanism that controls the pupil in the eye.
If indeed the connections of the attached hands undergo a pupil-like effect at the site of attachment, the thumbs and small fingers would be crossed. When the left cortex "looks for" the dominant right hand, it will find the right hand attached to the left side of the body. The small finger on the left hand, however, would be experienced as the thumb on the right hand, and vice versa. (Similarly, the top of the hand would be experienced as the bottom of the hand).
Such speculations attempt to connect the physics of the behavior of light passing through slits with the biological evolution of information transmission. There is a deeper possible connection having to do with models of brain computation. Briefly, the McCulloch and Pitts model of the nervous system starts with logical axioms and, through separation and relation (Boolean algebra) produces arbitrarily complex representations- the famous example being the working of the visual system in Hubel and Weisel's work. It has been suggested that the McCullock-Pitts computational model has practical limitations in representing spatial relationships, such as the rotation of mental images. Such a spatial limitation, if true, would carry over to space-like relational representation in the brain. In short, the current model of the brain identifies the interactions of neurons with a digital computer; the discovery of "pupil-like" effects would provide an additional computational model of the space-like representations of the brain as a camera obscura.
Researchers working with the hand transplants have a unique opportunity to test the "pupil hypothesis" experimentally. Functional MRI images of the left motor cortex of the subjects can be obtained while they repetitively flex their left thumb, and compared to the fMRI images when they repetitively flex their small left finger. In the motor cortex homunculus in normal subjects, the thumb is represented upside down below the small finger and on the opposite cortex. My prediction: the homunculus representation of the transplanted hands will be inverted again, with the thumb now "right side up" above the small finger.
Sincerely,
Charles Matthews M.D.
Neurologist
Raleigh N.C.
