I chose this topic from the scope of the neurosciences for a number of reasons.  First of all, perhaps because they are relatively rare (between 1985 and 1997, 563 therapeutic craniotomies were performed, of which only 20 were hemispherectomies; Smith et al., 1999), the subject of cerebral hemispherectomies (CHs) is not routinely covered in a normal neuroscience courses.  Second, CHs are, by nature, the direct application of neuroscience knowledge to human patients.  Third, the necessity of this procedure in humans provides a pool of experimental participants who can answer the kinds of experimental and introspective questions one could never address in non-human subjects.  Four, the dramatic recovery in the vast majority of patients bears scientific study for issues of plasticity, particularly in relation to traumatic brain injury.  This paper attempts to give an overview of the scientific perceptions of cerebral hemispherectomies and their significance to neuroscientific understanding.

Historical neglect of the concept of plasticity

One could make the case that our brain is who we are.  However, with a cerebral hemispherectomy, an individual can lose a significant portion of his brain mass and still retain or recover a remarkable amount of the original function.  This phenomenon of such incredible recovery following cerebral hemispherectomy addresses two of the central dogmas of neuroscience that happen also to be paradoxically contradictory.

Names such as Broca, Wernicke, Penfield, and many, many others are associated with the establishment of the concept of localization of function within the central nervous system.  However, a newer dogma, plasticity, has come into its own in recent years, most famously through the work of Hubel and Weisel (e.g., ocular dominance columns in cats) and Kandel (e.g., habituation in Aplasia), among others.  While the second concept does not negate the first, it presents occasions when function is not consistently or rigidly localized.

In a review of the historical literature Bach-y-Rita (1990) finds the plasticity concept was neglected for so many years for a number of reasons.  Perhaps the most salient of these is that many of the historically most prominent cases from the classical period of neuroscientists helped to establish localization of function.  The majority of these, as with the aforementioned Broca, examined older patients in whom plasticity yielded less sway than "hard-wired" function.  Thus, these patients did not typically recover and, indeed, this observation left such patients without a call for therapy, thereby creating a self-fulfilling prophecy of permanent loss of function.

Arguably, this (mis)perception persists to the present as is evident by the astonishment expressed in the case of Christopher Reeve's progress following his spinal injury.  Indeed, this famous case is not so usual as Franz (1915) noted that “significant” recovery from paralysis resulting from cerebral injury could persist “as late as 20 years after injury.”  Most strikingly, he commented that “[w]e should probably not (always) speak of permanent paralysis, but of uncared for paralysis” (Italics added).  Returning to Reeve’s case, his intense rehabilitation regimen has almost certainly made the progress that has continued to garner him headlines, but the prevailing neglect of the potential for plasticity has unfortunately made his level of treatment the exception.

As a result of the fact that plasticity has historically been underrepresented in the literature, Bach-y-Rita (ibid) contends that there has remained a poorly developed theoretical foundation available to guide researchers who might address these questions.  However, he does find that there is much evidence of plasticity in a number of forms in the literature and cites these as areas to consider in developing such theoretical guideposts.  Such areas include cellular and anatomic phenomena (e.g., sensory convergence), rehabilitation following traumatic injury (as covered above) and case studies in which recovery has been demonstrated.  Cerebral hemispherectomies represent opportunities to examine the course and mechanisms of recovery from several of these angles.

Issues surrounding cerebral hemispherectomies

Hemispherectomies are routinely performed in patients with intractable seizures (e.g., those induced by Rasmussen’s encephalitis, strokes, among other conditions).  This is a radical surgery, but one ultimately deemed necessary when all other therapies (e.g., drugs, special diets, etc.) have proven unsuccessful.  In spite of the apparently extreme nature of this procedure, functional recovery is commonly observed in these patients, particularly in the youngest where plasticity might be expected to be the greatest.

In the procedure, the hemisphere housing the focus of the seizure either is physically removed in what is known as an anatomical hemispherectomy or it may be surgically isolated from the rest of the brain.  This latter case is labeled a functional hemispherectomy.  One might view this procedure as the ultimate lesion study.  Unfortunately, because so much tissue is removed (or at least functionally deactivated), a number of confounds present themselves that might not otherwise be concerns in the controlled, focused lesions more commonly performed in order to study specific function.  For one thing, the serious nature of the surgery demands a relatively lengthy recovery period, thus preventing close monitoring of the progress of functional recovery in the earliest, perhaps most important stages.

Another problem with studying CH patients for general answers to hypotheses is that, primarily because of their relative scarcity, the majority of the literature tends to consist of case studies highlighting typically only one or two patients and reviews and meta-studies of this literature are difficult for a number of reasons.  To begin with, because this procedure affects nearly every conceivable aspect of the human condition, experiments address a plethora of areas (e.g., IQ, motor skills, language proficiency) and each of these may be tracked by any number of measures (particularly in the independently well-studied area of language where hundreds of instruments have been developed).  Contrast this with, say, the study of individuals with memory impairment.  The field is much more focused in those cases, aided as well by the fact that their subjects may be otherwise “intact” and high-functioning.  By this last point, I consider the possibility of, for example, impairment in language production.  Because an entire hemisphere was removed, it may be very difficult to determine whether the deficit is in accessing language or the motor skills to vocalize and/or write (I admit this is a flawed example, but I think the point is clear).

The issues of experimental design aside, the patients themselves present even greater hurdles to pooling data by virtue of the variance of the population and the idiosyncrasies inherent in their condition(s).  For example, while Rasmussen’s encephalitis (RE) is most frequently associated with CH, it is not the only cause for the procedure.  Different diseases may produce different effects prior to the surgery, so those must be taken into account.  In the case of RE, patients’ IQ diminishes as the disease progresses.  With a cerebral stroke, there may be pre-surgical loss of function in speech or motor areas.  Thus, the pathology of the original condition must be separated from the effects of the hemispherectomy.

Because CHs are performed most commonly on younger patients, developmental differences may also compound confusion.  The patient’s age at the onset of the disease, the severity of the disease, the age of the surgery, and the amount of post-surgical time that has passed before the patient is examined must be taken into account when analyzing data from this population.

The most obvious issue to consider in studying this CH patients is which hemisphere was excised.  This alone, of course, is not adequate to neatly partition data from subjects, as the general population is heterogeneous with respect to hemispheric localization of language (i.e., left, right, or even bilaterally).  Handedness is a general (though not perfect!) indicator of this factor, and is routinely noted in case studies.  Even then, some patients may be bilingual or have acquired a second language, further complicating standardized measures.  Actually, some of these “complications” may represent opportunities to address questions about language acquisition and some case studies have exploited some of these subpopulations to examine, for example, bilingual patients exclusively.

Cerebral hemispherectomy research in animals

The obvious solution to the issues of a limited human population and their inherent heterogeneity is to conduct animal research.  Surprisingly, as limited as the research on CH is in general, the vast majority of the publications on this topic tend to be case studies in humans.  This is likely due to a number of reasons.  First, CH is a relatively uncommon procedure, thus not attracting the same attention of more generalized topics like learning and memory.  (Incidentally, a parallel situation exists in traumatic spine injury research.  In her seminar last year at UNT, Dr. Lisa Rosenberg related that drug companies in general have little interest in this topic because the number of potential patients/customers for this type of injury is miniscule compared to their returns for developing, say, a new allergy medication.)

Animal research may also be avoided because CH is most relevant to humans, therefore they are studied directly in clinical settings primarily by physicians who naturally have the most contact with them.  Finally, as alluded to above, a CH does not target one specific region, thus making it difficult to draw conclusions about loss and/or recovery of function.

Cerebral hemispherectomy patients characterized

Immediately following hemispherectomy, patients generally have poor or virtually no use of contralateral limbs, although this tends to improve over time, with the extent of recovery depending on the age of the patient, course of the illness, and other factors mentioned above.

Patients often (thought not always) demonstrate "mirror movements" during the early stages of recovery.  These are movements made by the limbs contralateral to the excised hemisphere which mirror those performed by the unaffected limbs.  This mirroring is almost always of a lesser magnitude than the movements performed volitionally with the “working” limbs, and the phenomenon itself diminishes over a relatively brief period as the patient begins to acquire control of contralateral limbs.  These mirror movements have also been demonstrated in rats following CH while undergoing electrical stimulation (Machado et al., 2003).

Language is usually impaired to some extent in patients, with the typical pattern of recovery leaning to the notion of “younger is better.”  However, contrary to the conventional understanding of language acquisition, Curtiss et al. (2001) find that there may be multiple developmental windows for recovery with respect to language, and that the plasticity underlying the recovery may change in a nonlinear fashion.

Regardless, there is initially a more pronounced deficit which diminishes across time.  Generally, some deficits do persist, but patients several years beyond their surgeries often have developed coping strategies for avoiding, for example, complex grammar by breaking statements into smaller sentences.  Thus, they appear high functioning to the casual observer.  Interestingly, deficits tend to be generalized reductions in scores with respect to population norms across multiple aspects of language, rather than the aphasias of the more specific variety as were famously described by Broca and Wernicke.  This is all the more surprising given that, rather than a single, isolated lesion, an entire cerebral hemisphere has been removed!

Also surprising is that no other aphasias or primary sensory impairment are reported.  One might expect that perception would be altered permanently by such a sweeping alteration of brain architecture, but no such deficits are reported even in the short term.

Proposed mechanisms of recovery in cerebral hemispherectomy patients

The mechanisms underlying the remarkable recovery of patients following CH remain elusive.  Given that the main population studied under these conditions are human patients, it is not likely that the more invasive experimental manipulations (e.g., DNA chip gene screens, electrophysiological recordings, etc.) required to address the som of the necessary questions will be performed in the immediate future.  However, work certainly continues in other, parallel areas that may shed light on this apparently miraculous recovery of function.

The overwhelming view of the recovery of function following CH refers to the admittedly unknown mechanism as “plasticity.”   However, plasticity (as it is presently understood as the physical modification of existing synaptic connections and the establishment of new ones) almost certainly cannot completely explain the dramatic recovery of language, motor skills, and other aspects of function in such incredibly short time spans.  For example, the nearly complete (re)acquisition of language in the non-dominant hemisphere occurs over a period days.  One recalls Choamsky’s arguments against associative learning as the primary basis of language acquisition; that several words an hour would be learned, far too fast for a behaviorist model.  Similarly, “transference” of language to the remaining hemisphere happens much faster than even long term memory consolidation can realistically account for.

Based on this and other evidence, this idea of “transference” (though the term is routinely applied to this phenomenon) is clearly a misnomer in the implication that information is moved out of the damaged (and ultimately removed) hemisphere.  Instead, others have proposed that latent circuitry resides in each hemisphere, prepared to step in to restore function where the primary circuitry is rendered nonfunctional.   The existence of such hypothetical networks have yet to be demonstrated, although other hypotheses similarly lack concrete evidence to support them or refute this one.

Future directions

Conspicuously absent from this review is any comment on the effects on memory post-hemispherectomy, either in retention of pre-surgery memories or in the ability acquire new learning in any form.  This seemingly obvious area was neglected in the literature.  Why this is not tested directly and/or reported is not clear, but the primary emphases of this area of research tends to be on language and motor skills, perhaps because of well-developed rehabilitation programs, but this is only speculation.

Though less tangible in terms of quantitative measures, one might also expect the concept of personality to have been addressed in the literature, even anecdotally given the single-subject case study nature of many of the papers on these patients.  Unfortunately, this was never addressed.  Granted, the majority of patients were children when the surgeries were performed, but all the literature followed-up with then several years post-CH.  One might expect that the concepts of “left-brained” and “right-brained” personalities would be particularly exaggerated in these patients, as it has been reported that individuals have been known to undergo a personality change following a stroke or other unilateral brain injury, in which the personality tends to resemble that expected of the healthy, unaffected hemisphere.

While the subject of CH may appear esoteric to the general neuroscience community, it is more than an obscure surgical procedure.  Rather, these patients possess a unique set of data about a potential model system for certainly traumatic brain injury and possibly mechanisms for dramatic feats of plasticity.

Literature reviewed

Bach-y-Rita P. (1990). Brain plasticity as a basis for recovery of function in humans. Neuropsychologia. 28:547-54.

Chugani HT, Muller RA, Chugani DC. (1996).  Functional brain reorganization in children. Brain Dev. 18:347-56.

Curtiss S, de Bode S, Mathern GW.  (2001 ).  Spoken language outcomes after hemispherectomy: factoring in etiology.  Brain Lang. 79:379-96.

Franz SI, Sheetz M, Wilson A.  (1915).  The possibility of motor function recovery in long-standing hemiplegia. J. Am. Med. Assoc. 65:2150-2154.

Gemba H, Miki N, Sasaki K.  (1995).  Cortical field potentials preceding vocalization and influences of -erebellar hemispherectomy upon them in monkeys. Brain Res. 697:143-51.

Graveline CJ, Mikulis DJ, Crawley AP, Hwang PA.  (1998).  Regionalized sensorimotor plasticity after hemispherectomy fMRI evaluation. Pediatr Neurol. 19:337-42.

Helmstaedter C, Kurthen M, Linke DB, Elger CE.  (1997).  Patterns of language dominance in focal left and right hemisphere epilepsies: relation to MRI findings, EEG, sex, and age at onset of epilepsy.  Brain Cogn. 33:135-50.

Holloway V, Gadian DG, Vargha-Khadem F, Porter DA, Boyd SG, Connelly A.  (2000).  The reorganization of sensorimotor function in children after hemispherectomy. A functional MRI and somatosensory evoked potential study.  Brain. 123 Pt 12:2432-44.

Ishibashi H, Simos PG, Wheless JW, Baumgartner JE, Kim HL, Davis RN, Zhang W, Papanicolaou AC.  (2002).  Multimodality functional imaging evaluation in a patient with Rasmussen's encephalitis.  Brain Dev. 24:239-44.

Johnston, MV.  (2004?). Clinical disorders of brain plasticity.  Brain Development (In press; available on-line only at the time of this writing).

Machado AG, Shoji A, Ballester G, Marino R Jr.  (2003).  Mapping of the rat's motor area after hemispherectomy: The hemispheres as potentially independent motor brains. Epilepsia. 44:500-6.

Smith JR, Lee MR, Jenkins PD, King DW, Murro AM, Park YD, Lee GP, Meador KJ, Loring DW.  (1999).  A 13-year experience with epilepsy surgery.  Stereotact Funct Neurosurg. 73:98-103.

Stark RE, McGregor KK.  (1997).  Follow-up study of a right- and a left-hemispherectomized child: implications for localization and impairment of language in children. Brain Lang.  60:222-42.

Villablanca JR, Hovda DA.  (2000).  Developmental neuroplasticity in a model of cerebral hemispherectomy and stroke. Neuroscience. 95:625-37.