Few operations in epilepsy history sound as radical as hemispherectomy.
The name suggests removal of half the brain. In its original anatomical form, that is essentially what surgeons attempted: removing one cerebral hemisphere when catastrophic disease was confined to one side.
But the operation did not remain anatomically radical forever. Over the twentieth century, surgeons learned that the therapeutic goal was not necessarily to remove every piece of diseased tissue. The essential aim was to disconnect the pathological hemisphere from the rest of the brain so that seizures could no longer spread through its major pathways.
This shift — from removal to disconnection — produced functional hemispherectomy and later hemispherotomy.
The history is therefore not simply a story of increasingly aggressive surgery. It is almost the opposite: surgeons progressively removed less tissue while trying to preserve the same seizure-control effect.
Before epilepsy: Walter Dandy and hemisphere removal for tumour
The first major human series of hemisphere removal was not performed for epilepsy.
In 1928, American neurosurgeon Walter Dandy reported five patients who underwent removal of a cerebral hemisphere for infiltrating gliomas.
The operations were extraordinarily radical by the standards of the time, but they demonstrated something neurologically important: a person could survive removal of one cerebral hemisphere and remain ambulant.
The tumour results themselves were poor because malignant gliomas could not be cured simply by removing one hemisphere. But the operations established the technical possibility of hemispherectomy and influenced later epilepsy surgery.
The brain could reorganise more than expected
Dandy’s patients also challenged assumptions about how much neurological function depended on an intact hemisphere.
Motor and sensory deficits remained, of course, but survival and retained function showed that the human nervous system — particularly when damage had already occurred — could tolerate extraordinary anatomical loss.
This observation became especially relevant in children whose affected hemisphere had been severely damaged early in life.
If one hemisphere was already producing profound weakness and uncontrolled seizures, could removing it reduce seizures without creating an equivalent new loss of function?
1938: K. G. McKenzie and the first hemispherectomy for epilepsy
The first hemispherectomy specifically reported for epilepsy was performed by Canadian neurosurgeon Kenneth G. McKenzie.
McKenzie presented the case in 1938.
The patient was a 16-year-old girl who had suffered a severe early-life brain injury, developed left hemiplegia and later experienced disabling epilepsy.
McKenzie removed the right cerebral hemisphere.
Historical reviews report that the operation successfully relieved the seizures.
This was a very different surgical idea from removing a small cortical scar or tumour. Instead of identifying one limited epileptogenic focus, the entire damaged hemisphere was treated as the pathological system.
Why such a radical operation could make sense
Hemispherectomy was never intended as a general operation for epilepsy.
Its logic depended on a particular clinical situation.
The affected hemisphere had to be severely abnormal, the epilepsy had to be catastrophic or drug-resistant, and neurological function on that side was often already profoundly impaired.
In such circumstances, the diseased hemisphere might contribute relatively little useful function while generating repeated seizures that harmed the child or young person continuously.
The balance between loss and benefit therefore differed dramatically from surgery in an otherwise healthy hemisphere.
1950: Rowland Krynauw’s landmark series
The operation became widely known after South African neurosurgeon Rowland A. Krynauw published a major series in 1950.
His paper, Infantile Hemiplegia Treated by Removing One Cerebral Hemisphere, described 12 children who underwent hemispherectomy.
Ten had epilepsy as part of their clinical picture.
Krynauw reported marked seizure improvement in most survivors, together with behavioural and functional observations that attracted international attention.
The paper helped establish anatomical hemispherectomy as a possible treatment for severe childhood hemispheric epilepsy.
‘Infantile hemiplegia’ was a historical category
Krynauw’s terminology reflects its era.
‘Infantile hemiplegia’ was not one modern diagnosis.
It could include children with early destructive brain injuries, perinatal vascular damage, infection, developmental abnormalities and other causes of severe unilateral brain dysfunction.
Modern epilepsy surgery separates these causes much more precisely through MRI, pathology and genetics.
Historical records therefore should not be translated automatically into one contemporary disease category.
The anatomical hemispherectomy
In the classical anatomical operation, most or all of one cerebral hemisphere was removed while deeper structures such as the basal ganglia and thalamus were generally preserved.
The operation created a very large intracranial cavity.
At first, this seemed acceptable because seizure control could be dramatic.
Children who had endured frequent catastrophic seizures could become seizure-free or markedly improved.
But the long-term consequences of leaving such a large cavity were not yet understood.
Early success hid a delayed problem
The most serious complications did not always occur immediately after surgery.
Some patients recovered well and remained stable for years.
Then they deteriorated.
This delayed pattern made the problem harder to recognise because the operation could initially appear highly successful.
By the 1960s, reports began linking anatomical hemispherectomy with chronic intracranial bleeding, iron deposition, hydrocephalus and neurological decline.
Superficial cerebral hemosiderosis
One of the defining historical complications was superficial cerebral hemosiderosis.
Repeated small haemorrhages into the large postoperative cavity could release iron-containing blood products.
Over time, iron deposition accumulated along surfaces of the brain and spinal cord.
The process could damage neural tissue and was associated with progressive neurological deterioration.
In 1966, D. R. Oppenheimer and H. B. Griffith published an influential paper on persistent intracranial bleeding after hemispherectomy.
The report helped establish that the operation’s late morbidity was not simply an unfortunate coincidence but was related to the anatomy of the postoperative cavity.
Hydrocephalus became another major concern
Hydrocephalus — abnormal accumulation of cerebrospinal fluid within the ventricular system — was another important delayed complication.
Repeated haemorrhage, inflammation and changes in cerebrospinal-fluid circulation could contribute.
Some patients required shunt treatment.
The combination of late bleeding, superficial hemosiderosis and hydrocephalus caused enthusiasm for complete anatomical hemispherectomy to fall sharply.
A highly effective operation became difficult to justify
This created an unusual surgical dilemma.
The operation could be extremely effective for seizures.
But its delayed complications could emerge years later and could be devastating.
The question therefore became: could surgeons preserve the antiepileptic effect without leaving the same enormous resection cavity?
That question drove the next phase of innovation.
Theodore Rasmussen and the Montreal experience
Theodore Rasmussen, working at the Montreal Neurological Institute, became central to this transition.
Rasmussen had extensive experience with patients undergoing hemispheric surgery, including people with what would later become known as Rasmussen encephalitis.
By the 1970s, the Montreal group had recognised that complete anatomical removal was strongly associated with late complications.
Rasmussen observed that patients who had undergone large but incomplete hemispheric resections seemed less likely to develop superficial hemosiderosis.
This suggested that preserving some tissue might reduce the hazardous postoperative cavity.
The paradox: leave tissue behind, but make it electrically irrelevant
Simply leaving diseased cortex connected would not solve the epilepsy.
If the remaining hemisphere could still communicate with the opposite hemisphere, brainstem and other structures, seizures could continue to propagate.
Rasmussen’s solution was conceptually elegant.
Do not remove everything.
Instead, remove selected central and temporal regions and disconnect the remaining abnormal hemisphere from the rest of the brain.
This became known as functional hemispherectomy.
Functional hemispherectomy
In a functional hemispherectomy, substantial tissue is still removed, but less than in a classical anatomical hemispherectomy.
The remaining cortex is intentionally disconnected from major pathways.
The operation includes disconnection of callosal fibres linking the hemispheres and pathways connecting retained cortex with deeper structures.
The goal is functional isolation of the diseased hemisphere.
Historically, Rasmussen described and refined this strategy during the 1970s and later summarised the experience in his 1983 paper Hemispherectomy for Seizures Revisited.
Why the word ‘functional’ matters
The retained hemisphere is not ‘functional’ in the ordinary sense of continuing normal communication with the rest of the brain.
The term means that the surgery achieves the functional effect of a hemispherectomy — isolation of the diseased hemisphere — without anatomically removing all of it.
The distinction can be confusing.
A functional hemispherectomy is still a major destructive epilepsy operation.
It is simply less anatomically extensive than complete removal.
Hemidecortication and other intermediate approaches
Other surgeons also explored ways to reduce tissue removal.
Hemidecortication involved removing much of the cerebral cortex while preserving more underlying white matter and avoiding the same large cavity.
Several modifications appeared as surgeons attempted to balance three competing goals: complete seizure-producing network disconnection, lower blood loss and fewer late complications.
The terminology in older literature is not always consistent, and some authors used hemispherectomy, hemidecortication and functional hemispherectomy differently.
Modern readers therefore need to examine the actual surgical technique rather than relying only on the operation’s name.
From hemispherectomy to hemispherotomy
By the 1990s, the logic of disconnection had been pushed further.
If the therapeutic goal was to sever the abnormal hemisphere’s major connections, perhaps even less brain needed to be physically removed.
The term hemispherotomy came to describe procedures based predominantly on disconnection rather than resection.
In broad terms, hemispherectomy removes; hemispherotomy disconnects.
The difference is not absolute because hemispherotomy usually still requires some tissue removal to reach and divide critical pathways.
Delalande and vertical parasagittal hemispherotomy
French neurosurgeon Olivier Delalande and colleagues developed the vertical parasagittal hemispherotomy approach in the early 1990s.
The technique used a relatively limited surgical corridor near the midline to disconnect the pathological hemisphere.
The approach aimed to minimise cortical removal while completing the essential disconnections.
Later reports documented long-term outcomes in large paediatric series.
Delalande’s work was one of several major contributions to the modern hemispherotomy era.
Other hemispherotomy techniques
Surgeons including Jean-Guy Villemure and Johannes Schramm developed other disconnective approaches.
Lateral and vertical techniques differed in the route used to reach the corpus callosum, ventricular system, temporal structures and deep white-matter connections.
The procedures were technically different but shared the same central principle: preserve vascularised brain tissue where possible while disconnecting the pathological hemisphere completely.
The operation became a problem of network disconnection
This historical transformation is important because it mirrors a broader change in epilepsy thinking.
The earliest surgery often focused on removing a visible lesion.
Hemispherectomy treated an entire damaged hemisphere as the epileptogenic system.
Hemispherotomy went one conceptual step further: the pathological hemisphere did not necessarily have to be removed if its ability to participate in the wider seizure network could be abolished.
The operation became an exercise in disconnecting pathways.
Incomplete disconnection became a new form of surgical failure
Removing less tissue solved some complications but introduced another challenge.
If even a small but important pathway remained connected, seizures could continue or recur.
Modern hemispherotomy therefore depends on precise anatomical knowledge of commissural, projection and limbic pathways.
In some patients, recurrent seizures after surgery are found to result from incomplete disconnection and may require revision surgery.
This was a different failure mechanism from the late bleeding complications that had plagued anatomical hemispherectomy.
Which conditions drove the procedure’s development?
Over time, hemispheric surgery became associated particularly with severe unilateral epilepsies such as large perinatal strokes, hemimegalencephaly, extensive cortical malformations, Sturge–Weber syndrome and Rasmussen encephalitis.
The common feature is not the disease name itself.
It is that the epileptogenic abnormality is distributed so extensively through one hemisphere that a small focal resection is unlikely to control the seizures.
These modern indications were refined through imaging and electroclinical investigation, but the historical principle originated in the severely damaged hemispheres described by McKenzie and Krynauw.
Children became the main population
Hemispherectomy began in adults with tumours.
Its enduring place in epilepsy surgery became primarily paediatric.
There are several reasons.
Catastrophic hemispheric epilepsies often begin early in life.
Children may already have severe weakness and loss of useful hand function on the affected side.
And the developing brain has greater capacity for functional reorganisation than an adult brain, although plasticity is neither unlimited nor guaranteed.
The modern procedure is therefore strongly associated with specialised paediatric epilepsy centres.
Language and the developing brain
One of the most striking consequences of early hemispheric injury is that language function can reorganise.
Children with severe damage to one hemisphere may develop language networks that rely more heavily on the opposite hemisphere.
This does not mean that removing or disconnecting a hemisphere has no cognitive consequences.
Rather, it explains why carefully selected children can sometimes retain or develop substantial language abilities after an operation that would seem neurologically impossible if judged only by adult localisation models.
Motor outcome was never simply ‘half the body becomes paralysed’
Many children considered for hemispherectomy already had substantial hemiparesis before surgery.
Gross walking ability may remain possible after surgery, particularly when it was present beforehand.
Fine hand function on the opposite side is usually much more limited.
Historically, Krynauw’s reports were striking partly because some children remained ambulant and retained useful movement after such extensive surgery.
Those observations contributed to growing understanding of developmental neuroplasticity.
Why anatomical hemispherectomy did not disappear completely
The historical story is sometimes told as though anatomical hemispherectomy was simply abandoned and replaced.
That is too simple.
Disconnective techniques predominate in modern practice because they reduce several complications, but some centres still use anatomical hemispherectomy in carefully selected situations.
Modern surgical technique, haemostasis, imaging, intensive care and management of hydrocephalus differ greatly from the mid-twentieth century.
Recent series have therefore re-examined whether anatomical removal may still have a role in selected patients with unusual anatomy or a high risk of incomplete disconnection.
The historical trend is toward disconnection, not an absolute prohibition on resection.
The terminology still causes confusion
In modern literature, the word hemispherectomy is sometimes used as an umbrella term for both anatomical removal and functional disconnection procedures.
Hemispherotomy usually refers more specifically to predominantly disconnective operations.
That means a modern paper saying ‘hemispherectomy outcome’ may include patients who did not have an entire hemisphere physically removed.
This is one reason historical interpretation requires attention to surgical technique rather than name alone.
What changed over a century?
In 1928, Dandy demonstrated that a human could survive removal of a cerebral hemisphere.
In 1938, McKenzie applied the operation specifically to epilepsy.
In 1950, Krynauw’s paediatric series established the procedure as a serious epilepsy treatment.
In the 1960s, delayed bleeding, superficial hemosiderosis and hydrocephalus exposed the long-term hazards of the large postoperative cavity.
In the 1970s, Rasmussen developed the functional hemispherectomy concept: remove less, disconnect more.
In the 1990s, hemispherotomy techniques pushed that principle further by minimising resection while maximising disconnection.
Modern surgery retains that basic philosophy.
The central historical lesson
The history of hemispherectomy is not really the history of surgeons becoming more willing to remove brain.
It is the history of learning how little brain needed to be removed.
The original operation treated catastrophic hemispheric epilepsy by anatomical elimination.
Its success showed that severe unilateral epilepsy could sometimes be controlled surgically.
Its complications then forced surgeons to rethink what the operation was actually trying to accomplish.
The answer was not ‘remove half the brain’.
The answer was: isolate the pathological hemisphere from the networks through which its seizures spread.
That conceptual change turned hemispherectomy into hemispherotomy and transformed one of the most radical operations in neurosurgery into a sophisticated disconnection procedure.
Sources and further reading
Lew SM. Hemispherectomy in the treatment of seizures: a review. Translational Pediatrics. 2014/2015.
Anatomic hemispherectomy: historical perspective. Neurosurgical Focus. 2012.
Functional hemispheric disconnection procedures for chronic epilepsy: history, indications, techniques, complications and current practice in Europe. EANS consensus statement. 2024.
Krynauw RA. Infantile hemiplegia treated by removing one cerebral hemisphere. Journal of Neurology, Neurosurgery & Psychiatry. 1950.
Oppenheimer DR, Griffith HB. Persistent intracranial bleeding as a complication of hemispherectomy. Journal of Neurology, Neurosurgery & Psychiatry. 1966.
Rasmussen T. Hemispherectomy for seizures revisited. Canadian Journal of Neurological Sciences. 1983.
Delalande O et al. Vertical parasagittal hemispherotomy: surgical procedures and long-term outcomes in 83 children. Neurosurgery. 2007.
Functional hemispheric disconnection procedures for chronic epilepsy: EANS consensus statement. 2024.
Historical evidence checked: September 2026.