C H A P T E R

28

Future Developments
in Neurosurgery

Marshall B. Alien, Jr.

When this book was being planned, there was a consensus
among the authors that a chapter should be devoted to
anticipated developments in neurosurgery. One objective
might be to give the resident, the future neurosurgeon, some
background and ideas for subjects that might be fruitful for
investigation; another might he to help orient special train-
ing.

A logical approach to such an endeavor might be to

review current literature. Since it is impossible to survey
every area that might offer future contributions to neurosur-
gery, five topics have been selected because of their excep-
tional promise—in some cases because development in a
given area appears to be complete but applications to clinical
neurosurgery have not yet been generally instituted, and in
other cases because preliminary reports suggest promise but
techniques in the area have yet to be developed.

Without question, much of what is presented here will

prove to have been guesswork ten years from now, but at
this time some areas that appear promising include: (1) the
use of magnetic impulses to stimulate and record the ner-
vous system, (2) fetal surgery, (3) treatment of vasoconstric-
tion from vasospasm and arteriosclerosis, (4) transplantation
of neural tissue, and (5) computer-assisted stereotactic pro-
cedures that may be applied to open surgery and radiother-

apy.

One area that will not receive attention in this review is

the manipulation of genes, which in the future will play
without doubt a major role in the approach of the physician
to the diagnosis and management of patients who are or
might be candidates for certain neoplasms and degenerative
diseases. Genetic research and manipulation will influence
the course of neurosurgery, as well as oncology and medi-
cine in general. The implications of genetic manipulation
appear to have such a broad and profound impact that a
meaningful discussion in this text is not possible. For the
same reason, anticipated developments in reconstruction of
the spine will not be discussed either. Current developments

have been reviewed elsewhere in this text; future develop-
ments will depend on types of instrumentation that have yet
to be developed. Present-day technology and technique is
progressing so rapidly that it is difficult to remain current,
and it is equally difficult to analyze definitively the results
of any particular type of instrumentation. For future devel-
opments in reconstruction of the spine the reader is referred
to Chaps. 19 and 20.

This chapter ends with some general comments on the use

of computers in making clinical decisions. At present, com-
puters play a major role in imaging, orientation of stereotac-
tic probes, planning radiation therapy, and many administra-
tive hospital operations—to say nothing of the parts they
play in teaching and institutional accounting. Doubtless, in
future, computers will play an even greater part in the
planning and selection of therapies for patients, as well as in
the evaluation of physician services. The human elements of

judgment, ability, and experience must not be abandoned,

however.

THE USE OF MAGNETIC IMPULSES
TO STIMULATE AND RECORD

NERVOUS SYSTEM FUNCTION

Magnetic impulses may be used to stimulate and record the
electrophysiologic activity of the nervous system.''

2

Probes

emitting such magnetic impulses can deliver stimuli to either
peripheral nerves or the brain with effects similar to those
set off by electrical stimulation. One advantage of magnetic
stimulation is activation of the nervous system—either the
peripheral nerves or brain—without penetration of the skin;
the stimulation is virtually painless. Transcutaneous stimuli
may be used to provide evoked responses of the sensory and
motor pathways.

Magnetic impulses are generated by passing large electri-

-525

cal currents through a coil.

3

The strongest magnetic stimuli

are produced by loops at the periphery of the coil.

1

Weaker

impulses are generated by loops near the center of the coil,
and no energy is generated at the coil's center. This effect
within the coil results in imprecise sites of stimulation,
which is of significance in the stimulation of peripheral
nerves and functional cortical areas.

Peripheral nerve stimulation is accomplished with round

coils which have an outside diameter of up to 9.2 cm or with
pointed probes, which measure 8.6 x 10.6 cm.

4

They may be

applied to the skin overlying the nerve that is to be stimu-
lated in one of several different ways: orthogonally, longitu-
dinally, tangendally, off-center, symmetrically, obliquely, or
transversely. Coils do not require the use of substances to
connect them to the skin. They do not penetrate the skin,
although, for convenience, they usually are applied to the
skin surface. It is not necessary to remove clothing for
stimulation.

5

Magnetic impulses travel through body tissues painlessly

and without attenuation between the probe and the nervous
tissue being stimulated.

3

Stimulation may be applied to

peripheral nerves, nerve roots in the spinal canal, at the base

of the brain, or to the motor or sensory cortex... ,

1

.

3

-

12

Magnetic stimuli penetrate to considerable depths without
causing large electrical fields at the surface. By contrast,
electrical stimulation does cause pain, and impulses obtained
by recordings are attenuated and distorted, as are electrical
impulses applied for stimulation.

Magnetic impulses decrease in magnitude in relation to

the inverse square of the distance from the coil. Currents
induced at the surface of the skin by magnetic stimulation
will be oriented parallel to the skin and will thus activate
structures from a different orientation than that of electrical
stimuli.

3

Use of magnetic stimulation provides the capability

of obtaining evoked responses to percutaneous stimulation
of nerve roots as well as stimulation of the motor cortex.

12

-

13

Interestingly enough, latencies before motor responses
evoked by stimulation of the motor cortex are shortened
when muscles are slightly contracted voluntarily.

The capabilities of magnetic stimulation have already

demonstrated their usefulness in the diagnosis of multiple
sclerosis, degenerative ataxic disorders, and nerve root com-
pression by hemiated intervertebral disks.

12

'

14

'

15

Latencies

are significantly prolonged in patients with multiple sclero-
sis, and the degree of prolongation is greatest in patients
who have the most severe clinical involvement.

15

Significant abnormalities are found in patients with Fried-

reich's ataxia and other conditions involving demyelination
or degeneration of the CNS.

16

Stimulation of lumbosacral

roots by stimulation in the central sacral region reveals
reduction in the compound muscle action potentials and
prolongation of the latencies in patients with radiculopathies

proved b) surgery or imaging studies.

12

Cerebral stimulation for evoked potential monitoring to

identify sites of involvement with spondylosis and use of
electrical current during the course of intraspinal surgery are
currently recommended. Cortical stimulation is thought to be

superior to somatosensory evoked potential monitoring since
motor pathways may be monitored.

17

-

18

Activation of the

motor cortex with magnetic stimuli should prove superior in
this type of monitoring since the stimulation is more easily
applied and painless, but responses evoked in the lower
extremities by transcranial stimulation with magnetic stimuli
are of smaller amplitude and more difficult to elicit than
evoked responses in the hand.

3

Stimulation probes are bulky and the points of activation

of peripheral nerves are imprecise, rendering the use of
magnetic stimulation impractical for nerve conduction stud-
ies, except in those cases where proximal nerve trunks or
nerve roots are stimulated.!'

6

'

12

However, sites of stimulation

of peripheral nerves are not imprecise enough to influence
significantly the determination of latencies of evoked poten-
tials.

Magnetoencephalography (MEG) is the recording of mag-

netic signals from the brain. Electrical activity at the junc-
tion of the dendrite with the body of the neuron produces a

magnetic field.

19

Magnetic signals are weak, about 10-

13

tesia when compared to the normal geomagnetic field of 5 x

10-

8

tesla.

2

Typical urban noise is recorded as about 5 x 10-

7

tesla/Hz. Thus, magnetic signals from the brain are about
one billionth the strength of the earth's magnetic field.

19

MEG became feasible with the development of sensitive
superconducting quantum interference devices (SQUIDS)
and gradiometers, systems of detector coils that filter signals
originating within the body from distant sources of
noise.

2

'

19

'

20

Since 1975 there has been a steady development

of MEG systems, evolving from single-channel probes to
current systems that comprise up to 37 channels. It is specu-
lated that future systems may have as many as 100 channels,
allowing for monitoring of the entire brain with a single
probe placement.

1

'

21

Scalp electroencephalographic recordings indirectly re-

flect the electrical activity of the brain: Electrical currents
are transmitted and their volume conducted through the
brain, skull, and scalp, which are sources of inhomogeneous
impedance that result in attenuation and distortion of electri-
cal activity. MEG has the theoretical advantage that mag-
netic signals are not significantly affected by the medium
through which they pass.

2

Magnetic potentials are recorded from the cranial surface

as two finite regions of opposite polarity (extrema) that are

juxtaposed. The generator of these extrema lies deep along

the bisection of the line joining the extrema, and its direction
is perpendicular to the line. Its depth is linearly related to the
distance between the extrema.

2

The location of the generator is found by an inverse

solution. A number of spheres of known radii are used to
approximate mathematically the patient's recorded extrema.
The radius of whichever sphere most closely approximates
the measured extrema is then used to calculate the source
location.

22

The head is not an absolute sphere, particularly in

low-convexity regions such as the temporal. This renders the *
inverse solution potentially less accurate in these areas.
However, comparison of the spatial resolution of electrically

FUTURE DEVELOPMENTS IN NEUROSURGERY 527

and magnetically recorded epileptiform and evoked poten-
tials indicates that the inverse solution usually localizes
magnetic dipoles to within approximately 1 cm of electri-
cally recorded potentials.

MEG compares well with other functional imaging tech-

niques and depth electrode recordings for determining sites
of epileptic foci.

23

-

24

'

25

EEG and MEG recordings of interic-

tal events often show quite different wave forms and differ-
ent temporal relations, which may reflect different physio-
logical processes or generators. Single-channel MEG
recordings have demonstrated abnormal signals from epilep-
tic patients, even when abnormal signals were not present on
EEG.

23

Neuromagnetic signals have proven to be sufficient

to establish the location, depth, orientation, and polarity of
currents underlying paroxysmal discharges.

24

-

25

An orderly

magnetic field pattern accompanies interictal spikes.

26

Si-

multaneous multichannel MEG recordings (e.g., 37-channel
MEG) may resolve some of the spatiotemporal discrepancies
between EEG and limited-channel MEG recordings.

EEG and MEG compliment each other in the investigation

of sources of epilepsy. In both EEG and MEG, the signal
recorded depends on the relative orientation of the discharg-
ing dipolc and the recording apparatus. When combined, the
EEG and MEG offer good possibilities for localizing sites of
cerebral activity, either spontaneous or induced.

MEG recordings of evoked responses may aid in the

localization of segments of the somatosensory cortex acti-
vated by peripheral stimulation.

27

Various components of the

somatosensory response to stimulation of nerves in the arms
or legs provide different dipole sources in the projection
areas. Both ipsilateral and contralateral responses are

recorded from secondary somatosensory areas.

28

Neuromagnetic recordings of the human auditory cortex

have demonstrated that sounds of different frequencies pro-
duce responses at different sites in the primary auditory
cortex.

29

There is an orderly progression in the tonographic

map within the auditory cortex. This finding supports the
hypothesis that there is a direct mapping of the cochlea on
the cortex. Responses vary in amplitude according to the
alertness of the subject.

MEG studies of patients during attacks of migraine have

shown changes similar to those of spreading depression in
laboratory animals, causing a biphasic slow wave followed
by reduction of activity for as much as 10 min.

This brief review concludes that magnetic stimulation

will provide an effective means of obtaining evoked poten-
tials in alert patients being investigated for various CNS
diseases, particularly multiple sclerosis, and other degener-
ative disorders in the future. Magnetic stimulations may
also be used increasingly as a means of investigating radi-
culopathies. If the instrumentation can be significantly im-
proved, magnetic stimulations will surely play a major role
in nerve conduction studies. Magnetoencephalography
shows great promise as a complimentary modality for the
localization of sources of epileptic activity, and it may play
a role in diagnosing a number of other illnesses such as
migraine. It presents a most valuable research tool for the

topographic localization of various functional areas of the
human cerebral cortex.

FETAL SURGERY

Major changes in pediatric neurosurgery during the next
decade will involve the delineation of indications and the
development of techniques for the general institution of
surgery in the correction of congenital defects in fetuses. A
review of fetal surgery with its expectations during the
current decade has been published recently, and this section
will draw heavily from that review.

30

Hydrocephalus and myelomeningoceles are the two most

common congenital defects involving the nervous system.
There are indications that the earlier hydrocephalus is
treated, the less severe the deficits may be. It could be that
prenatal treatment of dysraphism could reduce the conse-
quent deformities since a more physiological development
might result.

Prenatal detection of abnormalities in the development of

the neural tube has been made possible by the availability of
ultrasound and examinations of amniotic fluid for karyo-
types, alphafetoprotein, and acetylcholinesterase—tech-
niques which are relatively common now; however, it must
be recognized that even with detailed examination 20 to 40
percent of abnormalities will be missed.

30

There have been reports of attempts to treat hydrocephalus

in utero since 1981.

30

-

35

These have primarily consisted of

inserting ventriculostomies so that fluid from the ventricles
can drain into the amniotic cavity—the hypothesis being that
the earlier the intraventricular pressure is reduced, the less
damage there will be to the cerebral mantle.

Evans et al. indicate that as of July 1, 1989, there had

been 45 cases reported to the international registry, the
shunts being performed at 25 ±2.73 weeks with a range of

18 to 31 weeks of gestation.

30

Thirty-four out of forty-one

treated fetuses had survived. Of the seven fetal deaths, four
were due to trauma at the time of surgery or the onset of
premature labor, which occurred within 48 hours of the
operative procedure. Of the 38 survivors with a follow-up of

12.2 ±5.8 months (range, 6-36 months), 14 are reported to

be normal. The remaining 24 have varying degrees of neuro-
logical deficits with 18, or 53 percent, of the survivors
having severe handicaps. This indicates that about one-third
of the entire series was normal; however, ventriculostomies
have allowed the remainder of the abnormal fetuses to
survive.

Most of the intellectual and neurological deficits were due

to associated lesions, cortical blindness, seizure disorders,
and spastic diplegia.

30

In this series of patients there were

cases which could have been rejected for shunting had
selection criteria been more rigidly followed. Individuals
with holoprosencephaly and autosomal trisomies should
have been excluded from surgery. All were late in reaching
their developmental milestones and had tested less than 60

528 CHAPTER 28

in developmental quotients. The conclusion is that the re-

sults have been quite disappointing, but efforts should con-
tinue with stricter selection criteria since the history of

untreated hydrocephalus is so poor. There should be contin-
ued attempts at refining techniques in the selection and
treatment of these devastating lesions.

There are reasons to believe that at least some of the

severe effects of myelomeningocele might be reversed by

prenatal repair. For the selection of cases for fetal surgery
there must be a knowledge of the natural history of the
lesion which is being treated, an assessment of the danger to
the mother and infant, and a practiced team with good
equipment for carrying out the surgical procedure in utero.
One of the greatest problems with such prenatal surgery is
the induction of premature labor. A combination of sedatives
and prostaglandin inhibitors is used to suppress uterine con-
tractions.

Much experience and success in the treatment of

diaphragmatic hernia in utero is being accumulated despite
an exceptionally high mortality rate in the initial tries.

38

Normally, there is a very high postnatal mortality rate with
this lesion.

Since these are such complex procedures with so many

implications—including the lives of two individuals—and
since the candidates for fetal surgery are so rare, institutions
offering such therapies will surely be limited, hut fetal
surgery for congenital defects will undoubtedly become
available in this decade.

30

-

38

TREATMENT OF IMPAIRED
CEREBRAL PERFLSION FROM
CHRONIC VASOSPASM AND
ARTERIOSCLEROTIC NARROWING

that cerebral vasospasm due to subarachnoid hemorrhage
and trauma to the vessel walls responds to transluminai
dilatation of the vessels by microballoons. This is encourag-
ing and consistent with the reports which suggest that vaso-
spasm is a pathological condition of the vessel walls.

44

-"

Evaluation of cerebral vasospasm by use of transcranial
Doppler techniques offers hope that we will soon have
specific indications for such therapy.

46

The finding that anastomosis of extracranial blood vessels

to branches of a middle cerebral artery does not provide

statistically significant evidence of clinical improvement was
a disappointment to most surgeons who were charged with
the care of stroke patients.

47

'

48

Other techniques designed for

revascularization or for interrupting the relentless course of
degenerative lesions of blood vessels supplying the brain
must be analyzed for effectiveness in preventing cerebral
ischemia. Techniques that have appeared include (1) omental
transplantation, (2) laser angioplasty, and (3) balloon dilita-
tion.

Goldsmith, Chen, and Duckett recommended revasculari-

zation of the brain by intact omentum almost 20 years ago.

49

Through a series of subsequent reports Goldsmith has
presented a strong case for the use of such techniques to
prevent or treat cerebral ischemia.

50

-

53

On superficial obser-

vation, there is a strong reason to question whether this
technique offers much that has not been offered by the
by-pass procedure. The technique has been supported by
the Chinese efforts and certainly deserves comprehensive
evaluation.

54

-

57

Of considerable interest, also, are a series of reports

indicating success with the use of lasers in transluminai
angioplasty. Most of these reports are related to angioplasty
in coronary vessels, but the techniques have been applied to
other vessels and may be applicable to vessels supplying the
brain.

58

"

60

CEREBROVASCULAR SURGERY

For many years chronic vasospasm has presented a great
challenge to surgeons treating subarachnoid hemorrhage.
The incidence of neurological deficits and even mortality in
patients who have recently experienced subarachnoid hem-
orrhage has been disappointing and has rendered surgery
hazardous during the period between 2 days and 2 weeks
following a subarachnoid hemorrhage. An almost incalcula-
ble number of drugs have been tried in the treatment of
chronic vasospasm, without success. The only therapy that
has enjoyed consistent success has been the use of artificial
hypertension and volume expansion, approaches that would
appear to be contraindicated while defects in vessel walls are
present.

39

^

3

Treatment of ruptured aneurysms in the early

period after hemorrhage is now common.

The pathophysiology of chronic vasospasm has been de-

bated, with many contending that the clinical state is due to
muscle contraction, but growing evidence suggests that the
basic lesion is a pathological state. Recent reports indicate

TRANSPLANTATION OF
NEURAL TISSUES

The history of transplantation of neural tissues in mammals
extends over a hundred years.

60

-

61

Thompson reported on the

implantation of a segment of cortex in a dog's brain in

1890.

62

He speculated that transplantation of neural tissues

would have a place in future research. The first well-docu-
mented survival of a graft to the CNS in newborn rats was
reported in 1917 by Dunn.

63

In 1930, May reported on

intraocular grafts of tissue removed from the CNS, and in

1940, Le Gros dark gave detailed descriptions of the sur-

vival of cortical transplants.

64

-

65

The anterior chamber of the eye became a favorite target

for implanting grafts since the transplanted tissues could be
observed there. This was the site that Greene and Arnold
chose to transplant embryonic grafts of CNS tissue, which
they reported in 1945,

66

and in 1970, it was the site that

Olson and a subsequent cadre of colleagues used.

67

In an-

FUTURE DEVELOPMENTS IN NEUROSURGERY

529

other report of the same year Olson also demonstrated ax-
onal growth and secretion of neurotransmitters by the trans-
planted neurons.

68

It was this series of reports that sparked

great interest and investigation into the possibilities of neural
tissue transplantation.

Lund and Hauschka, using fetal transplants of CNS tissue,

reported on the synaptic specificity in mammalian brain, and

Stenevi, Bjorklund, and Syendgard demonstrated that neu-

rons taken from fetal brains could survive transplantation
into adult mammalian brains. They outlined the factors im-
portant for successful neural transplantation in adult host rat
brains.

69

-

70

Perlow et al., shortly thereafter, showed that the

transplanted fetal neurons could survive, grow axons, and
produce neurotransmitters with resultant reversal of motor
abnormalities in host animals with movement disorders that

resembled Parkinson's disease.

71

In subsequent reports, evi-

dence has been presented that animal models with lesions
resulting in learning deficits somewhat resembling Alz-
heimer's disease would respond to the implantation of dopa-
minergic and cholinergic neurons, and they reported signifi-
cant improvement in swimming performance in old animals
that had undergone grafting.

72

These and many other reports of experimental successes

in the treatment of animal models with lesions of the septo-
hippocampal system by grafting with monoaminergic and
cholinergic neurones lead to interest in clinical trials.

73

-

75

Even more dramatic successes in the treatment of animals
having experienced lesions in the nigrostriatal pathways by
grafting with mesencephalic-dopamine-rich neurons taken
from 16-17 day old fetuses support this.

71

'

76

Evidence is that tissues most likely to provide appropriate

neurotransmitters leading to improvement in learning or the
treatment of movement disorders would be those that con-
tain monoaminergic or cholinergic neurons or neurons capa-
ble of producing dopamine or acetylcholine.

61

Sources for

donor cells which have shown promise are: (1) the septum or
brain stem of fetuses, from which most experimental mate-
rial has come; (2) the peripheral nervous system; and (3) cell
lines.

There is evidence that the sympathetic ganglia and adrenal

medulla of mature individuals produce the neurotransmitters
necessary for successful reimplantation, and that mature
neurons from these organs would grow axons and produce
neurotransmitters. Thus the peripheral nervous system pro-
vides for the capability of making autologous grafts. Neur-
oblastomas, tumors of immature cells of the nervous system,
reportedly, can be converted into mature cells in culture, and
it is possible to differentiate some lines of cells which are
cholinergic, others which are adrenergic, and still other lines
which are mixed. Thus it is theoretically possible to develop
lines of cells which might be selected for maturation and
implantation.

77

'

61

While animal fetal tissues may be easy to obtain, many

ethical and moral questions are raised when one considers
using cells of aborted human fetuses for transplantation.
Chances for a successful program using such tissues would
appear to be remote. Transplantations using adrenal medulla

have been attempted in many patients with varying degrees
of success reported. There have been major complications,
resulting in part from the fact that two major operative
procedures are being performed on a debilitated patient. The
procedure has been abandoned in most centers in the United
States at present. So far, there has been very little success in
the development of a cell line from neoplasms suitable for
transplantation, and we are aware of very few attempts to
implant such cells. However, the technique is appealing, and
we are anticipating some continued experimental trials if a
suitable cell line becomes available.

In summary, it appears likely that neural tissue transplan-

tation may have much to offer in the treatment of degenera-
tive lesions of the nervous system, and it would appear that
alternative sources of transplantable tissue should be investi-

gated.

NEW DIRECTIONS FOR
STEREOTAXIC SURGERY USING

COMPUTER ASSISTANCE

The association of late-generation computers, computerized
tomography (CT), magnetic resonance imaging (MRI), and
lasers as well as radiation therapy with stereotaxis has revi-
talized and expanded the capabilities for stereotactic surgery.
The first stereotactic instrument for use in animals was
described by R.H. Clarice and V. Horsely in 1906,

78

and this

was followed 2 years later by the classic investigation in
which they reported on their examination of deep cerebellar
nuclei in small animals.

79

Stereotactic instruments for ani-

mals were designed so that measurements were based upon a
reference plane connecting the external auditory meatus and
the inferior rims of the orbit. After considerable success with
their anatomical studies, Clarke concluded that stereotaxic
methods might be applied to human beings. A disagreement
with Horsley over this recommendation ended their long and
productive relationship.

Other devices and instruments for use in animals and

humans were eventually developed. In 1947 Drs. Speigel
and Wycis and their associates reported on the application of
an instrument for stereotactic operations in humans.

80

Sev-

eral different instruments were developed.

81

'

82

Stereotactic

surgery became very popular, primarily for the treatment of

movement disorders but also for the treatment of pain and
psychiatric conditions. Stereotactic biopsies of brain sub-

stance and the production of lesions deep within the brain
were possible even from the time instruments were first
developed. Initially, human stereotactic atlases related deep
cerebral target structures to anatomic landmarks in the dien-

cephalon identified by ventriculography.

83

-

85

Problems arose relating to the ethics of treating psychiat-

ric disorders. With the use of levodopa for the treatment of
Parkinson's disease in the 1960s, there was a great decline in
the numbers of stereotactic procedures performed.

The development of computerized tomography (CT) led

to the adaptation of stereotactic instruments using CT guid-
ance.

86

-

88

This provided the capability of localizing lesions

deep within the cerebral tissue, open biopsy of which might
produce increasing neurological deficits.

88

-

90

Biopsy of such

lesions, using stereotactic methods for precision, became
commonplace.

91

-

92

This increased the number of indications

for stereotactic procedures.

Development of the carbon dioxide laser has added

new capabilities to the stereotactic approach to deep-seated
intracranial lesions. Now neoplasms could be treated.

93

Kelly and his associates have reported on the treatment of
large series of patients with deep—and even superficial—in-
tracranial lesions treated with stereotactic instrumentation,
resections being accomplished with the carbon dioxide
laser.

94

-

97

Kelly has also used these techniques to resect

vascular lesions, radiation necrosis, and epileptic foci. De-
tails of the technique as used for volumetric resection of
neoplasms are discussed in Chap. 23.

• Other areas of development in stereotactic surgery such as

stereotactic robotics, frameless stereotaxis, and holographic

guidance of stereotaxis are also discussed in Chap. 23.

THE ROLE OF COMPUTERS IN

CLINICAL DECISIONS

The past decade has witnessed a phenomenal growth in the
role of computers in daily life. In hospitals, computers are
indispensable in imaging, teaching, the distribution of infor-
mation, and in accounting. Computers are playing an ever-
increasing role in teaching, especially in teaching us how to
make decisions. Now computers are assisting in the clinical
decision-making process. ;''

Computers have the ability to expand the basis upon

which decisions are made and to separate facts from emo-

tions. In a recent series of reports, emanating from the

Department of Surgery at the Riyadh Armed Forces Hospital
in Saudi Arabia, patients who were considered possible

candidates for termination of life support were evaluated by
their physicians, their nurses, and by a computer program
based upon an "acute physiology and chronic health evalua-
tion" (APACHE II).

98

A false positive anticipation of death

of between 6.7 percent and 16.7 percent occurred among the
human evaluators, but never by the computer analysis.

99

Use

of daily APACHE II scores improved the predictive capabil-
ity of the computerized system by a factor of 4. The system
was not recommended for clinical decisions at the time of

publication, but one could clearly envision benefits from
utilizing computers in decision-making processes.

Similarly, cardiologists have reported that:

Computers can aid in avoiding distorting influences to which
clinicians are subjected: (1) by placing the clinician's limited

personal experience into broader perspective through comparison
with a larger repository of clinically relevant information; (2) by
making explicit the assumptions implied by his or her decisions;
and (3) by alerting the clinician whenever the decisions made do
not appear consistent with these assumptions, with the available
information or with the conventional rules of logic.

100

Experiences relating to the prediction of deaths are clearly

applicable to the practice of neurosurgery in critical care

.units at the present, especially where trauma patients are

involved. Similarly, computerized predictions could be ap-
plicable to the indications for surgery, and for the selection
of pre- and postoperative care. While the institution of
computerized systems in decision making might be slower in
coming about and would obviously require updating almost
on a daily basis, it is clear that computers will play an
ever-increasing role in physicians' decision-making pro-

cesses in the future, and, equally important, they will play a
major part in the evaluation of the services that physicians
deliver.

It is imperative that physicians who will be practicing

medicine 5 to 10 years from now become thoroughly
acquainted with the computer's capabilities and that the
programs for their applications be developed.

REFERENCES

1. Hallett M, Cohen LG: Magnetism: A new method for stimu-

lation of nerve and brain. JAMA 262:538-541, 1989.

2. Editorial: Magnetoencephalography. Lancet 335:576-577,

1990.

3. Barker AT, Freeston IL, Jalinous R, Jarratt JA: Magnetic

stimulation of the human brain and peripheral nervous sys-
tem: An introduction and the results of an initial clinical
evaluation. Neurosurgery 20:100-109, 1987.

4. Maccabee PJ, Amassian VE, Cracco RQ, Cadwell JA: An

analysis of peripheral motor nerve stimulation in humans
using the magnetic coil. Electroencephalogr Clin Neurophy-
siol 70:524-533, 1988.

5. Barker AT, Jalinous R, Freeston IL: Noninvasive stimulation

of human motor cortex. Lancet 2:1106-1107, 1985.

6. Evans BA, Litchy WJ, Daube JR: The utility of magnetic

stimulation for routine peripheral nerve conduction studies.
Muscle Nerve 11:1074-1078, 1988.

7. Schriefer TN, Mills KR, Murray NMF, Hess CW: Evaluation