01 Where before what

Long before neurology had a single effective treatment, it had a method, and that method is still the most distinctive thing it offers the rest of medicine. A cardiologist can often reason straight from a symptom to its cause. A neurologist first has to ask where in a long, branching system something has gone wrong. The rule is to find the lesion before you name the disease.

The earliest surviving record of neurologic localization is an Egyptian surgical papyrus describing a man who could not answer questions after a wound to his temple (Brazis 2011, p. 1). In the nineteenth century, Broca and Wernicke matched what patients had lost in life with what their brains showed at autopsy, an approach that became the clinicopathologic method (Brazis 2011, pp. 1-2). Hughlings Jackson added the caution that a lesion shows you where a circuit can be broken, which is a different thing from where a faculty lives.

The four steps

Clinical diagnosis in neurology has four steps. You recognize that a function is impaired, decide which part of the nervous system is affected, list the likely causes, and only then use tests to choose among them (Brazis 2011, p. 2). The first step raises three questions, which are whether the deficit is real, whether it is serious enough to need a workup, and whether it is neurologic at all, since a child may limp because of a sore hip (Brazis 2011, p. 3). Be generous on the first question, and unless there are clear psychiatric features, take signs at face value (Brazis 2011, p. 3).

Why the scan comes second

Imaging ordered without a localization finds things, and whatever it finds tends to get blamed. Congenital brain cysts are easy to see and are often blamed for symptoms they did not cause, while the real disease goes untreated (Brazis 2011, p. 1). Many disease processes sit below the resolution of any scanner, so the images have to be read alongside the clinical picture, and the two sometimes disagree (Pearl 2014, p. 1). In children, an incidental cyst or a small bright spot on MRI is common and frightening to parents. A localization made before the scan is what lets you tell them with confidence that it is not the cause.

Finding a lesion is like finding a house. You can locate it by its street address, or by the intersection of two streets (Brazis 2011, p. 3). Either way you need a map with named stops, and this curriculum uses eleven of them. From the top down, they are the cerebral cortex, subcortex and deep gray matter, brainstem, cerebellum, spinal cord, anterior horn cell, nerve root, plexus, peripheral nerve, neuromuscular junction and muscle.

An intersection can pin a lesion down precisely. A third-nerve palsy on its own could sit anywhere along the nerve's course, from the brainstem to the orbit. Add a hemiparesis on the opposite side, and only one place carries both the third-nerve fibers and the corticospinal tract. That place is the cerebral peduncle of the midbrain, on the same side as the palsy (Brazis 2011, pp. 3-4).

Use the figure above as an elimination game. Each finding you switch on removes the levels that cannot produce it and tells you why. Weakness of the face, arm and leg removes the spinal cord at once, because the cord does not supply the face. A sensory level on the trunk removes every level except the cord. A finding earns its keep when it excludes a level, far more than when it merely fits one.

How precise each level can be

Localization is most precise low in the nervous system and least precise in the cortex (Brazis 2011, p. 4). A muscle lesion weakens that muscle, and a nerve lesion weakens its muscles and numbs its skin. A cortical lesion produces a mix of motor, sensory and behavioral findings that varies with the area involved. Low-level signs are also consistent, so a wasted hand muscle is there every time you look. Higher-level signs can come and go within a single examination. For language and other higher functions, the useful question is whether the patient performs a task consistently and normally each time it is tested (Brazis 2011, p. 4).

With children this matters, because a toddler's language, attention and praxis are still being built and have to be measured against age. The signs of motor unit disease already look much as they will in adulthood. The lower the level you are examining, the more you can trust a single clean sign.

Students tend to assume that a good history will tell them where the lesion is. For finding where, though, the examination matters far more, and a lesion can often be localized from the examination alone (Brazis 2011, p. 4). The examination includes the sensory findings a patient reports during it, so a complaint of numbness can be as objective as a wrist drop (Brazis 2011, p. 4).

The history contributes time, and the tempo suggests the kind of lesion. If a left hemiparesis arrived over minutes, think first of a vascular event or a seizure. If it evolved over a few days, think of infection or demyelination, and if it crept in over months, a tumor or a degenerative process (Brazis 2011, p. 4). Pairing where from the exam with when from the story is the backbone of this curriculum, and Module 11 builds it into a working differential.

Three questions at the bedside

By the end of every examination, you should be able to say where the lesion is, what the phenomenology is (the kind of disorder you are looking at), and what the likely cause is (Morris 2012, p. xxix). Commit to a diagnosis, with a short differential if you must. Listing every possible cause for fear of being wrong helps no one, and if you prove wrong you will have learned something (Morris 2012, p. xxix). Neurology relies on localization, or on recognizing the phenomenon, before it can reach a cause (Morris 2012, p. xxv). The middle question carries the disorders that have no focal lesion. A tremor, a seizure type or a tic is recognized by what it looks like.

In a child

Children add a second clock, so besides asking how long this has been happening, ask since when in development. A weakness present since early infancy, a skill that never arrived, and a skill that was there and has since been lost are three different stories even when today's examination looks the same, as the last section of this module shows.

Suppose two ten-year-olds have EEGs on the same afternoon, and both reports describe focal epileptiform discharges. The first has seizures that turn her eyes to the left and jerk her left face and hand, the signature of a right frontal focus, and that is where her discharges sit. The second has headaches and a normal examination, and his EEG was ordered for thoroughness. Only in the first child does the report confirm a prediction.

A test moves a belief by an amount that depends on where the belief started. The rule, named for the 18th-century English minister Thomas Bayes, is used daily in the form pretest odds × likelihood ratio = post-test odds. A test is best thought of as an "opinion modifier" (Gill 2005). An identical report means different things in different children because it lands on different priors.

The exam sets the prior

Every localization is a prediction, and a finding at the predicted address inherits the weight of the exam, and a finding at an address the exam excluded starts near zero. Clinical diagnosis should come before ancillary tests, since only careful localization avoids the false positives of "gunshot" imaging (Brazis 2011, p. 1).

What healthy children's tests show

When the scan and the exam disagree

A child weak from disease of the nerves or the neuromuscular junction has a normal brain MRI, because much disease lies outside the brain or below the resolution of imaging (Pearl 2014, p. 1). An abnormal scan explains the symptoms only if the exam put the lesion there.

Use the figure above by choosing a test and then a prior. The false-positive rates are sourced and the priors illustrative.

Faced with several abnormal findings, the localizer's first instinct is parsimony, which means looking for one lesion that explains them all. The instinct is usually right, and it drives the intersection method. A third-nerve palsy and a hemiparesis are far more likely to come from one midbrain lesion than from two unrelated ones.

Parsimony is still only a strategy, and nature doesn't promise to follow it. Hickam's dictum, usually attributed to the physician John Hickam, holds that patients can have as many diseases as they please. Saint's triad records the same observation, that some patients genuinely carry several unrelated conditions at once (Hilliard 2004). The skill lies in recognizing when to stop forcing a single explanation onto findings that resist it.

Did the findings begin together?

Timing gives the cleanest test, since a third-nerve palsy and a hemiparesis that appear together point to the midbrain. If the palsy dates from childhood and the hemiparesis appeared decades later, the examination is probably showing two lesions that arrived at different times (Brazis 2011, p. 4). Before you merge findings into one lesion, ask whether they began together.

Discrete lesions and system lesions

Most localization teaching is built on discrete lesions, such as an infarct, a demyelinating plaque, a tumor or an abscess, each of which damages everything in one place. Other diseases attack a functional system instead, an array of related neurons scattered across several levels (Brazis 2011, p. 4). Parkinson disease is a system disorder that still localizes neatly to the substantia nigra, whereas a spinocerebellar degeneration has no single address. When the findings refuse to settle at one level, consider in turn several discrete lesions (as in demyelinating disease), a diffuse toxic or metabolic process, and a system degeneration. Module 10 takes up the case where the answer is "everywhere".

A child with a static motor disability from early life can acquire a new and unrelated problem. The old findings have to be subtracted before the new ones can be localized, which makes the parents' account of the child's baseline part of the localization itself.

An adult's nervous system holds still while you examine it, but a child's keeps changing. That makes development the fourth dimension of pediatric localization. The same finding can be normal at one age and a sign of disease at another, and a fixed lesion can produce signs that change as the brain around it grows.

When age changes the meaning

The upgoing big toe is a cornerstone sign of corticospinal tract damage, yet it can be normal in the first year of life, while the corticospinal tracts are still being myelinated (Utah PediNeuroLogic Exam). The closer an infant is to one year, the more consistently the toe goes down (DeMyer, p. 282). Handedness runs the other way, since a hand preference before 12 months is abnormal and points to weakness of the other hand (Utah PediNeuroLogic Exam). The baby who reaches only with the right hand may be telling you about the left.

Fixed lesions with changing signs

Cerebral palsy is defined as a group of permanent disorders of movement and posture attributed to non-progressive disturbances in the developing fetal or infant brain (Rosenbaum 2007). The clinical picture changes as the child grows, and infants with cerebral palsy may look hypotonic before spasticity appears (DeMyer, p. 273). A floppy infant who becomes a stiff toddler need not have a progressive disease. The maturing motor system has revealed a lesion that was there all along.

Localization in time

In children you localize along a time axis as well as a spatial one, and part of the work is deciding which of three stories the child is in.

No single examination can tell these three apart. The history, the developmental record and repeated examinations can, which is why in child neurology the second visit is often more informative than the first.