Chapter 9. Other Somatic Sensation

Sensory and motor functions are interdependent, as was dramatically illustrated by the early animal experiments of Claude Bernard and Charles Sherrington, in which practically all effective movement of a limb was abolished by eliminating only its sensory innervation (sectioning of posterior roots). Interruption of other sensory pathways and destruction of the parietal cortex also has profound effects on motility. To a large extent, human motor activity depends on a constant influx of sensory impulses (most of them not consciously perceived). Sensory motor integration is therefore necessary for normal nervous system function but disease may affect motor or sensory functions independently. There may be loss or impairment of sensory function, and this can represent the principal manifestation of neurologic disease.

This chapter deals with general somatic sensation, i.e., afferent impulses that arise in the skin, muscles, or joints. One form of somatic sensation—pain—was discussed in Chap. 8. Because of its overriding clinical importance, pain has been accorded a chapter of its own, but that chapter and this chapter are of one piece. The special senses—vision, hearing, taste, and smell—are considered in the next section (Chaps. 12, 13, 14, and 15), and visceral sensation, most of which does not reach consciousness, is considered with the disorders of the autonomic nervous system (Chap. 26).

An understanding of sensory disorders depends upon knowledge of functional anatomy. It is necessary to be familiar with the sensory receptors in the skin and deeper structures, the distribution of the peripheral nerves and roots, and the pathways by which sensory impulses are conveyed from the periphery and through the spinal cord and brainstem to the thalamus and cerebral cortex. These aspects of sensory anatomy and physiology were touched upon in Chap. 8 in relation to the perception of pain and are elaborated upon here to include all forms of somatic sensation. An appropriate starting point is Fig. 9-1 that shows the cutaneous distribution of the peripheral nerves.

All sensation depends on impulses that are excited by stimulation of receptors and conveyed to the central nervous system by afferent (sensory) fibers. Sensory receptors are of two general types: those in the skin, mediating superficial sensation (exteroceptors), and those in the deeper somatic structures (proprioceptors). Skin receptors are particularly numerous and transduce four types of sensory experience: warmth, cold, touch, and pain; these are conventionally referred to as sensations or senses, e.g., tactile sensation or sense of touch. Proprioceptors inform us of the position of our body or parts of our body; of the force, direction, and range of movement of the joints (kinesthetic sense); and a sense of pressure, both painful and painless. Histologically, a wide variety of sensory receptors have been described, varying from simple, free dendrite terminals to highly branched and encapsulated structures, the latter bearing the names of the anatomists who first described them (see below). These are called "dendrites" because the direction of flow of physiologic activity and of sensory information from these structures in the periphery is toward the cell body.

Mechanisms of Cutaneous Sensation

As indicated in the preceding chapter, it had been thought that each of the primary modalities of cutaneous sensation is subserved by a morphologically distinct end organ, each with its separate peripheral nerve fibers. According to this formulation, postulated by von Frey, each type of end organ was thought to respond only to a particular type of stimulus and to subserve a specific modality of sensation: Meissner corpuscles (named after Georg Meissner), touch; Merkel discs (named after Friedrich Sigmund Merkel), pressure; Ruffini plumes (names after Angelo Ruffini), warmth and skin stretch; Krause end bulbs (named after Wilhelm Krause), cold; Pacini (pacinian) corpuscles (named after Filippo Pacini), vibration and tickle; and for pain, nerve endings that not associated with transducer ("free nerve endings"). The last of these are also termed "naked" because they are surrounded by Schwann cells but are not myelinated (Fig. 9-2).

This specificity theory, as it came to be called, has held up best in respect to the peripheral mechanisms for pain, insofar as certain primary afferent fibers, namely the C and A-δ fibers and their free nerve ending receptors, respond maximally to noxious stimuli. Even these freely branching receptor endings and their pain fibers convey considerable non-noxious information; that is, their specificity as pain fibers is not absolute (Chap. 8). Nor has it been possible to ascribe a specific function to each of the many other types of receptors. Thus, Merkel discs and Meissner corpuscles within nerve plexuses around the hair follicles, and free nerve endings can all be activated by moving or stationary tactile stimuli. Conversely, a single type of receptor seems capable of generating more than one sensory modality. Lele and Weddell found that with appropriate stimulation of the cornea, each of the four primary modalities of somatic sensibility (touch, warmth, cold, pain) could be recognized, even though the cornea contains only fine, free nerve endings. In the outer ear, which is also sensitive to these four modalities, only two types of receptors—freely ending and perifollicular—are present. The lack of organized receptors—e.g., the end bulbs of Krause and Ruffini—in the cornea and ear makes it evident that these types of receptors are not essential for the recognition of cold and warmth as von Frey and other early anatomists had postulated.

Particularly instructive have been the observations of Kibler and Nathan, who studied the responses of warm and cold spots to different stimuli. (Warm and cold spots are those small areas of skin that respond most consistently to thermal stimuli with a sensation of warmth or cold.) They found that a cold stimulus applied to a warm spot gave rise to a sensation of cold and that a noxious stimulus applied to a warm or cold spot gave rise only to a painful sensation; they also noted that mechanical stimulation of these spots gave rise to a sensation of touch or pressure. These observations indicate that cutaneous receptors, some not easily distinguishable from one another on morphologic grounds, are probably endowed with only a relative degree of specificity, in the sense that each responds preferentially (i.e., has a lower threshold) to one particular form of stimulation. However, among the freely branching nociceptors there is some degree of specialization, as discussed in Chap. 8. Such end organs can then be classed as mechanoreceptors, thermoreceptors, or nociceptors, depending on their selective sensitivity to mechanical, thermal, or noxious stimuli respectively (Sinclair; Light and Perl).

Physiologic studies have shown that the quality of sensation depends on the type of nerve fiber that is activated. Different diseases therefore evoke a variety of sensory symptoms. Microstimulation of single sensory fibers in a peripheral nerve of an awake human subject arouses different sensations, depending on which fibers are stimulated and not on the frequency of stimulation. On the other hand, the frequency of stimulation governs the intensity of sensation, as does the number of sensory units that are stimulated. Stated somewhat differently, afferent impulse frequency (temporal summation) is encoded by the brain in terms of magnitude or intensity of sensation. In addition, as the intensity of stimulation increases, more sensory units are activated (spatial summation).

Localization of a stimulus was formerly thought to depend on the simultaneous activation of overlapping sensory units. Tower defined a peripheral sensory unit as a dorsal root ganglion cell, its central and distal processes, and all the sensory endings in the territory supplied by those distal processes (the receptive field of the sensory cell). In the very sensitive pulp of the finger, where the error of localization is less than 1 mm, there are 240 overlapping, low-threshold mechanoreceptors per square centimeter. Highly refined physiologic techniques have demonstrated that activation of even a single sensory unit is sufficient to localize the point stimulated and that the body map in the parietal lobe is capable, by its modular columnar organization, of encoding such refined topographic information. Also, each point in the skin that is stimulated may involve more than one type of receptor. To gain access to its receptor, a stimulus must pass through the skin and possess sufficient energy to transduce, i.e., depolarize, the nerve ending.

Not only does the threshold of stimulation vary but the nerve impulse that is generated is a graded one, not an all-or-none phenomenon like an action potential in nerve. This poorly understood peripheral generator potential determines the freqsuency of impulses in the nerve and to what degree it is sustained or fades out (i.e., adapts to the stimulus or fatigues). However, the mechanism by which a stimulus is translated into a sensory experience, i.e., is encoded, is now partially understood. Special molecules transduce physical alterations by opening cationic channels. As mentioned in Chap. 8, these changes lead to the opening of voltage-gated sodium channels, which generates an action potential. It is probably fair to say that each type of specialized ending has a membrane structure that facilitates the transduction process for a particular type of stimulus. In general, the encapsulated endings, which are more highly myelinated, are of low-threshold type, variably adaptable to continued stimulation (Meissner and pacinian corpuscles are rapidly adapting; Merkel discs and Ruffini endings are slowly adapting) and are connected to large sensory fibers (see Lindblom and Ochoa). The pacinian receptors are the most deeply situated (Fig. 9-1).

Sensory Pathways

Sensory Nerves

Fibers that mediate superficial sensation are located in cutaneous sensory or mixed sensorimotor nerves. In cutaneous nerves, unmyelinated pain and autonomic fibers exceed myelinated fibers by a ratio of 3 or 4:1. The myelinated fibers are of two types: small, lightly myelinated, A-δ fibers for pain and cold, as discussed in Chap. 8, and larger, faster-conducting A-α fibers for touch and pressure. Nonmyelinated autonomic fibers are efferent (postganglionic) and innervate piloerector muscles, sweat glands, and blood vessels. The differing conduction velocities of the fibers of these are discussed in Chap. 45.

A single cutaneous (touch, pain, and temperature) type of afferent fiber connects to several receptors, all of one type, which are irregularly distributed in the skin and account for sensory spots. Stimuli from a given spot are therefore conveyed by two or more fibers. The proprioceptive fibers subserve pressure sense and, with endings in articular structures, the sense of position and movement; they enable one to discriminate the form, size, texture, and weight of objects. Sensations of tickle, itch, and wetness are believed to arise from combinations of several types of receptors.

Itch is a distinctive sensation that can be separated on clinical and neurophysiologic grounds from touch and from pain. Two aspects are recognized: a brief primary localized itch at the site of stimulus or injury, and a subsequent more diffuse sensation that is greatly intensified by gentle touch. Itch sensation is transmitted by specific C fibers, not by touch mechanisms, for regions of analgesia no longer can be stimulated to itch but areas anesthesia retain this sensation. There is, however, no specialized peripheral itch receptor, the sensation depending instead on the spinal connections to itch pathways. Several forebrain regions are activated by itch, i.e., there is no central "itch center." From the perspective of neurologic diseases, the main cause of itch is postherpetic neuralgia, but cases of paroxysmal itch of central origin are known to be caused by multiple sclerosis. The pathophysiology of itching has been discussed by Greaves and Wall and is reviewed further by Yosipovitch and colleagues.

Spinal Roots

All the sensory neurons have their cell bodies in the dorsal root ganglia. The peripheral extensions of these cells constitute the sensory nerves; the central projections of these same cells form the posterior (dorsal) roots and enter the spinal cord. Each dorsal root contains all the fibers from skin, muscles, connective tissue, ligaments, tendons, joints, bones, and viscera that lie within the distribution of a single body segment (somite). This segmental innervation has been amply demonstrated in humans and animals by observing the effects of lesions that involve one or two spinal nerves, such as (1) herpes zoster, which also causes visible vesicles in the corresponding area of skin; (2) the effects of a prolapsed intervertebral disc, which causes hypalgesia in a single root zone; and (3) surgical section of several roots on each side of an intact root (method of remaining sensibility). Maps of the dermatomes derived from these several types of data are shown in Figs. 9-3 and 9-4. It should be noted that there is considerable overlap from one dermatomal segment to the other, more so for touch than for pain. By contrast, there is less overlap between adjacent peripheral nerves and almost none between the divisions of the trigeminal nerve. Also, the maps differ somewhat according to the methods used in constructing them. In contrast to most dermatomal charts, those of Keegan and Garrett (based on the injection of local anesthetic into single dorsal root ganglia) show bands of hypalgesia to be continuous longitudinally from the periphery to the spine (Fig. 9-4). The distribution of pain fibers from deep structures, although not exactly corresponding to that of pain fibers from the skin, also follows a segmental pattern. In Chap. 8 it was commented that the areas of projection of referred pain from the visceral organs and musculoskeletal structures roughly correspond to the overlying dermatomes but they have distinctive patterns, termed sclerotomes (Fig. 8-5).

Posterior Root Entry Zone, Dorsal Horns and Posterior Columns

In the dorsal roots, the sensory fibers are first rearranged according to function. Large and heavily myelinated fibers enter the cord just medial to the dorsal horn and divide into ascending and descending branches. The descending fibers and some of the ascending ones enter the gray matter of the dorsal horn within a few segments of their entrance and synapse with nerve cells in the posterior horns as well as with large ventral horn cells that subserve segmental reflexes. Some of the ascending fibers run uninterruptedly in the dorsal columns of the same side of the spinal cord, terminating in the gracile and cuneate nuclei in the upper cervical spinal cord and medulla (Fig. 9-5). The central axons of the primary sensory neurons are joined in the posterior columns by other secondary neurons whose cell bodies lie in the posterior horns of the spinal cord (see below). The fibers in the posterior columns assume a medial position as new fibers from each successively higher root are added laterally, thereby creating somatotopic laminations (see Fig. 8-3).

Of the long ascending posterior column fibers, which are activated by mechanical stimuli of skin and subcutaneous tissues and by movement of joints, only about 25 percent (from the lumbar region) reach the gracile nuclei at the upper cervical cord. The remaining fibers send collaterals to, or terminate in, the dorsal horns of the spinal cord, at least in the cat (Davidoff). An estimated 20 percent of ascending fibers originate from cells in Rexed layers IV and V of the posterior horns (see Fig. 8-1) and convey impulses from low-threshold mechanoreceptors that are sensitive to hair movement, skin pressure, or noxious stimuli. There are also descending fibers in the posterior columns, including fibers from cells in the dorsal column nuclei.

The posterior columns contain a portion of the fibers for the sense of touch as well as the fibers mediating the senses of touch-pressure, vibration, direction of movement and position of joints, and stereoesthesia—recognition of surface texture, shape, numbers, and figures written on the skin and two-point discrimination—all of which depend on patterns of touch-pressure (see Fig. 8-2). The nerve cells of the nuclei gracilis and cuneatus and accessory cuneate nuclei give rise to a secondary afferent path, which crosses the midline in the medulla and ascends as the medial lemniscus to the posterior thalamus. However, the fiber pathways in the posterior columns are not the sole mediators of proprioception in the spinal cord (see "Posterior [Dorsal] Column Syndrome" further on).

In addition to the well-defined posterior column pathways, there are cells in the more loosely structured "reticular" part of the dorsal column that receive secondary ascending fibers from the dorsal horns of the spinal cord and from ascending fibers in the posterolateral columns. These dorsal column fibers project to brainstem nuclei, cerebellum, and thalamic nuclei. Many other cells of the dorsal horn nuclei are interneurons, with both excitatory and inhibitory effects on local reflexes or on the primary ascending neurons. They are also under the influence of the sensorimotor cortex. Some act on descending corticospinal motor neurons. The functions of many of the extrathalamic projections of dorsal column cells are unknown (Davidoff).

Thinly myelinated or unmyelinated fibers, subserving mainly pain sensibility, but some sensitive to touch and pressure, enter the cord on the lateral aspect of the dorsal horn and synapse with dorsal horn cells, mainly within a segment or two of their point of entry into the cord. The dorsal horn cells, in turn, give rise to secondary sensory fibers, some of which may ascend ipsilaterally but most of which decussate and ascend in the spinothalamic tracts, as described in Chap. 8 (see Figs. 8-1 and 8-2). Observations based on surgical interruption of the anterolateral funiculus indicate that fibers mediating touch and deep pressure occupy the ventromedial part (anterior spinothalamic tract). Also as remarked in Chap. 8, an ascending tract of secondary sensory axons lies in or medial to the descending corticospinal system.

After the posterior columns terminate in the gracile and cuneate nuclei of the rostral cervical cord and medulla, synapses are made with fibers that cross the midline and ascend to form the medial lemniscal tracts in the brainstem. The lemniscal system is situated in a paramedian position, changing orientation slightly at different levels of the brainstem, and joining the spinothalamic system in the rostral midbrain to terminate in the posterior thalamic nuclei (see below and also Fig. 8-2).

Trigeminal Connections

The pathways mediating cutaneous sensation from the face and head—especially touch, pain, and temperature—are conveyed to the brainstem by the trigeminal nerve. After entering the pons, the pain and temperature fibers turn caudally and run through the ipsilateral medulla as the descending spinal trigeminal tract, terminating in the long, vertically oriented nucleus caudalis, or spinal trigeminal nucleus, that lies beside it and extends to the second or third cervical segment of the cord, where it becomes continuous with the posterior horn of the spinal gray matter. Axons from the neurons of this nucleus cross the midline and ascend as the trigeminal quintothalamic tract (also termed, somewhat imprecisely, trigeminal lemniscus) along the medial side of the spinothalamic tract (see Fig. 8-2), of which it is the equivalent.

Thalamocortical Connections

The ventral posterior thalamic nucleus receives fibers from the medial lemniscal, spinothalamic, and trigeminal (fibers from the principal sensory and spinal trigeminal nuclei) tracts, and projects mainly to two somatosensory cortical areas. The first area (S1) corresponds to the post-central cortex or Brodmann areas 3, 1, and 2. S1 afferents are derived primarily from the ventral posterolateral nucleus (VPL, the terminus of medial lemniscal and spinothalamic fibers) and the ventral posteromedial nucleus (VPM, the terminus of trigeminal fibers) and are distributed somatotopically, with the leg represented uppermost and the face lowermost (face and hand are juxtaposed). Electrical stimulation of this area yields sensations of tingling, numbness, and warmth in specific regions on the opposite side of the body. The information transmitted to S1 is tactile and proprioceptive, derived mainly from the dorsal column–medial lemniscus system and concerned mainly with sensory discrimination. The second somatosensory area (S2) lies on the upper bank of the sylvian fissure, adjacent to the insula. Localization of function is less discrete in S2 than in S1, but S2 is also organized somatotopically, with the face rostrally and the leg caudally. The sensations evoked by electrical stimulation of S2 are much the same as those of S1 but, in distinction to the latter, may be felt bilaterally. Of note is the failure of cortical stimulation to evoke pain sensation.

Undoubtedly, the perception of sensory stimuli involves more of the cerebral cortex than the two discrete areas described above. Some sensory fibers probably project to the precentral gyrus and others to the superior parietal lobule. Moreover, S1 and S2 are not purely sensory in function; motor effects can be obtained by stimulating them electrically. It has been shown that sensory neurons in VPL, cuneate and gracile nuclei, and sensory neurons in the dorsal horns of the spinal cord all receive descending as well as ascending cortical projections. This reciprocal arrangement probably influences movement and the transmission and interpretation of pain, as discussed in Chap. 8.

Provided that the subcortical structures, especially the thalamus, are intact, certain sensations such as pain, touch, pressure, and extremes of temperature can reach consciousness. Their accurate localization, however, as well as the patient's ability to make other fine sensory discriminations, depends to a large extent on the integrity of the sensory cortex. This clinical distinction is elaborated in the discussion of the sensory syndromes further on.

Figure 9-6 ("sensory homunculus") shows the cortical representation of sensory information in the postcentral gyrus. As in the case of the motor representation (see Fig. 3-4), a disproportionate area is devoted to localization in the fingers, lips, and face.

From this brief account of the various channels of sensory information, one must conclude that at every level there is the possibility of feedback control from higher levels. Most external and some internal stimuli are highly complex and induce activity in more than one sensory system. In every system there is sufficient redundancy to allow lesser-used systems to compensate partially for the deficits incurred by disease.

Most neurologists would agree that sensory testing is the most difficult part of the neurologic examination. For one thing, test procedures are relatively crude and are unlike those natural modes of stimulation with which the patient is familiar. It is also fair to say that few diagnoses are made solely on the basis of the sensory examination; more often the exercise serves to complement the motor examination. Quite often, no objective sensory loss can be demonstrated despite symptoms that suggest the presence of such an abnormality. Only rarely does the opposite pertain, i.e., one discovers a sensory deficit when there has been no complaint of sensory symptoms. Sensory symptoms in the nature of paresthesia or dysesthesia may be generated along axons of nerves not sufficiently diseased to impair or reduce sensory function; in the latter instance, loss of function may have been so mild and gradual as to pass unnoticed. Always there is some difficulty in evaluating the response to sensory stimuli, as it depends on the patient's interpretation of sensory experiences. This, in turn, will depend on the patient's general awareness and responsiveness and ability to cooperate, as well as degree of suggestibility. Children, by virtue of their simple and direct responses, are often better witnesses than more sophisticated individuals who are likely to analyze their feelings minutely and report small and insignificant differences in stimulus intensity.

General Considerations

Before proceeding to sensory testing, the physician should question patients about their symptoms; this, too, poses particular problems. Patients may be confronted with derangements of sensation that are unlike anything they have previously experienced, and they have limited terms to describe what they feel. They may say that a limb feels "numb" and "dead" when in fact they mean that it is weak. Occasionally a loss of sensation is discovered almost accidentally, e.g., by a lack of pain on touching an object hot enough to blister the skin or unawareness of articles of clothing and other objects in contact with the skin. More often, disease induces new and unnatural sensory experiences such as a band of tightness, a feeling of the feet being encased in cement, lancinating pains, an unnatural feeling when stroking the skin, a sensation as if walking on pebbles, and so on.

If nerves, sensory roots, or spinal tracts are damaged or partially interrupted, the patient may complain of tingling or prickling feelings ("like Novocain" or like the feelings in a limb that has "fallen asleep," the common colloquialism for nerve compression), cramp-like sensations, or burning or cutting pain occurring either spontaneously or in response to stimulation.

Experimental data support the view that partially damaged touch, pressure, thermal, and pain fibers become hyperexcitable and generate ectopic impulses along their course, either spontaneously or in response to a natural volley of stimulus-evoked impulses (Ochoa and Torebjork). These abnormal sensations are experienced as paresthesias, or dysesthesias if they are severe and distressing, as noted in Table 8-2. Another positive sensory symptom is allodynia, referring to a phenomenon in which one type of stimulus evokes another type of sensation—e.g., touch may induce pain.

The clinical description of a sensation may divulge the particular sensory fibers involved (Table 9-1). It is known that stimulation of touch fibers gives rise to a sensation of tingling and buzzing; of muscle proprioceptors, to pseudocramp (the sensation of cramping without actual muscle contraction); of thermal fibers, to heat (including burning) and coldness; and of A-δ fibers, to prickling and pain. Paresthesia arising from ectopic discharges in large sensory fibers can be induced by nerve compression, hypocalcemia, hypomagnesemia, certain medications (niacin foremost among them), and diverse diseases of nerves. Band-like sensations on a limb or the trunk are the result of dysfunction in large sensory fibers, either in the periphery or their continuation in the posterior columns. Certain sensory symptoms suggest an anatomic location of nerve disease; for example, lancinating pains that radiate to the back or neck implicate root or, less often, sensory ganglion disease.

The presence of persistent paresthesiae incriminates a lesion involving sensory pathways in nerves, spinal cord, or higher structures. Most often, the large fibers in the peripheral nerves or posterior columns are involved. Evanescent paresthesia, of course, is usually of no significance. Every person has had the experience of resting a limb on the ulnar, sciatic, or peroneal nerve and having the extremity "fall asleep." This is because of compressive interruption of axonal transport and not of ischemia of the nerves or other components of the limb as is commonly assumed. The hyperventilation of anxiety may cause paresthesia of the lips and hands (sometimes unilateral) from diminution of CO2 and thereby of ionized calcium; tetany may also occur, with carpopedal spasms. However, these sensory experiences are transient and should not be confused with the persistent, albeit frequently fluctuating, paresthesia of structural disease of the nervous system. Severe acral and peripheral paresthesias with perversion of hot and cold sensations are characteristic of certain neurotoxic shellfish poisonings (ciguatera) and some toxins, such as mercury.

Also worth comment are vibratory paresthesias, which we have encountered in only a handful of patients. One articulate physician described the sensation as a high-amplitude, low-frequency "buzz" that was distinctly different from the more common prickling paresthesia, burning, numbness, etc. We have the impression that these sensations are almost always a manifestation of central sensory disease, in one case probably attributable to the posterior columns and in another to cerebral disease. Beyond this, little is known about this symptom.

Effect of Age on Sensory Function

A matter of importance in the testing of sensation is the progressive impairment of sensory perception that occurs with advancing age. This requires that sensory thresholds, particularly in the feet and legs, be assessed in relation to age standards. The effect of aging is most evident in relation to vibratory sense, but proprioception, the perception of touch, and fast pain are also diminished with age. Sweating and vasomotor reflexes are reduced as well. These changes, which are discussed further in Chap. 29, are probably caused by neuronal loss in dorsal root ganglia and are reflected in a progressive depletion of fibers in the posterior columns. Receptors in the skin and special sense organs (taste, smell) also wither with age.

Terminology of Sensory Signs and Symptoms

(See also Table 8-2)

A few additional terms require definition, as they are encountered in discussions of sensation. Some of these, relating to pain sensation, were mentioned in Chap. 8. Anesthesia refers to a complete loss and hypesthesia to a partial loss of all forms of sensation. Loss or impairment of specific cutaneous sensations may be indicated by an appropriate prefix or suffix, e.g., thermoanesthesia or thermohypesthesia (loss and reduction in temperature sense), analgesia (loss of pain sense), hypalgesia (reduction in pain sensibility), tactile anesthesia, and pallanesthesia, or apallesthesia (loss of vibratory sense). The term hyperesthesia, as explained in Chap. 8, refers to an increased sensitivity to various stimuli and is usually used with respect to cutaneous sensation. It implies a heightened activity of the sensory apparatus. Under certain conditions (e.g., sunburn), there does appear to be an enhanced sensitivity of cutaneous receptors, but usually the presence of hyperesthesia betrays an underlying sensory defect. Careful testing will demonstrate an elevated threshold to tactile, painful, or thermal stimuli; but once the stimulus is perceived, it may have a severely painful or unpleasant quality (hyperpathia). Some clinicians use this last term to denote an exaggerated response to a painful stimulus (hyperalgesia which is subtly different from hyperpathia, denotes an abnormally painful reaction to a painful stimulus; Table 8-2). In alloesthesia, or allesthesia, a tactile or painful stimulus delivered on the side of hemisensory loss is experienced in a corresponding area of the opposite side or at a distant site on the same side. This phenomenon is observed most frequently with right-sided putaminal lesions (usually hemorrhage) and with anterolateral lesions of the cervical spinal cord; it presumably depends on the existence of an uncrossed ipsilateral spinothalamic pathway (see the original studies of Ray and Wolff).

Testing of Sensory Function

The detail with which sensation is tested is determined by the clinical situation. If the patient has no sensory complaints, it is sufficient to test vibration and position sense in the fingers and toes and the perception of pinprick over the extremities, and to determine whether the findings are the same in symmetrical parts of the body. A rough survey of this sort occasionally detects sensory defects of which the patient was unaware. More thorough testing is in order if the patient has complaints referable to the sensory system or if one finds localized atrophy or weakness, ataxia, trophic changes of joints, or painless ulcers.

A few other general rules are useful. It is easier for a patient to perceive the boundary of an abnormal area of sensation if the examiner proceeds from an area of reduced sensation toward the normal area. One should not press the sensory examination in the presence of fatigue, for an inattentive patient is a poor witness. Also, the examiner must avoid suggesting symptoms to the patient. After explaining in the simplest terms what is required, the examiner interposes as few questions and remarks as possible. Patients should generally not be asked, "Do you feel that?" each time they are touched; they are simply told to say "yes" or "sharp" every time they are touched or feel pain. Repetitive pinpricks within a small area of skin should be avoided, as this will make inapparent a subtle hypalgesia, because of the phenomenon of temporal summation as discussed earlier. In patients who may be overinterpreting slight changes in pinprick, differentiating between warm and cold is often more informative than differentiating between "sharp" and "dull." The patient should not directly observe the part under examination. A cooperative patient may, if asked to use a pin or his fingertips, outline an analgesic or anesthetic area or determine whether there is a graduated loss of sensation in the distal parts of a leg or arm. Finally, the findings of the sensory examination should be accurately recorded in narrative or on a chart by shading affected regions on a preprinted figure of the body or a sketch of a hand, foot, face, or limb.

Described below are the usual bedside methods of testing sensory function. These tests are sufficient for most clinical purposes. For clinical investigation and research into the physiology of pain, which require the detection of threshold values and quantification of sensory impairment, a wide range of instruments are available. Their use has been described by Dyck and colleagues (see also Lindblom and Ochoa and Bertelsmann et al).

Testing of Tactile Sensation

This is usually done with a wisp of cotton or light touch of a finger. The patient is first acquainted with the stimulus by applying it to a normal part of the body. Then, with eyes closed, he is asked to indicate if the sensation feels "natural" or to say "yes" each time various other parts are touched. A patient who is simulating sensory loss may say no in response to a tactile stimulus. Cornified areas of skin, such as the soles and palms, will require a heavier stimulus than other areas and the hair-clad parts a lighter one because of the numerous nerve endings around the follicles. Patients will be more sensitive to a moving stimulus of any kind than to a stationary one. The deft application of the examiner's or the patient's roving fingertips is a useful refinement and aids in demarcating an area of tactile loss, as Trotter and Davies originally showed.

More precise testing is possible by using a von Frey hair. By this method, a stimulus of constant strength can be applied and the threshold for tactile sensation determined by measuring the force required to bend a hair of known length. Spurious areas of hypesthesia may arise when a series of contactual stimuli lead to a decrement of sensation, either through adaptation of the end organ or because the initial sensation outlasts the stimulus and seems to spread.

Testing of Pain Perception

This is most efficiently assessed by pinprick, although it may be evoked by a variety of noxious stimuli. Patients must understand that they are to report the feeling of sharpness, not simply the feeling of contact or pressure of the pinpoint. If pinpricks are applied rapidly in one area, their effect may be summated and a heightened perception of pain may result; therefore they should be delivered about one per second and not over the same spot. Small differences in intensity can be discounted. An effective approach is to ask the patient to compare the pin sensation in two areas on a scale of 1 to 10; a report of "8 or 9" as compared to "10" is usually insignificant.

It is almost impossible, using an ordinary pin, to apply each stimulus with equal intensity. A pinwheel (Wartenberg wheel) is sometimes more effective because it allows the application of a more constant pressure, but risk of transmission of blood-borne infection from patient to patient has made this method outdated. This difficulty can be overcome to some extent by the use of an algesimeter, which delivers stimuli of constant intensity. Quantification of small-fiber sensation can be better assessed in the clinical laboratory by the use of thermal stimuli delivered by a computerized device as described below.

If an area of diminished or absent touch or pain sensation is encountered, its boundaries should be demarcated to determine whether it has a segmental or peripheral nerve distribution or is lost below a certain level on the trunk. As mentioned, such areas are best delineated by proceeding from the region of impaired sensation toward the normal. The changes may be confirmed by dragging a pin lightly over the parts in question. Areas of reduced pinprick sensation can be profitably corroborated by the thermal sense examination.

Testing of Deep-Pressure Pain

One can estimate the perception of this modality simply by lightly pinching or pressing deeply on the tendons, muscles, or bony prominences. Pain can often be elicited by heavy pressure even when superficial sensation is diminished; conversely, in some diseases, such as tabetic neurosyphilis, loss of deep-pressure pain may be more striking than loss of superficial pain. This uncomfortable examination can be omitted in routine neurological cases.

Testing of Thermal Sense

A quick but rough way to assess thermal loss (or to corroborate a previously found zone of hypalgesia) is to warm one side of a tuning fork by rubbing it briskly against the palm and apply its alternate sides to the patient's skin and asking the patient which side is colder (or warmer). This suffices for most bedside examinations.

If more careful examination is required, the skin should first be exposed to room air for a brief time before the examination. The test objects should be large, ideally two stoppered test tubes containing hot (45°C/113°F) and cold (20°C/68°F) tap water with thermometers that extend into the water through the flask stoppers. Extreme degrees of heat and cold may be employed first to delineate roughly an area of thermal sensory impairment. The side of each tube is applied successively to the skin for a few seconds and the patient is asked to report whether the flask feels "less hot" or "less cold" in comparison to a normal part. The qualitative change can then be quantitated as far as possible by recording the differences in temperature that the patient is able to recognize as the difference in temperature between the two tubes is gradually reduced. A normal person can detect a difference of 1°C (1.8°F) or even less in the range of 28°C (82.4°F) to 32°C (89.6°F); in the warm range, differences between 35°C (95°F) and 40°C (104°F) can be recognized, and in the cold range, between 10°C (50°F) and 20°C (68°F). If the temperature of the test object is below 10°C (50°F) or above 50°C (122°F), sensations of cold or heat become confused with pain. This technique has been largely supplanted by commercially manufactured electronic devices that can present a series of slightly differing thermal stimuli in sequence to a probe placed on the finger or toe. Special algorithms are used for the order and magnitude of temperature change and to determine whether the patient's reports are consistent and valid. The results are reported in the form of a "just noticeable difference" (JND) between temperatures or intensities of pain.

Testing of Proprioceptive Sense

Awareness of the position and movements of our limbs, fingers, and toes is derived from receptors in the muscles, tendons (Golgi tendon organs, according to Roland and Ladegaard-Pederson) and joints and is probably facilitated by the activation of skin receptors (Moberg). The two modalities comprising proprioception, i.e., sense of movement and of position, are usually both impaired, although clinical situations do arise in which perception of the position of a limb or digits is lost while that of passive and active movement (kinesthesia) of these parts is retained. The opposite occurs but is infrequent.

Abnormalities of position sense may be disclosed in several ways. When the patient has his arms outstretched and eyes closed, the affected arm will wander from its original position; if the patient's fingers are spread apart, they may undergo a series of changing postures ("piano-playing" movements, or pseudoathetosis); in attempting to touch the tip of his nose with his index finger, the patient may miss the target repeatedly, but the performance is corrected when the eyes are open.

Perception of passive movement is first tested in the fingers and toes as the defect, when present, is reflected maximally in these parts. It is important to grasp the digit at the sides, perpendicular to the plane of movement; otherwise the pressure applied by the examiner may allow the patient to identify the direction of movement. This applies as well to testing of the more proximal segments of the limb. The patient should be instructed to report each movement as being "up" or "down" from the previous position (directional kinesthesia). It is useful to demonstrate the test with a large and easily identified movement, but once the idea is clear to the patient, the smallest detectable changes should be determined. The part being tested should be moved rapidly. Normally, a very slight degree of movement is appreciated in the digits (as little as 1 degree of an arc). The test should be repeated enough times to eliminate chance (50 percent of responses). Defective perception of passive movement is judged by comparison with a normal limb or, if perception is bilaterally defective, on the basis of what the examiner has learned through experience to be normal. Slight impairment may be disclosed by a slowness of response or, if the digit is displaced very slowly, by an unawareness or uncertainty that movement has occurred; or, after the digit has been displaced in the same direction several times, the patient may misjudge the first movement in the opposite direction; or, after the examiner has moved the toe, the patient may make a number of small voluntary movements of the toe in an apparent attempt to determine its position or the direction of the movement. Inattentiveness will also cause some of these errors.

The lack of position sense in the legs can also be demonstrated by displacing the limb from its original position and having the patient, with eyes closed, place the other leg in the same position or point to the great toe. If proprioception if abnormal in axial structures, the patient will be unable to maintain his balance with feet together and eyes closed (Romberg sign). This test is often interpreted imprecisely. In the Romberg position, even a normal person whose eyes are closed will sway slightly, and the patient who lacks balance because of cerebellar ataxia or some other motor disorder will sway considerably more if his visual cues are removed. Only a marked discrepancy in balance with eyes open and with eyes closed qualifies as a Romberg sign. The most certain indication of abnormality is the need to step to the side or backward to avoid falling. Mild degrees of unsteadiness in an anxious or suggestible patient may be overcome by diverting his attention, e.g., by having him touch the index finger of each hand alternately to his nose while standing with eyes closed or by following the examiner's finger with his eyes. Patients with authentic proprioceptive problems will sway when there gaze is diverted from the ground and then become more unsteady when the eyes are closed. Patient with factitious unsteadiness, usually will remain stable when they look at the ceiling or a distant object and then become very unsteady when the eyes are closed. It is important to remember that any defect in proprioception (e.g. peripheral neuropathy, myelopathy or vestibulopathy) will lead to a Romberg sign even though the sign, described by Moritz Romberg, was meant to diagnose tabes dorsalis.

Testing of Vibratory Sense

This is a composite sensation comprising touch and rapid alterations of deep-pressure sense. The only cutaneous structure capable of registering such stimuli of this frequency is the rapidly adapting pacinian corpuscle. The conduction of vibratory sense depends on both cutaneous and deep afferent fibers that are carried in the muscle spindle afferent fibers and ascend mainly in the dorsal columns of the cord. Consequently, it is rarely affected by lesions of single nerves but will be disturbed in patients with disease of multiple peripheral nerves, dorsal columns, medial lemniscus, and thalamus. Vibration and position sense are usually impaired in similar conditions, although one of them (most often vibration sense) may be affected disproportionately. With advancing age, vibration is the sensation most commonly diminished, especially at the toes and ankles (see further on).

Vibration sense is tested by placing a tuning fork with a low rate and long duration of vibration (generally 128 Hz) over the bony prominences, making sure that the patient responds to the vibration, not simply to the pressure of the fork, and that he is not trying to listen to it. As with thermal and pain testing, there are mechanical devices that quantitate vibration sense. Quantitative tuning forks with a 0-8 scale are available but it is sufficient for most clinical purposes to compare the point tested with a normal part of the patient or the examiner. The examiner may detect the vibration after it ceases for the patient by holding a finger under the distal interphalangeal joint, the handle of the tuning fork being placed on the dorsal aspect of the joint. Or the vibrating fork is allowed to run down until the moment that vibration is no longer perceived, at which point the fork is transferred quickly to the corresponding part of the examiner and the time to extinction is noted. There is a small degree of accommodation to the vibration stimulus, so that slight asymmetries detected by rapid shifting from a body part on one side to the other should be interpreted accordingly. The perception of vibration at the patella after it has disappeared at the ankle or at the anterior iliac spine after it has disappeared at the knee is indicative of a length-dependent peripheral neuropathy. The approximate level of pinprick loss from a spinal cord lesion can be corroborated by testing vibratory sensation over the iliac crests and successive dorsal vertebral spines.

Testing of Discriminative (Parietal Lobe Cortical) Sensation

Damage to the parietal lobe sensory cortex or to the thalamocortical projections results in a particular type of disturbance—namely, an inability to make sensory discriminations and to integrate spatial and temporal sensory information (see further under "Sensory Loss Caused by Lesions of the Parietal Lobe" and Chap. 22). Lesions in these structures usually disturb complex sensory perception but leave the primary modalities (touch, pain, temperature, and vibration sense) relatively little affected. The integrity of discriminative sensory functions can be assessed only if it is first established that the primary sensory modalities on which they depend (mainly touch) are largely normal. If a cerebral lesion is suspected, discriminative function should be tested further in the following ways.

Two-Point Discrimination

The ability to distinguish two points from one is tested by using a compass, the points of which should be blunt and applied simultaneously and painlessly. The distance at which such stimuli can be recognized as a distinct pair varies but is roughly 1 mm at the tip of the tongue, 2 to 3 mm on the lips, 3 to 5 mm at the fingertips, 8 to 15 mm on the palm, 20 to 30 mm on the dorsal hands and feet, and 4 to 7 cm on the trunk. It is characteristic of the patient with a lesion of the sensory cortex to mistake two points for one, although occasionally the opposite occurs.

Cutaneous Localization and Figure Writing (Graphesthesia)

Localization of cutaneous tactile or painful stimuli is tested by touching various points on the body and asking the patient to place the tip of his index finger on the point stimulated or on the corresponding point of the examiner's limb. Recognition of numbers or letters traced on the skin (these should be larger than 4 cm on the palm) with a pencil or similar object or the direction of a line drawn across the skin also depends on localization of tactile stimuli. Normally, traced numbers as small as 1 cm can be detected on the pulp of the finger if drawn with a pencil. According to Wall and Noordenbos, these are also the most useful and simple tests of posterior column function.

Appreciation of Texture, Size, and Shape

Appreciation of texture depends mainly on cutaneous impressions, but recognition of the shapes and sizes of objects is based on sensory experience from deeper receptors as well. Inability to recognize shape and form is frequently a manifestation of cortical disease, but a similar clinical defect will occur if tracts that transmit proprioceptive and tactile sensation are interrupted by lesions of the spinal cord and brainstem (and, of course, of the peripheral nerves). This type of sensory defect is called stereoanesthesia (see further on, under "Posterior [Dorsal] Column Syndrome") and is distinguished from astereognosis, which connotes an inability to identify an object by palpation, even though the primary sense data (touch, pain, temperature, and vibration) are intact. In practice, pure astereognosis is rarely encountered, and the term is employed when the impairment of superficial and vibratory sensation in the hands seems to be of insufficient severity to account for the defect in tactile object identification. Defined in this way, astereognosis is either right- or left-sided and, with the qualifications mentioned below, is the product of a lesion in the opposite hemisphere, involving the sensory cortex, particularly S2 or the thalamoparietal projections.

The traditional doctrine that somatic sensation is represented only in the contralateral parietal lobe is not absolute. Beginning with the report by Oppenheim in 1906, there have been sporadic patients who showed bilateral astereognosis or loss of tactile sensation as a result of an apparently unilateral cerebral lesion. The correctness of these observations was corroborated by Semmes and colleagues, who tested a large series of patients with traumatic lesions involving either the right or left cerebral hemisphere. They found that the impairment of sensation (particularly discriminative sensation) following right-and left-sided lesions was not strictly comparable; the left hand as well as the right tended to be impaired by injury to the left sensorimotor region, whereas only the left hand tended to be affected by injury to the right sensorimotor region. These observations, with minor qualifications, were also confirmed by Carmon and by Corkin and associates, who investigated the sensory effects of cortical excisions in patients with focal epilepsy. Thus it appears that certain somatosensory functions in some patients are mediated not only by the contralateral hemisphere but also by the ipsilateral one, although the contribution of the former is undoubtedly the more significant.

The traditional concept of left hemispheric dominance in respect to tactile perception has been questioned by Carmon and Benton, who found that the right hemisphere is particularly important in perceiving the direction of tactile stimuli. Also, Corkin and associates observed that patients with lesions of the right hemisphere show a consistently greater failure of tactile-maze learning than those with left-sided lesions, pointing to a relative dominance of the right hemisphere in the mediation of tactile performance involving a spatial component. Certainly the phenomenon of sensory inattention or extinction is more prominent with lesions of the right as opposed to the left parietal lobe and is most informative if the primary and secondary sensory cortical areas are spared. These matters are considered further on in this chapter and in Chap. 22.

Finally, there is a distinction between astereognosis and tactile agnosia. Some authors (e.g., Caselli) have defined tactile agnosia as a strictly unilateral disorder, right or left, in which the impairment of tactile object recognition is unencumbered by a disturbance of the primary sensory modalities. Such a disorder would be designated by others as a pure form of astereognosis (see above). In our view, tactile agnosia is a disturbance in which a one-sided lesion lying posterior to the postcentral gyrus of the dominant parietal lobe results in an inability to recognize an object by touch in both hands. According to this view, tactile agnosia is a disorder of apperception of stimuli and of translating them into symbols, akin to the defect in naming parts of the body, visualizing a plan or a route, or understanding the meaning of the printed or spoken word (visual or auditory verbal agnosia). These and other agnosias are discussed in Chap. 22.

Sensory Changes Caused by Interruption of a Single Peripheral Nerve

In these cases, the distribution of sensory loss will vary, depending on whether the nerve involved is predominantly muscular, cutaneous, or mixed. Following division of a cutaneous nerve, the area of sensory loss is always less than its anatomic distribution because of overlap from adjacent nerves. That the area of tactile loss is greater than that for pain relates both to a lack of collateralization (regeneration) from adjacent tactile fibers (in contrast to rapid collateral regeneration of pain fibers) and to a greater overlap of pain sensory units. If a large area of skin is involved, the sensory defect characteristically consists of a central portion in which all forms of cutaneous sensation are lost, surrounded by a zone of partial loss, which becomes less marked as one proceeds from the center to the periphery. Perceptions of deep pressure and passive movement are intact because these modalities are mediated by nerve fibers from subcutaneous structures and joints. Along the margin of the hypesthetic zone, the skin becomes excessively sensitive (hyperesthetic); light contact may be felt as smarting and mildly painful, more so as one proceeds from the periphery of the area to its center. According to Weddell, the dysesthesias are attributable to the greater sensitivity of collateral regenerating fibers that have made their way from surrounding healthy pain fibers into the denervated region.

Particular types of lesions have differing effects on sensory nerve fibers, as discussed earlier, but they are nearly always to some extent multimodal. Compression of a nerve ablates mainly the function of large touch and pressure fibers and leaves the function of small pain, thermal, and autonomic fibers intact; procaine and ischemia have the opposite effect. The tourniquet test is instructive in this respect. A sphygmomanometer cuff is applied above the elbow, inflated to a point well above the systolic pressure, and maintained there for as long as 30 min. (This is not particularly painful if the patient does not contract the forearm and hand muscles.) Paresthesiae appears within a few minutes, followed by sensory loss—first of touch and vibration, then of cold, fast pain, heat, and slow pain, in that order and spreading centripetally. Physiologic studies have confirmed the theory of Lewis and colleagues that compression blocks the function of nerve fibers in order of their size. Release of the cuff results in postcompression paresthesia, which has been shown to arise from spontaneous activity that is generated along the myelinated nerve fibers from ectopic sites at a distance from the compression. Within seconds of releasing the cuff, other changes appear—the cold, blanched hand becomes red and hot and there is an array of tingling, stinging, cramp-like sensations that reach maximum intensity in 90 to 120 s and slowly fade (Lewis et al). Sensory function is recovered in an order inverse to that in which it was lost. Similar spontaneous and ectopic discharges probably explain the paresthetic symptoms early in the acute demyelinating neuropathies, even before the appearance of sensory loss or numbness. It is worth emphasizing that these features of compression are not because of nerve ischemia, as commonly stated; instead, they result from reversible physiologic changes in the myelin and underlying axon.

Certain maneuvers for the provocation of positive sensory phenomena—for example, the Tinel sign of tingling upon percussion of a regenerating peripheral nerve and the Phalen sign of paresthesia in the territory of the median nerve on wrist flexion—typify the susceptibility of a damaged nerve to pressure. In the case of a severed nerve, regeneration from the proximal end begins within days. The thin, regenerating sprouts are unusually sensitive to mechanical stimulation, which produces tingling, or the Tinel sign.

Sensory Changes Caused by Involvement of Multiple Nerves (Polyneuropathy) (Table 9-2)

In most instances of polyneuropathy, the sensory changes are accompanied by varying degrees of motor and reflex loss. Usually the sensory impairment is bilaterally symmetrical. Because in most types of polyneuropathy the longest and largest fibers are the most affected, sensory loss is most severe over the feet and legs and, if the upper limbs are affected, over the hands. The term glove-and-stocking, employed to describe the distribution of sensory loss of polyneuropathy, draws attention to the predominantly distal pattern of involvement but fails to indicate that the change from normal to impaired sensation is characteristically gradual. (In psychogenic sensory loss, by contrast, the border between normal and absent sensation is usually sharply defined.) The abdomen, thorax, and face are spared except in the most severe cases, in which sensory changes may be found over the anterior thoracoabdominal escutcheon and around the mouth.

When the neuropathy is primarily demyelinating rather than axonal, paresthesia is an early feature. It is reported in the distal territories of nerves; when short nerves (e.g., the trigeminal) are involved, the paresthesia appears in proximal body parts, such as the perioral region.

The sensory loss of polyneuropathy usually involves all modalities of sensation, and—although it is difficult to equate the degrees of impairment of pain, touch, temperature, vibration, and position senses—one modality may be impaired disproportionately to the others. This clinical feature is explained by the fact that particular diseases of the peripheral nerves selectively damage sensory fibers of different size. For example, degeneration or demyelination of the large fibers that subserve kinesthetic sense causes a loss of vibratory and position sense and relative sparing of pain, temperature, and, to some degree, tactile perception. When extreme, such a polyneuropathy results in pseudoathetoid movements of the outstretched fingers or toes; it may also result in a sensory ataxia because of affection of the large-diameter nerves destined for the spinocerebellar tracts.

By contrast, involvement of the small-caliber myelinated and unmyelinated axons affects pain, temperature, and autonomic sensation, with preservation of proprioceptive, vibration, and tactile sense—producing a syndrome called "pseudosyringomyelia," because it simulates the dissociated pain from tactile sensory loss that is seen in this disease of the spinal cord (see further on, under "Sensory Spinal Cord Syndromes"). Prolonged analgesia may lead to trophic ulcers and Charcot joints. These patterns of sensory loss, as well as those produced by the plexopathies and mononeuritis multiplex, are discussed further in Chap. 46.

Sensory Changes from Involvement of Nerve Roots (Radiculopathy) (Figs. 9-1, 9-2, and 9-3 and Table 9-2)

The surface innervation of the sensory nerve roots serves as one of the most useful and dependable guides to localization in neurology and the main dermatomes are known to all physicians. Because of considerable overlap from adjacent roots, division of a single sensory root does not produce complete loss of sensation in any area of skin. However, compression of a single sensory cervical or lumbar root (e.g., by a herniated intervertebral disc) can cause a segmental impairment of cutaneous sensation. When two or more contiguous roots have been completely divided, a zone of sensory loss can be demonstrated; this is surrounded by a narrow zone of partial loss in which a raised threshold accompanied by excessive sensitivity may or may not be evident. For reasons not altogether clear, partial sensory loss from root lesions is easier to demonstrate by the use of a painful stimulus than by a tactile or pressure stimulus.

Disease of the nerve roots frequently gives rise to "shooting" (lancinating) pains and burning sensations that project down the course of their sensory nerves. The common examples are sciatica, from lower lumbar or upper sacral root compression, and sharp pain radiating from the shoulder and down the upper arm, from cervical disc protrusion.

When multiple roots are affected (polyradiculopathy) by an infiltrative, inflammatory, or compressive process, the syndrome is more complex and must be differentiated from polyneuropathy. The distinguishing features of a polyradiculopathy, aside from pain, are asymmetrical muscle weakness that involves both proximal and distal parts differentially in each limb and a pattern of sensory loss that is consistent with affection of several roots, not necessarily contiguous ones. Details are found in Chap. 46.

Sensory Changes from Involvement of Sensory Ganglia (Sensory Neuronopathy, Ganglionopathy) (Table 9-2)

Widespread disease of the dorsal root ganglia produces many of the same sensory defects as disease of the posterior nerve roots, but it is unique in that proximal areas of the body also show pronounced sensory loss; the face, oral mucosa, scalp, trunk, and genitalia may be sites of hypesthesia and hypalgesia. Proprioception is diminished or lost in distal and, to some extent, proximal body parts, giving rise also to ataxic movements, often quite severe, and to pseudoathetosis. Tendon reflexes are lost. Sometimes there are additional features of dysautonomia, but strength is entirely spared. Recognition of this unusual pattern of pansensory loss is of considerable diagnostic importance, because it raises for consideration a number of underlying diseases that might otherwise be overlooked; these diseases are discussed in Chap. 46. The main causes of this syndrome are paraneoplastic, connective tissue disease, particularly Sjögren syndrome, toxic exposure, and idiopathic inflammation.

Tabetic Syndrome

The tabetic syndrome may be considered a type of polyradiculopathy or ganglionopathy, but it is often considered with diseases of the spinal cord because it produces wallerian degeneration of posterior columns. It results from damage to the large proprioceptive and other fibers of the posterior lumbosacral (and sometimes cervical) roots. It was in the past typically caused by neurosyphilis but also by diabetes mellitus, and other diseases that involve the posterior roots or dorsal root ganglia. Numbness or paresthesia and "lightning" or lancinating pains are frequent complaints; areflexia, abnormalities of gait (gait of sensory ataxia), and hypotonia without significant muscle weakness are found on examination. The sensory loss may involve only vibration and position senses in the lower extremities, but loss or impairment of superficial or deep pain sense or of touch may appear in severe cases. In this case, loss of deep tendon reflexes distinguishes the sensory root syndrome from a lesion in the posterior columns. The feet and legs are most affected in tabes, much less often the arms and trunk. The Romberg sign is prominent. Frequently, atonicity of the bladder with retention of urine and trophic joint changes (Charcot joints) and crises of abdominal ("gastric") pains are associated.

Cases of congenital absence of all cutaneous sensation resulting from the lack of development of small sensory ganglion cells may produce the tabetic syndrome; this is discussed in Chap. 8. A similar but partial defect may be found in the Riley-Day syndrome (Chap. 26). There are also forms of hereditary polyneuropathy that cause universal insensitivity.

Sensory Spinal Cord Syndromes (Fig. 9-7)

See also Chap. 44.

Complete Spinal Sensory Syndrome

With a complete transverse disruption of the spinal cord, the most striking loss is of power; most characteristic, however, is a loss of all forms of sensation below a level that corresponds to the lesion. There may be a narrow zone of hyperesthesia at the upper margin of the anesthetic zone. Loss of pain, temperature, and touch sensation begins one or two segments below the level of the lesion; vibratory and position senses have less-discrete levels but they can be detected by careful examination. The sensory (and motor) loss in spinal cord lesions that involve both gray and white matter is expressed in patterns corresponding to bodily segments or dermatomes. These are shown in Figs. 9-3 and 9-4 and are most obvious on the trunk, where each intercostal nerve has a transverse distribution.

Also, it is important to remember that during the subacute evolution of a transverse spinal cord lesion, there may be a discrepancy between the level of the lesion and that of the sensory loss, the latter ascending as the lesion progresses. This can be understood if one conceives of a lesion as evolving from the periphery to the center of the cord, affecting first the outermost fibers carrying pain and temperature sensation from the legs. Conversely, a lesion advancing from the center of the cord will affect these modalities in the reverse order, in a pattern of sacral sparing, meaning that sensation is preserved over the buttocks and anal region but is absent over the trunk and legs.

Hemisection of the Spinal Cord (Brown-SéQuard Syndrome)

Disease may be confined to or predominate on one side of the spinal cord; pain and thermal sensation are affected on the opposite side of the body, and proprioceptive sensation is affected on the same side as the lesion. Although rarely present in its entirety, a partial Brown-Séquard syndrome is common in practice. The loss of pain and temperature sensation begins one or two segments below the lesion. An associated spastic motor paralysis on the side of the lesion completes the syndrome (Fig. 9-7). Tactile sensation is not greatly affected, as the fibers from one side of the body are distributed in tracts (posterior columns, anterior and lateral spinothalamic) on both sides of the cord.

Syringomyelic Syndrome (Lesion of the Central Gray Matter)

Because fibers conducting pain and temperature sensation cross the cord in the anterior commissure, a lesion of considerable vertical extent in this location characteristically abolishes these modalities on one or both sides over several segments (dermatomes) but spares tactile sensation (Fig. 9-7). This type of dissociated sensory loss usually occurs in a segmental distribution, and because the lesion frequently involves other parts of the gray matter, varying degrees of segmental amyotrophy and reflex loss are usually present as well. If the lesion has spread to the white matter, corticospinal, spinothalamic, and posterior column signs will be conjoined. The most common cause of such a lesion in the cervical region is the centrally situated developmental syringomyelia; less-common causes are intramedullary tumor, trauma, and hemorrhage. A pseudosyringomyelic syndrome was mentioned earlier in relation to small-fiber neuropathies that simulate syringomyelia.

Posterior (Dorsal) Column Syndrome

Paresthesias in the form of tingling and pins-and-needles sensations or girdle- and band-like sensations are common complaints with posterior column disease. In some cases there may be the additional feature of a diffuse, burning, unpleasant sensation in response to pinprick. Loss of vibratory and position sense occurs below the level of the lesion, but the perception of pain and temperature is affected relatively little or not at all. Because posterior column lesions are caused by the interruption of central projections of the dorsal root ganglia cells, they may be difficult to distinguish from a process that affects large fibers in sensory roots (tabetic syndrome); however, the tendon reflexes are spared in the former and eliminated in tabes. In some diseases that involve the dorsal columns, vibratory sensation may be involved predominantly, whereas in others position sense is more affected.

With complete posterior column lesions, only a few of which have been verified by postmortem examinations, not only is the patient deprived of knowledge of movement and position of parts of the body below the lesion but all types of sensory discrimination are impaired (see Nathan et al for a review of the subject). If the lesion is in the high cervical region, there is clumsiness in the palpation of objects and an inability to recognize the qualities of objects by touch, even though touch-pressure sensation is relatively intact. Stereoanesthesia is also expressed by impaired graphesthesia and tactile localization. There may be unusual disturbances of touch and pressure, manifesting as lability of threshold, persistence of sensation after removal of the stimulus, and sometimes tactile and postural hallucinations. Nathan and colleagues confirmed this counterintuitive observation that lesions of the posterior columns cause only slight defects in touch and pressure sensation and that lesions of the anterolateral spinothalamic tracts also cause minimal or no defects in these modalities. This is not easily reconciled with clinical observations of patients with lesions in the posterior columns. However, a combined lesion in both pathways causes a total loss of tactile and pressure sensibility below the lesion.

The loss of sensory functions that follows a posterior column lesion—such as impaired two-point discrimination; figure writing; detection of size, shape, weight, and texture of objects; and ability to detect the direction and speed of a moving stimulus on the skin—may simulate a parietal "cortical" lesion, but differs in that vibratory sense is also lost in spinal cord syndromes. In several cases on record, interruption of the posterior columns by surgical incision or other type of injury did not cause a permanent loss of the sensory modalities thought to be subserved by these pathways (Cook and Browder, Wall and Noordenbos). Because postmortem studies of these cases were not performed, it is possible that some of the posterior column fibers had been spared. Also, it should be realized that not all proprioceptive fibers ascend to the gracile and cuneate nuclei; some proprioceptive fibers leave the posterior columns in the lumbar region and synapse with secondary neurons in the spinal gray matter and ascend in the ipsilateral posterolateral funiculus. Only cutaneous fibers continue to the gracile and cuneate nuclei.

The usual causes of the posterior column syndrome are multiple sclerosis, vitamin B12 deficiency, copper deficiency, tabes dorsalis, and HIV and human T-lymphotropic virus (HTLV) type 1 infection.

Anterior Myelopathy (Anterior Spinal Artery Syndrome)

With infarction of the spinal cord in the territory of supply of the anterior spinal artery or with other lesions that affect the ventral portion of the cord predominantly, as in some cases of myelitis, one finds a loss of pain and temperature sensation below the level of the lesion but with relative or absolute sparing of proprioceptive sensation. Because the corticospinal tracts and the ventral gray matter also lie within the area of distribution of the anterior spinal artery, spastic paralysis is a prominent feature (Fig. 9-7).

Disturbances of Sensation from Lesions of the Brainstem

A characteristic feature of medullary lesions is the occurrence, in many instances, of a crossed sensory disturbance, i.e., a loss of pain and temperature sensation on one side of the face and on the opposite side of the body. This is accounted for by involvement of the descending spinal trigeminal tract or its nucleus and the crossed lateral spinothalamic tract on one side of the brainstem and is nearly always caused by a lateral medullary infarction (Wallenberg syndrome). In the upper medulla, pons, and midbrain, the crossed trigeminothalamic and lateral spinothalamic tracts run together; a lesion at these levels causes loss of pain and temperature sense on the opposite half of the face and body. There are no tactile paresthesias, only thermal or painful dysesthesias. In the upper brainstem, the spinothalamic tract and the medial lemniscus become confluent, so that an appropriately placed lesion causes a contralateral loss of all superficial and deep sensation. Cranial nerve palsies, cerebellar ataxia, and motor paralysis are almost invariably associated, as indicated in the discussion of strokes in this region (Chap. 34). In other words, a lesion in the brainstem at any level is unlikely to cause an isolated sensory disturbance.

Hemisensory Loss Caused by a Lesion of the Thalamus (Syndrome of Dejerine-Roussy)

Involvement of the VPL and VPM nuclei of the thalamus, usually because of a vascular lesion, causes loss or diminution of all forms of sensation on the opposite side of the body. Position sense is affected more frequently than any other sensory function and is usually, but not always, more profoundly reduced than loss of touch and pinprick. With partial recovery of sensation, or with an acute but incomplete lesion, spontaneous pain or discomfort (thalamic pain), sometimes of the most distressing type, may appear on the affected side of the body and any stimulus may then have a diffuse, unpleasant, lingering quality (anterior and proximal syndromes). Thermal—especially cold—stimuli, emotional disturbance, loud sounds, and even certain types of music may aggravate the painful state. In spite of this overresponse to stimuli, the patient usually shows an elevated pain threshold, i.e., a stronger stimulus than normal is necessary to produce a sensation of pain (hypalgesia with hyperpathia). The same type of pain syndrome may occasionally accompany lesions of the white matter of the parietal lobe, the medial lemniscus, or even the posterior columns of the spinal cord.

It should be pointed out that a symptomatic hemisensory syndrome, usually with few objective changes, occurs frequently without manifest evidence of thalamic or spinal cord damage. This particularly occurs in young women, as pointed out by Toth. A number of our patients with this benign condition have had migraine, and one, as in Toth's series, had the antiphospholipid syndrome, but the connections between all these entities is tenuous and many cases are considered to be psychogenic.

Sensory Loss Caused by Lesions of the Parietal Lobe

In the anterior parietal lobe syndrome (Verger-Dejerine syndrome), there are disturbances mainly of discriminative sensory functions of the opposite arm, leg, and side of the face without impairment of the primary modalities of sensation (unless the lesion is extensive and deep). Loss of position sense and sense of movement, impaired ability to localize touch and pain stimuli (topagnosia), widening of two-point threshold, and astereognosis are the most prominent findings, as described earlier in this chapter and in Chap. 22, "Clinical Effects of Parietal Lobe Lesions."

Another characteristic manifestation of parietal lobe lesions is sensory inattention, extinction, or neglect. In response to bilateral simultaneous testing of symmetrical parts, using either tactile or painful stimuli, the patient may acknowledge only the stimulus on the sound side; or, if the face and hand or foot on the affected side is touched or pricked, only the stimulus to the face may be noticed. Apparently cranial structures command more attention than other less richly innervated parts. Yet each stimulus, when applied separately to each side or to each part of the affected side, is properly perceived and localized. In the case of sensory neglect, the patient ignores one side of the body and extrapersonal space contralateral to the parietal lesion, which is usually in the nondominant hemisphere. Left parietal lesions may also cause (right) sensory neglect, but less frequently and less profoundly. Sensory neglect or extinction, which may also occur occasionally with posterior column and medial lemniscus lesions, may be detected in persons who disclaim any sensory symptoms. These phenomena and other features of parietal lobe lesions are considered further in Chap. 22.

Yet another parietal lobe syndrome, Dejerine-Mouzon, is featured by a severe impairment of the primary modalities of sensation (pain, thermal, tactile, and vibratory sense) over half of the body. Motor paralysis is variable; with partial recovery, there may be a clumsiness that resembles cerebellar ataxia. Because the sensory disorder simulates that caused by a thalamic lesion, it was called pseudothalamic by Foix and coworkers. Hyperpathia, much like that of the Dejerine-Roussy syndrome (see above), has also been observed in patients with cortical–subcortical parietal lesions. The pseudothalamic syndrome was related by Foix and colleagues to a sylvian infarct; Bogousslavsky and associates traced it to a parietal infarct caused by occlusion of the ascending parietal branch of the middle cerebral artery. In each of the aforementioned parietal lobe syndromes, if the dominant hemisphere is involved, there may be an aphasia, a bimanual tactile agnosia, or a Gerstmann syndrome; with nondominant lesions, there may be anosognosia (see Chap. 22).

Often with parietal lesions, the patient's responses to sensory stimuli are variable. A common mistake, as emphasized by Critchley, is to attribute this abnormality to hysteria (see also below). A lesion confined to only a part of the parietal cortex (the best examples have been caused by glancing bullet or shrapnel wounds of the skull) may result in a circumscribed loss of superficial sensation in an opposite limb, mimicking a root or peripheral nerve lesion.

Sensory Loss Due to Suggestibility and Hysteria

(See also Chap. 51)

The possibility of suggesting sensory loss to a patient is a very real one, as has already been indicated. Hysterical patients may complain of a complete hemianesthesia—sometimes with the overtly hysterical findings of reduced hearing, sight, smell, and taste on one side—as well as impaired vibration sense over only half the skull and sternum, most of these being anatomic impossibilities. A frequently used test to disclose this feature is performed by placing a vibrating tuning fork on each side adjacent to the middle of the forehead. The transmission of vibration through the bone assures that loss of sensation on one side of the midline is not possible. Anesthesia of an entire limb or a sharply defined sensory loss over part of a limb, not conforming to the distribution of a root or cutaneous nerve, may also be observed. The diagnosis of hysterical hemianesthesia is best made by eliciting the other relevant symptoms of hysteria or, if this is not possible, by noting the discrepancies between the sensory loss displayed by the patient and that which occurs as part of the known, anatomically verified sensory syndromes.

Sometimes, in a patient with no other neurologic abnormality or in one with a definite neurologic syndrome, one is dismayed by sensory findings that are completely inexplicable and discordant. In such cases, one must try to reason through to the diagnosis by disregarding the sensory findings or approach the finding as revealing a second disorder such as a neurofibroma of a nerve root.

Laboratory Diagnosis of Somatosensory Syndromes

Affirmation of the clinical sensory syndromes is often possible by the application of electrophysiologic testing. Slowing and reduced amplitude of sensory nerve conduction is found with lesions of nerve, but only if the lesion lies distal to (or within) the sensory ganglion. Severe sensory loss in a neuropathic pattern with preserved sensory nerve action potentials therefore indicates a radiculopathy. Loss or slowing of the H reflex and F responses corroborates the presence of lesions in proximal parts of nerves, plexuses, and roots. By the use of somatosensory evoked potentials, it is possible to demonstrate slowing of conduction in the peripheral nerves or roots, in the pathways from spinal cord to a point in the lower medulla, in the medial lemniscus to the thalamus, and in the pathway from the thalamus to the cerebral cortex. In the context of regional sensory loss, evoked potentials find their greatest utility in demonstrating root disease when sensory nerve conduction studies are normal; otherwise, they are used most frequently to support the diagnosis of multiple sclerosis, in which case there may or may not be corresponding sensory features. (See Chap. 2 for discussion of evoked potential testing.)

In practice, it is seldom necessary to examine all modalities of sensation and perception. With single peripheral nerve lesions, touch and pinprick testing are the most informative. With spinal cord disease, pinprick and thermal stimuli are most revealing of lateral column lesions; testing the senses of vibration, position, and movement, and particularly the sense of direction of a dermal stimulus, reliably indicates posterior column lesions. Touch is the least useful. In brainstem lesions, all modes of sensation, including touch, may be affected, and this applies in general to thalamic lesions. Thus, one is guided in the selection of tests by the suspected locale of the disease.