23: Touch

• Active and Passive Touch Evoke Similar Responses in Mechanoreceptors

• The Hand Has Four Types of Mechanoreceptors

  • Receptive Fields Define the Zone of Tactile Sensitivity

  • Two-Point Discrimination Tests Measure Texture Perception

  • Slowly Adapting Fibers Detect Object Pressure and Form

  • Rapidly Adapting Fibers Detect Motion and Vibration

  • Both Slowly and Rapidly Adapting Fibers Are Important for Grip Control

• Tactile Information Is Processed in the Central Touch System

  • Cortical Receptive Fields Integrate Information from Neighboring Receptors

  • Neurons in the Somatosensory Cortex Are Organized into Functionally Specialized Columns

  • Cortical Columns Are Organized Somatotopically

• Touch Information Becomes Increasingly Abstract in Successive Central Synapses

  • Cognitive Touch Is Mediated by Neurons in the Secondary Somatosensory Cortex

  • Active Touch Engages Sensorimotor Circuits in the Posterior Parietal Cortex

• Lesions in Somatosensory Areas of the Brain Produce Specific Tactile Deficits

• An Overall View

In this chapter on the sense of touch, we focus on the hand because of its importance in the sensory appreciation of object properties and its role in skilled motor tasks. The hand is one of evolution's great creations. The fine manipulative capacity provided by our fingers is possible because of their fine sensory capacity; if we lose tactile sensation in our fingers we lose manual dexterity.

When we become skilled in the use of a tool, such as a scalpel or a pair of scissors, we feel conditions at the working surface of the tool as though our fingers were there because two groups of mechanoreceptors monitor the vibrations and forces produced by those distant conditions. When we scan our fingers across a surface we feel its form and texture because another group of mechanoreceptors has high spatial and temporal acuity. A blind person uses this capacity to read Braille at a hundred words per minute. When we grip and manipulate an object we do so delicately, with only as much force as is required, because yet another group of mechanoreceptors continually monitors slip and adjusts our grip appropriately.

We are also able to recognize objects placed in the hand from touch alone. When we are handed a baseball we recognize it instantly without having to look at it because of its shape, size, weight, density, and texture. We do not have to think about the information provided by each finger to deduce that the object must be a baseball; the information flows to memory and instantly matches previously stored representations of baseballs. Even if we have never previously handled a baseball, we perceive it as a single object, not as a collection of discrete features. The somatosensory pathways of the brain have the daunting task of integrating information from thousands of sensors in each hand and transforming it to a form suitable for cognition.

Sensory information is extracted for the purpose of motor control as well as cognition, and different kinds of information are extracted for those purposes. We can, for example, shift our attention from the baseball's shape to its location in the hand to readjust our grip for an effective throw. This selective attention to different aspects of the sensory information is brought about by cortical mechanisms.

Touch is defined as direct contact between two physical bodies. In neuroscience touch refers to the special sense by which contact with the body is perceived consciously. Touch can be active, as when you move your hand or some other part of the body against another surface, or passive, as when someone or something else touches you.

Active and passive modes of tactile stimulation excite the same population of receptors in the skin and evoke similar responses in afferent fibers. They differ somewhat in cognitive features that reflect attention and behavioral goals during the period of stimulation. Passive touch is used for naming objects or describing sensations; active touch is used when the hand manipulates objects. The sensory and motor components of touch are intimately connected anatomically in the brain and are important functionally in guiding motor behavior.

The distinction between active and passive touch is important clinically when patients have deficits in hand use. Motor deficits such as weakness, stiffness, or clumsiness may result from sensory loss, which is why passive sensory testing is important in the neurological examination (see Appendix B). Common neurological tests for touch include measurements of detection thresholds, vibration sense, two-point or texture discrimination, and the ability to recognize form through touch (stereognosis). The neural mechanisms underlying these tests are discussed in this chapter. Other common tests of somatosensory function—tendon reflexes, pinprick, and thermal tests—are discussed in other chapters.

Nearly all of these neurological tests involve passive touch; the clinician stimulates the body in a controlled and reproducible fashion. The tests measure the sensitivity and function of various receptors for touch. Deviations from the expected values may help diagnose the sensory deficits or lesions that underlie sensory dysfunction.

The softness and compliance of the skin plays a major role in the sense of touch. When an object contacts the hand, the skin conforms to its contours, forming a mirror image of the object surface. The resultant displacement and indentation of the skin stretches the tissue, thereby stimulating the sensory endings of mechanoreceptors at or near the region of contact.

These receptors are highly sensitive and are continually active as we manipulate objects and explore the world with our hands. They provide information to the brain about the position of the stimulus on the skin, its shape and surface texture, the amount of force applied at the contact point, and how these features change over time when the hand or the object moves.

Tactile sensations in the human hand arise from four kinds of mechanoreceptors: Meissner corpuscles, Merkel cells, Pacinian corpuscles, and Ruffini endings (Figure 23–1). The sense of touch can be understood as the combined result of the information provided by these four systems acting in concert.

Each receptor responds in a distinctive manner depending on its morphology, innervation pattern, and depth in the skin. Receptors are innervated by either slowly adapting or rapidly adapting axons. Slowly adapting (SA) fibers respond to steady skin indentation with a sustained discharge, whereas rapidly adapting (RA) fibers stop firing as soon as the indentation is stationary. Sustained mechanical sensations from the hand must accordingly arise from the SA fibers. The sensation of motion across the skin is signaled by RA fibers. The receptors are further subdivided into two types based on size and location in the skin; each type includes both rapidly and slowly adapting fibers.

Thus tactile sensation in the hand is mediated by four functional units: rapidly adapting type 1 (RA1), slowly adapting type 1 (SA1), rapidly adapting type 2 (RA2), and slowly adapting type 2 (SA2). Each unit consists of an afferent fiber, the fiber's distal branches, and the receptor organ(s) that surround the axon terminals (Table 23–1).

Type 1 fibers terminate in clusters of small receptors in the superficial layers of the skin at the margin between the dermis and epidermis (Figure 23–1). RA1 fibers are the most numerous tactile afferents in the hand, reaching a density of approximately 2 per mm² at the fingertip in man and monkey. The RA1 receptor organ, the Meissner corpuscle, is a globular, fluid-filled structure that encloses a set of flattened, lamellar cells originating from the myelin sheath (see Figure 21–6). The lamellae are coupled mechanically to the edge of the papillary ridge by collagen fibers, a relationship that confers fine mechanical sensitivity to frictional forces as the hand is moved across surfaces (Box 23–1). An RA1 axon typically innervates 10 to 20 Meissner corpuscles, integrating information from several adjacent papillary ridges. Each Meissner corpuscle is innervated by 2 to 5 RA1 axons (Figure 23–3A).

SA1 fibers are also widely distributed in the skin, particularly in the fingertips. The SA1 receptor organs, the Merkel cells, consist of small epithelial cells that surround the terminal branches of an axon. Each Merkel cell encloses a semirigid structure that transmits compressive strain to the sensory nerve ending. Because there are synapse-like junctions between the Merkel cells and the SA1 axon terminals, it has been proposed that the mechanosensitive ion channels reside in the Merkel cells rather than in the nerve endings. Merkel cells are densely clustered in the center of each papillary ridge in glabrous skin (Figure 23–3A), placing them in an excellent position to detect deformation of the overlying skin, either from pressure above or lateral stretch. In hairy skin Merkel cells are localized in small clusters called touch domes (Figure 23–3B).

Type 2 fibers innervate the skin sparsely and terminate in single large receptors in the deeper layers of the dermis or in the subcutaneous tissue (Figure 23–1). The receptors are larger and less numerous than the receptor organs of the type 1 fibers. The large size of these receptors allows them to sense mechanical displacement at some distance from the sensory nerve endings.

The RA2 fibers terminate in Pacinian corpuscles located in the subcutaneous tissue (Figure 23–1). Each RA2 axon terminates without branching in a single Pacinian corpuscle, and each Pacinian corpuscle receives but a single RA2 axon. Pacinian corpuscles are large onion-like structures in which successive layers of connective tissue are separated by fluid-filled spaces. These layers surround the unmyelinated RA2 ending and its myelinated axon up to one or more nodes of Ranvier. The capsule amplifies high-frequency vibration, a role that is important for tool use. Estimates of the number of Pacinian corpuscles in the human hand range from 2,400 in the young to 300 in the elderly.

The SA2 fibers innervate Ruffini endings concentrated at the finger and wrist joints and along the skin folds in the palm; they are relatively rare in the fingertips. The Ruffini endings are elongated fusiform structures that enclose collagen fibrils extending from the subcutaneous tissue to folds in the skin at the joints, in the palm, or in the fingernails. The SA2 nerve endings are intertwined between the collagen fibers in the capsule, and are excited by stimuli that stretch the receptor along its long axis.

Receptive Fields Define the Zone of Tactile Sensitivity

Individual mechanoreceptor fibers convey information from a limited area of skin called the receptive field. Tactile receptive fields have been determined in the human hand using microneurography.

Åke Vallbo and Roland Johansson inserted microelectrodes through the skin into the median or ulnar nerves in the human hand and recorded the responses of individual afferent fibers. They found that in humans, as in other primate species, there are important differences between touch receptors, both in their physiological responses and in the structure of their receptive fields.

Type 1 fibers have small, highly localized receptive fields with multiple spots of high sensitivity that reflect the branching patterns of their axons in the skin (Figure 23–4). Receptive fields on the fingertips are the smallest on the body, averaging 11 mm² for SA1 fibers and approximately 25 mm² for RA1 fibers. The fields are small because of the high density of receptors in the fingertips. Receptive fields become progressively larger on the proximal phalanges and the palm, consistent with the lower density of mechanoreceptors in these regions.

In contrast, type 2 fibers innervating the deep layers of skin are connected to only a single Pacinian corpuscle or Ruffini ending. Because these receptors are large they collect information from a broader area of skin. Their receptive fields contain a single "hot spot" where sensitivity to touch is greatest; this point is located directly above the receptor (Figure 23–4).

The receptive fields of type 1 fibers are significantly smaller than most objects that contact the hand. Thus RA1 and SA1 fibers detect small pieces of an object, signaling the properties of only a portion of its surface. As in the visual system, the spatial features of objects are distributed across a population of stimulated receptors with responses that are integrated in the brain to form a unified percept.

Two-Point Discrimination Tests Measure Texture Perception

The ability of humans to resolve spatial details of textured surfaces depends on which part of the body is contacted. Tactile acuity is highest on the fingertips and the lips, where receptive fields are smallest (Figure 23–5A). The separation that defines performance midway between chance and perfect discrimination, the threshold for spatial acuity, is approximately 1 mm on the fingertips of young adults; by the sixth or seventh decade of life it declines on average to approximately 2 mm. When we grasp an object we can discriminate features of its surface separated by as little as 0.5 mm. Humans are able to distinguish horizontal from vertical orientation of gratings with remarkably narrow spacing of the ridges (Figure 23–5B).

Tactile acuity is slightly greater in women than in men and varies between fingers but not between hands. The distal pad of the index finger has the keenest sensitivity; spatial acuity declines progressively from the index to the little finger and declines rapidly at locations proximal to the distal finger pads. Tactile spatial resolution is 50% poorer at the distal pad of the little finger and six to eight times coarser on the palm.

Tactile acuity on proximal parts of the body decreases in parallel with the growing size of receptive fields of SA1 and RA1 fibers (Figure 23–5A). When a pair of probes is spaced several millimeters apart on the hand, each of them is perceived as a distinct point because they produce separate dimples in the skin and stimulate nonoverlapping populations of SA1 and RA1 receptors. As the probes are moved closer together, the two sensations become blurred because both probes are contained within the same receptive field. The spatial interactions between tactile stimuli form the basis of neurological tests of two-point discrimination and texture recognition.

Blind individuals use the fine spatial sensitivity of SA1 and RA1 fibers to read Braille. The Braille alphabet represents letters as simple dot patterns that are easy to distinguish by touch (Figure 23–6). A blind person reads Braille by moving the fingers over the dot patterns. This hand movement enhances the sensations produced by the dots.

Because the Braille dots are spaced approximately 3 mm apart, a distance greater than the receptive field diameter of an SA1 fiber, each dot stimulates a different set of SA1 fibers. An SA1 fiber fires a burst of action potentials as a dot enters its receptive field and is silent once the dot leaves the field (Figure 23–6). Specific combinations of SA1 fibers that fire synchronously signal the spatial arrangement of the Braille dots. RA1 fibers also discriminate the dot patterns, enhancing the signals provided by SA1 fibers.

Slowly Adapting Fibers Detect Object Pressure and Form

The most important functional feature of the slowly adapting fibers (SA1 and SA2) is their ability to signal skin deformation and pressure. The sensitivity of an SA1 receptor to edges, corners, points, and curvature provides information about object shape, size, surface texture, and compliance. We perceive an object as hard or rigid if it indents the skin, and soft if the skin surface instead deforms the object.

Paradoxically, as an object's diameter increases, the responses of individual SA1 fibers become weaker and the sensation less distinct. For example, the tip of a pencil pressed 1 mm into the skin feels sharp, unpleasant, and highly localized at the contact point, whereas a 1 mm indentation by the eraser feels blunt and broad. The weakest sensation is evoked by a flat surface pressed against the finger pad (Figure 23–7).

To understand why these objects evoke different sensations, we need to consider the physical events that occur when the skin is touched. When a pencil tip is pressed against the skin it dimples the surface at the contact point and forms a shallow, sloped basin in the surrounding region (approximately 4 mm in radius). Although the indentation force is concentrated in the center, the surrounding region is also perturbed by local stretch, called tensile strain. SA1 receptors at both the center and the surrounding "hillsides" of skin are stimulated, firing spike trains proportional to the degree of local stretch.

If a second probe is pressed close to the first one, more SA1 fibers are stimulated but the neural response of each fiber is reduced because the force needed to displace the skin is shared between the two probes. Ken Johnson and his colleagues have shown that as more probes are added within the receptive field, the response intensity of each fiber becomes progressively weaker because the displacement forces on the skin are distributed across the entire contact zone. Thus the skin mechanics results in a case of "less is more." Individual SA1 fibers respond more vigorously to a small object than to a large one because the force needed to indent the skin is concentrated at a small contact point. In this manner each SA1 fiber integrates the local skin indentation profile within its receptive field.

The sensitivity of SA1 receptors to local stretch of the skin also enables them to detect edges, the places where an object's curvature changes abruptly. SA1 firing rates are many times greater when a finger touches an edge than when it touches a flat surface.

The indentation force of a flat or gently curved surface is distributed symmetrically within the central contact zone, whereas the force applied by an object boundary displaces the skin asymmetrically, beyond the edge as well as at the edge. This asymmetric distribution of force produces enhanced responses from receptive fields located along the edges of an object. As edges are often perceived as sharp, we tend to grasp objects on flat or gently curved surfaces rather than by their edges.

The SA2 fibers that innervate Ruffini endings respond more vigorously to stretch of the skin than to indentation because the receptors are located along the palmar folds or at the finger joints. The SA2 fibers therefore provide information about the shape of large objects grasped with the entire hand, the "power grasp" in which all five fingers press an object against the palm. They also provide information about hand shape and finger movements when the hand is empty. If the fingers are fully extended and abducted we feel the stretch in the palm and proximal phalanges as the glabrous skin is flattened. Similarly, if the fingers are fully flexed, forming a fist, we feel the stretch of the skin on the back of the hand, particularly over the metacarpal-phalangeal and proximal interphalangeal joints.

The SA2 system may play a central role in stereognosis—the recognition of three-dimensional objects using touch only—as well as other perceptual tasks in which skin stretch is a major cue. Benoni Edin has shown that SA2 innervation in the hairy skin plays a substantial role in the perception of hand shape and finger position. The SA2 fibers aid the perception of finger joint angle by detecting skin stretch around the knuckles. The Ruffini endings near these joints are aligned such that different groups of receptors are stimulated as the fingers move in specific directions (Figure 23–4A). In this manner the SA2 system provides a neural representation of skin stretch over the entire hand, a proprioceptive rather than exteroceptive function.

Rapidly Adapting Fibers Detect Motion and Vibration

The RA1 receptor organ, the Meissner corpuscle, detects events that produce low-frequency, low-amplitude skin motion. This includes hand motion over the surface of objects, the detection of microscopic surface features, and low-frequency vibration. RA1 fibers contribute to detection of Braille patterns because they sense the change in skin indentation as individual dots pass over their receptive fields (Figure 23–6). They can detect irregularities and bumps as small as 10 m. We use the sensitivity of RA1 fibers to motion to adjust grip force when we grasp an object.

The RA2 receptor, the Pacinian corpuscle, is the most sensitive mechanoreceptor in the somatosensory system. It is exquisitely responsive to high-frequency (30–500 Hz) vibratory stimuli, and can detect vibration of 250 Hz in the nanometer range (Figure 23–8). The buzzing sensation experienced when a tuning fork is pressed against the skin in a neurological examination is mediated by the synchronized firing of RA2 units. It is a useful measurement of dynamic sensitivity to touch, particularly in cases of localized nerve damage.

The Pacinian corpuscle's filtering and amplifying of high-frequency vibration allows us to feel conditions at the working surface of a tool in our hand as if our fingers themselves were touching the object under the tool. The clinician uses this exquisite sensitivity to guide a needle into blood vessel and to probe tissue stiffness. The auto mechanic can use vibratory sense to position wrenches on unseen bolts. We can write in the dark because we feel the vibration of the pen as it contacts the paper and transmits the resulting frictional forces from the surface roughness to our fingers.

Both Slowly and Rapidly Adapting Fibers Are Important for Grip Control

In addition to their role sensing the size, shape, and texture of objects, mechanoreceptors provide important information concerning the actions of the hand during skilled movements. Roland Johansson and Gören Westling used microneurography to determine the role of mechanoreceptors when objects are grasped in the hand. By placing microelectrodes in the median nerve, they were able to record the firing patterns of individual mechanosensitive fibers as an object was initially contacted by the fingers, grasped between the thumb and index finger, lifted, held above a table, lowered, and returned to the rest position.

They found that all four classes respond to grasp and that each fiber type monitors a particular function. The RA1, RA2, and SA1 fibers detect contact when an object is first touched (Figure 23–9). The SA1 fibers signal the amount of grip force applied by each finger, and the RA1 fibers sense how quickly the grasp is applied. The RA2 fibers detect the small shock waves transmitted through the object when it is lifted from the table and placed on another surface. We know when an object makes contact with the table top because of these vibrations and therefore can manipulate the object without looking at it. The RA1 and RA2 fibers cease responding after grasp is established. The SA2 fibers signal flexion or extension of the fingers during grasp or release of the object and thereby monitor the hand posture as these movements proceed.

Signals from the hand that report on the shape, size, and texture of an object are important factors governing the application of force during grasping. Johansson and his colleagues have shown that we lift and manipulate an object with delicacy—with grip forces that just exceed the forces that result in overt slip—and that the grip force is adjusted automatically to compensate for differences in the friction coefficient between surfaces. Subjects predict how much force is required to grasp and lift an object and modify the grip force based on the tactile information provided by SA1 and RA1 afferents. Objects with smooth surfaces are grasped more firmly than those with rough textures, properties coded by RA1 afferents during initial contact of the hand with an object. The significance of the tactile information in grasping is seen in cases of nerve injury or during local anesthesia of the hand; patients apply unusually high grip forces, and the coordination between the grip and load forces applied by the fingers is poor.

The information supplied by the RA1 receptors to monitor grasping actions is critical for grip control, allowing us to hold on to objects when perturbations cause them to slip unexpectedly. RA1 fibers are silent during steady grasp and usually remain quiet until the object is returned to rest and the grasp released. However, if the object is unexpectedly heavy or jolted by external forces and begins to slip from the hand, the RA1 fibers fire in response to the small tangential movements of the object. The net result of this RA1 activity is that grip force is increased by signals from the motor cortex.

As described in Chapter 22, the central touch system comprises the dorsal column tracts of the spinal cord, relay nuclei in the brain stem and thalamus, and a hierarchy of regions in the cerebral cortex (see Figure 22–11). Tactile information enters the cerebral cortex through the primary somatosensory cortex (S-I) in the postcentral gyrus of the parietal lobe.

The neurons in S-I are at least three synapses beyond the receptors in the skin. Their inputs represent information processed in the dorsal column nuclei, the thalamus, and the cortex itself. Each cortical neuron receives inputs arising from receptors in a specific area of the skin, and these inputs together are its receptive field. We perceive that a particular location on the skin is touched because a specific population of neurons in the brain is activated. This experience can be experimentally induced by electrical stimulation of the same cortical neurons.

The primary somatic sensory cortex comprises four cytoarchitectural areas: Brodmann's areas 3a, 3b, 1, and 2 (Figure 23–10A). These areas are extensively interconnected such that processing of sensory information at this level involves both serial and parallel processing. Neurons in areas 3a and 3b receive inputs from the lateral and medial zones of the ventral posterior nucleus of the thalamus and project their axons to areas 1 and 2 (Figure 23–10B). Areas 3b and 1 receive information from receptors in the skin, whereas areas 3a and 2 receive proprioceptive information from receptors in muscles, joints, and the skin.

Somatosensory information is conveyed in parallel from the four areas of S-I to higher centers in the cortex, including the secondary somatosensory cortex (S-II), the posterior parietal cortex, and the primary motor cortex.

Cortical Receptive Fields Integrate Information from Neighboring Receptors

The receptive fields of cortical neurons are much larger than those of mechanoreceptive fibers in peripheral nerves. For example, the receptive fields of SA1 and RA1 fibers innervating the fingertip are tiny spots on the skin, whereas those of the cortical neurons receiving these inputs cover an entire fingertip or several adjacent fingers (Figure 23–11). The receptive field of a neuron in area 3b represents a composite of inputs from 300 to 400 sensory nerve fibers.

Receptive fields in higher cortical areas are even larger, spanning functional regions of skin that are activated simultaneously during motor activity. These include the tips of several adjacent fingers, or both the fingers and the palm. Receptive fields of neurons in the posterior parietal cortex and S-II are often bilateral, including symmetric positions on the contralateral and ipsilateral hands.

The size and position of receptive fields on the skin are not fixed permanently but can be modified by experience or by injury to sensory nerves (see Chapter 17). Cortical receptive fields appear to be formed during development and maintained by simultaneous activation of the input pathways.

The receptive fields of cortical neurons usually have an excitatory zone surrounded by or superimposed upon inhibitory zones (Figure 23–12). Stimulation of regions of skin outside the excitatory zone may reduce the neuron's responses to tactile stimulation within the receptive field. Similarly, repeated stimulation within the receptive field may also decrease neuronal responsiveness because the excitability of the pathway is diminished by inhibition.

Inhibitory receptive fields created by feed-forward and feedback connections through interneurons in the dorsal column nuclei, the thalamus, and the cortex itself limit the spread of excitation. Inhibition generated by strong activity in one circuit reduces the output of nearby neurons that are only weakly excited. The inhibitory networks ensure that the strongest of several competing responses is transmitted, permitting a winner-take-all strategy. These circuits prevent blurring of tactile details such as texture when large populations of touch neurons are stimulated. In addition, higher centers in the brain use inhibitory circuits to focus attention on relevant information from the hand when it is used in skilled tasks, by suppressing unwanted, distracting inputs during such behaviors.

Neurons in the Somatosensory Cortex Are Organized into Functionally Specialized Columns

In a series of pioneering studies of the cerebral cortex, Vernon Mountcastle discovered that the cortex is organized into vertical columns or slabs. Each column is 300 to 600 m wide and spans all six cortical layers from the surface to the white matter. All of the neurons within a column receive inputs from the same local area of the receptor sheet and respond to the same class or classes of receptors (Figure 23–13). A column is therefore an anatomical structure that, by organizing inputs that convey related information on location and modality, comprises an elementary functional module of the cortex.

Although the receptive fields of the neurons comprising a column in the somatosensory cortex are not precisely congruent, they share a common center that is most clearly evident in layer IV. In addition, the neurons in a column usually process only one submodality, such as pressure or vibration. This is not surprising, as the somatosensory submodalities are conveyed in anatomically separate pathways, and within these pathways the different types of mechanoreceptive fibers are also segregated (see Chapter 22).

The columnar organization of the cortex is a direct consequence of the intrinsic cortical circuitry and the migration pathway of neuroblasts during cortical development. The pattern of connections within the cerebral cortex is oriented vertically, perpendicular to the cortical surface. Thalamocortical axons terminate mainly on clusters of stellate cells neurons in layer IV, the axons of which project vertically toward the surface of the cortex (Figure 23–13B). The apical dendrites and axons of cortical pyramidal cells are also oriented vertically, parallel to the stellate cell axons. Thus the thalamocortical input is relayed to a narrow column of pyramidal cells with apical dendrites that are contacted by the stellate cell axons. This means that the same information is relayed up and down through the thickness of the cortex in a columnar fashion.

The sensory properties computed in each column are output in parallel to multiple regions of the brain. Neurons in each of the six cortical layers project to specific targets. Horizontal connections within layers II and III link neurons in neighboring columns, allowing them to share information when activated simultaneously by the same stimulus. For this reason, neurons in these supragranular layers have larger receptive fields than neurons in layer IV. Neurons in layers II and III also project to other cortical regions, both on the same side of the brain and at mirror locations in the other hemisphere. These feed-forward connections to higher cortical areas allow complex signal integration, as detailed later in this chapter.

Neurons in layer V receive inputs from layers II and III in the same and adjacent columns; they project to subcortical structures, including the basal ganglia, the pontine and other brain stem nuclei, the spinal cord, and the dorsal column nuclei. Layer VI neurons project to the thalamus.

In addition to feed-forward signals of information from mechanoreceptors, recurrent signals from layers II and III of higher brain areas are provided to layer I in each column. These feedback signals originate not only in somatosensory areas of the brain, but also in sensorimotor areas of the posterior parietal cortex, frontal motor areas, limbic areas, and regions of the medial temporal lobe involved in memory formation and storage. The recurrent signals are thought to play a role in the selection of sensory information for cognitive processing (by the mechanisms of attention) and in short-term memory tasks. Feedback pathways may also be involved in the gating of sensory signals during motor activity.

Cortical Columns Are Organized Somatotopically

The columns within the primary somatic sensory cortex are arranged such that there is a complete somatotopic representation of the body in each of the four areas (3a, 3b, 1, and 2). The cortical map of the body corresponds roughly to the spinal dermatomes (see Figure 22–9). Sacral segments are represented medially, lumbar and thoracic segments centrally, cervical segments more laterally, and the trigeminal representation of the head at the most lateral portion of S-I cortex (Figure 23–14). Knowledge of the neural map of the body in the brain is important for localizing damage to the cortex from stroke or head trauma.

Another important feature of somatotopic maps is the amount of cerebral cortex devoted to each body part. A neural map of the body in the brain does not duplicate exactly the spatial topography of the skin. Rather, it has disproportionately large areas devoted to certain body regions, particularly the hand, foot, and mouth, and relatively smaller areas devoted to more proximal body parts (see Figure 16–6). Each part of the body is represented in proportion to its importance to the sense of touch. Thus more cortex is devoted to the fingers than to the entire trunk (Figure 23–14C).

The amount of cortical area devoted to a unit area of skin—called the cortical magnification —varies by more than a hundredfold across different body surfaces. It is closely correlated with the innervation density and thus the spatial acuity of the mechanoreceptors in an area of skin. The areas with greatest magnification—the lips, tongue, fingers, and toes—have tactile spatial thresholds of 0.5, 0.6, 1.0, and 4.5 mm, respectively.

Brain regions activated by tactile stimuli have been visualized with high-resolution optical imaging of intrinsic signals generated by capillary blood flow. Cortical responses to air puffs applied to each finger of a monkey's hand are strongest in narrow transverse bands across areas 3b and 1 (Figure 23–15). Although inputs from neighboring fingers are organized in sequential bands, there is significant overlap between them, especially for the areas activated by digits 3 through 5, which are generally used in concert. The thumb and index finger, which often operate independently, have distinct territories and occupy a slightly larger cortical area, reflecting their importance in hand function.

Jon Kaas and his collaborators noted that somatotopic maps are important neurologically because they allow functionally related neurons to interconnect efficiently. For example, horizontal connections between cortical columns representing adjacent fingers enable us to perceive the surfaces contacted by different fingers as a continuous surface. The proximity of neurons representing the hand and mouth is likely to be useful in feeding behaviors, particularly in humans and other primates that bring food to the mouth with their hands.

The body surface has at least 10 distinct neural maps in the parietal lobe: four in S-I, four in S-II, and at least two in the posterior parietal cortex. Microelectrode mapping studies in monkeys have shown that the four areas of S-I (3a, 3b, 1, and 2) each have a complete, separate somatotopic representation of the body surface specific to a particular somatic sensory modality (Figure 23–14B). Area 3a receives input primarily from muscle stretch receptors; area 3b receives input from both SA1 and RA1 fibers; area 1 receives input primarily from RA1 and RA2 fibers; and area 2 contains a map of both touch and proprioception (Figure 23–11A).

As a result, these regions mediate different aspects of somatic sensation. Areas 3b and 1 are involved in sensing the details of surface texture, whereas area 2 is responsible for sensing the size and shape of objects. These attributes of somatic sensation are further elaborated in S-II and the posterior parietal cortex, where they form the basis of cognitive acts (object discrimination) and motor acts (object manipulation).

As we have seen, the responses of mechanoreceptors to Braille dot patterns or embossed letters touched by the fingers faithfully encode the stimulus contours (see Figure 23–6) and specify precisely where on the hand touch occurs. The sharp sensory images encoded by receptors in the skin are preserved up to the first stage of cortical processing in area 3b.

As information flows toward higher-order cortical areas, specific combinations of stimuli or stimulus patterns are needed to excite individual neurons. Neurons in areas 1 and 2 of S-I are concerned with more abstract features than just the location of a tactile stimulus. Neurons with receptive fields that include more than one finger fire at higher rates when several fingers are touched, and thus are concerned with the size of objects held in the hand.

Signals from neighboring neurons are combined in higher cortical areas to discern global properties of objects such as their orientation on the hand, the direction of motion across the skin (Figure 23–16), and their shape. In general, cortical neurons are concerned with sensory features that are independent of the stimulus position in their receptive field, allowing the brain to represent patterns common to stimuli of a particular class.

A cortical neuron is able to detect the orientation of an edge or the direction of motion because of the spatial arrangement of the presynaptic receptive fields. The receptive fields of the excitatory presynaptic neurons are aligned along a preferred axis; a stimulus orientation that matches the alignment elicits a strong excitatory response. In addition, the receptive fields of inhibitory presynaptic neurons at one side of the excitatory fields reinforce the orientation and direction selectivity of the postsynaptic cell (Figure 23–17).

Convergent inputs from different sensory modalities allow neurons in higher cortical areas to detect the size and shape of objects. Whereas neurons in areas 3b and 1 respond only to touch, and neurons in area 3a respond to muscle stretch, many of the neurons in area 2 receive both inputs. Thus neurons in area 2 can integrate information on the hand posture used to grasp an object, the grip force applied by the hand, and the tactile stimulation produced by the object, and this integrated information allows us to recognize the object.

Cognitive Touch Is Mediated by Neurons in the Secondary Somatosensory Cortex

An S-I neuron's response to touch depends almost exclusively on input within the neuron's receptive field. This feed-forward pathway is often described as a bottom-up process because the receptors in the periphery are the principal source of excitation of S-I neurons.

Higher-order somatosensory areas receive information not only from peripheral receptors, but also are strongly influenced by top-down processes, such as goal setting and attentional modulation. Data obtained from single-neuron studies in monkeys, neuroimaging studies in humans, and clinical observations of patients with lesions in higher-order somatosensory areas suggest that the ventral and dorsal regions of the parietal lobe serve complementary functions in the touch system similar to the dorsal (what) and ventral (where) pathways of the visual system (see Chapter 25).

Like S-I, the S-II cortex contains four distinct anatomical subregions with separate maps of the body. In both humans and monkeys S-II is located on the upper bank and adjacent parietal operculum of the Sylvian or lateral fissure (see Figure 23–10A). The central zone—consisting of S-II proper and the adjacent parietal ventral area—receives its major input from areas 3b and 1, largely tactile information from the hand and face. A more rostral region, the parietal rostroventral area, receives information from area 3a about active hand movements (Figure 23–18). The most caudal somatosensory region of the lateral sulcus extends onto the parietal operculum. This region abuts the posterior pa integrating somatosensory and visual properties of objects.

Physiological studies indicate that S-II plays a key role in the use of touch to recognize objects placed in the hand. S-II neurons are essential for distinguishing spatial features such as shape and texture and temporal properties such as vibratory frequency. Although neurons in S-II respond to textures such as Braille dots, embossed letters, or gratings, their pattern of firing does not encode the spatial or temporal patterns of the stimuli. Instead, they simply fire at different rates for different patterns. Similarly, they do not represent vibration as periodic spike trains linked to the oscillatory frequency, as do the sensory fibers from the skin. Instead, their firing rates also depend on behavioral context or motivational state.

In an elegant study Ranulfo Romo and his colleagues compared responses of neurons in S-I, S-II, and various regions of the frontal lobe of monkeys while the animals performed a two-alternative forced-choice task (Figure 23–19). The animals were rewarded if they correctly recognized which of two vibratory stimuli was higher in frequency. Neurons in S-I faithfully represented the vibratory cycles each stimulus, firing a brief burst in response to each. In contrast, S-II neurons responded to the first stimulus with spike trains proportional to its frequency but responded to the second stimulus with a signal that combined the frequencies of both stimuli. Thus the same vibratory stimulus could evoke different firing rates in S-II, depending on whether the preceding stimulus was higher or lower in frequency.

Romo's group found that neurons in prefrontal and premotor cortex that receive inputs from S-II preserve a memory of the first stimulus, continuing to fire after it ends. They proposed that the memory signal is fed back to S-II from these higher brain regions, modifying the response of S-II neurons to the direct tactile signals from the hand. In this manner sensorimotor memories of previous stimuli influence sensory processing in the brain, allowing us to make cognitive judgments about newly arriving tactile stimuli.

S-II is the gateway to the temporal lobe via the insular cortex. Regions of the medial temporal lobe, particularly the hippocampus, are vital to the storage of explicit memory (see Chapter 67). We do not store in memory every scintilla of tactile information that enters the nervous system, only that which has some behavioral significance. In light of the demonstration that the firing patterns of S-II neurons are modified by selective attention, S-II could make the decision whether a particular bit of tactile information is remembered.

Active Touch Engages Sensorimotor Circuits in the Posterior Parietal Cortex

The posterior parietal regions surrounding the intraparietal sulcus play an important role in the sensory guidance of movement rather than in discriminative touch. In monkeys these regions include areas 5 and 7 and in humans they include the superior parietal (Brodmann's areas 5 and 7) and inferior parietal cortex (areas 39 and 40).

Neurons in medial portions of area 5 receive postural information from SA2 fibers in the skin overlying the wrist, elbow, and shoulder, from receptors deep in these joints, and from muscle spindles that provide information about the movements of the arm. In monkeys these neurons are particularly sensitive when the animal extends its hand to grasp an object. Other cells located more laterally in area 5 integrate tactile and postural information from the hand. They respond most vigorously when the monkey shapes the hand in anticipation of grasping an object or plucks food morsels from a small container.

Neurons in area 7 of the monkey integrate tactile and visual stimuli that overlap in space and thus play an important role in eye-hand coordination. They respond more vigorously when the monkey is able to observe its hand while manipulating objects of interest than when simply looking at the object or handling it in the dark. Firing patterns in area 7 are also correlated with the hand posture used to grasp different objects.

It was originally believed that areas 5 and 7 were higher-order somatosensory areas that processed tactile information for object recognition and shape discrimination, and used proprioceptive signals for internal representations of integrated postures of the limbs. These theories were supported by anatomical data that showed that the principal sensory inputs to area 5 originate in area 2 of S-I cortex. However, these ideas about the role of areas 5 and 7 had to be altered following the discovery by Vernon Mountcastle, Juhani Hyvärinen, and others that these areas are involved in motor control.

In fact, during reaching and grasping, neural activity in the posterior parietal cortex coincides with activation of neurons in motor and premotor areas of the frontal cortex and precedes activity in S-I. There is strong evidence that area 5 receives convergent central and peripheral signals that allow it to compare central motor commands with peripheral sensory feedback during reaching and grasping behaviors. Sensory feedback from the hand to the posterior parietal cortex is used to confirm the goal of the planned action, thereby reinforcing a previously learned skill or correcting those plans when errors occur.

Predicting the sensory consequences of hand actions is an important component of active touch. For example, when we view an object and reach for it, we predict how heavy it should be and how it should feel in the hand; we use such predictions to initiate grasping. Daniel Wolpert and Randy Flanagan have proposed that during active touch the motor system controls the afferent flow of somatosensory information so that subjects can predict when tactile information should arrive in S-I and reach consciousness. Convergence of central and peripheral signals allows neurons to compare planned and actual movements. Corollary discharge from motor areas to somatosensory regions of the cortex may play a key role in active touch. It provides posterior parietal neurons with information on intended actions, allowing these neurons to compare planned and actual neural responses to tactile stimuli. Such mechanisms may explain why it is so difficult to tickle oneself.

Loss of tactile sensation in the hand produces significant motor as well as sensory deficits. Local anesthesia of the hand provides a direct way to appreciate the sensorimotor role of touch.

Loss of touch does not cause paralysis or weakness because much of skilled movement is predictive, relying on sensory feedback for adjustment if necessary. The motor system compensates for the absence of tactile information by generating more force than necessary. Under local anesthesia hand movements are clumsy and poorly coordinated, and force generation during grasping is abnormally slow. With the loss of tactile sensibility one is completely reliant on vision for directing the hand.

These motor problems are further exacerbated by long-term, chronic loss of tactile function because of injury to peripheral nerves or dorsal column lesions. Deafferentation produces major changes in the afferent connections in the brain (see Chapter 18), as do certain diseases. Myelinated afferent fibers in the dorsal columns degenerate in patients with demyelinating diseases, such as multiple sclerosis. In late-stage syphilis the large-diameter neurons in the dorsal root ganglia are destroyed (tabes dorsalis). These patients have severe chronic deficits in touch and proprioception but often little loss of temperature perception and nociception. The somatosensory losses are accompanied by motor deficits: clumsy and poorly coordinated movements and dystonia.

Similar impairments occur in patients with damage to S-I caused by stroke or head trauma, or following surgical excision of the postcentral gyrus. The severity and extent of the deficit depends on the site of the lesion. The sensory and motor deficits resulting from lesions in various parts of the parietal lobe have been compared in clinical studies.

Patients with lesions in the anterior parietal cortex have severe difficulties with simple tactile tests—touch thresholds, vibration and joint position sense, and two-point discrimination (Figure 23–20A). The patients also perform poorly on more complex tasks, such as texture discrimination, stereognosis, and visual-tactile matching tests. Motor deficits are less pronounced than sensory losses, particularly during tests of force and position control. Exploratory movements and skilled tasks, such as catching a ball or pinching small objects between the fingertips, are also somewhat abnormal.

In contrast, patients with lesions in the posterior parietal cortex have only mild difficulty with simple tactile tests. However, they have profound difficulty with complex tactile recognition tasks, and use few exploratory and skilled movements (Figure 23–20B). They display kinematic deficits when interacting with objects, failing to shape and orient the hand properly to grasp them, and misdirect the arm during reaching. They use too much grip force when an object is placed in their hand and are unable to direct the fingers properly when asked to evaluate its size and shape. These deficits are described clinically as the "useless hand" syndrome (tactile apraxia).

Humans with lesions localized to S-II also cannot perform complex tactile discrimination tasks such as stereognosis, but the deficits appear to be cognitive rather than sensorimotor. Mel Goodale and David Milner found that patients that are unable to discriminate the size of objects in clinical psychophysical tests are able to use visual information to match the shape of the hand to that of an object they must grasp. They propose that the dorsal parietal areas bordering the intraparietal sulcus serve a sensorimotor function, guiding hand movements in handling objects. The ventral somatosensory areas bordering the lateral sulcus serve a cognitive function, allowing us to recognize and name objects and thereby remember them.

Studies of sensory deficits in human patients are complicated by the fact that disease states or trauma rarely produce a "clean" lesion confined to one localized brain area. For this reason analyses of experimentally controlled lesions in animals have been useful for understanding the etiology of the sensory deficits observed in human patients. For example, macaque monkeys with a lesion of the cuneate fascicle show chronic losses in tactile discrimination, such as higher touch thresholds, impaired vibration sense, and poor two-point discrimination. They also display major deficits in the control of fine finger movements during grooming, scratching, and manipulation of objects.

A similar deficit in the execution of skilled movements can be produced experimentally in monkeys by inhibiting the neurons in the hand-representation region of area 2. The animal has great difficulty coordinating movements of the fingers because tactile feedback is absent (Figure 23–21).

Experimental ablation of somatosensory areas of the cortex has provided valuable information about the function of these areas. Small lesions limited to area 3b produce major deficits in touch sensation from a particular part of the body. Lesions in area 1 produce a defect in the assessment of the texture of objects, whereas lesions in area 2 alter the ability to differentiate the size and shape of objects. The resulting damage to tactile function is less severe when such lesions are made in infant animals, apparently because in the developing brain S-II cortex may take over functions normally assumed by S-I.

Removal of S-II cortex in monkeys causes severe impairment in the discrimination of both shape and texture and prevents the animals from learning new tactile discriminations. Ablation of area 5 produces deficits in roughness discrimination but few other alterations in passive touch. However, motor performance is severely compromised as these animals misdirect reaches to objects and fail to preshape the hand to grasp objects skillfully.

The similarity between the impairments observed in humans and monkeys is an important basis for understanding clinical losses of somatosensory function. We shall learn in later chapters that lesioning studies of other cortical areas in monkeys have also provided insight into higher-order sensory and motor functions of the brain.

When we explore an object with our hands a large part of the brain may become engaged by the sensory experience, by the thoughts and emotions it evokes, and by motor responses to it. These sensations result from the parallel actions of multiple cortical areas engaged in feed-forward and feedback networks.

At the first touch the peripheral sensory apparatus deconstructs the object into tiny segments, distributed over a large population of approximately 20,000 sensory nerve fibers. The SA1 system provides high- fidelity information about its spatial structure that is the basis of form and texture perception. The RA1 system conveys information about motion of the object in the hand, which enables us to manipulate it skillfully. The RA2 receptors transmit information about vibration of objects that allows us to use them as tools. The SA2 system provides information about the hand conformation and posture during grasping and other hand movements.

The information from these mechanoreceptors is conveyed to consciousness by the dorsal column fiber tracts of the spinal cord, relay nuclei in the brain stem and thalamus, and a hierarchy of intracortical pathways. By analyzing patterns of activity across the entire population, the brain constructs a neural representation of the object and the actions of the hand.

Central pathways have four important functions as they process the information from these mechanoreceptors. First, they convey the information provided by the receptors to the cognitive mechanisms responsible for perception. For this function the organization of these pathways is critical. A central theme is specificity: Computations of different kinds are kept separate by parallel pathways and by segregation of function at synaptic relays within single pathways.

Computations in these pathways are complex and accomplished serially, beginning in the dorsal column nuclei, progressing through the thalamus and several cortical stages, and terminating in regions of the medial temporal cortex concerned with memory and perception and in motor areas of the frontal lobe that mediate voluntary movements.

The brain's processing of touch is aided by the somatotopic organization of the neurons involved at each relay. Adjacent skin areas that are stimulated together are linked anatomically and functionally in central relays. Body parts that are especially sensitive to touch—the hands, feet, and mouth—are represented in large areas of the brain, reflecting the importance of tactile information conveyed from these regions.

The second function of the central pathways is transformation of the disaggregated representation of object properties in thousands of neurons to an integrated representation of complex object properties in a few neurons. Convergent excitatory connections between neurons representing neighboring skin areas and intracortical inhibitory circuits enable higher-order cortical cells to integrate global features of objects. In this manner the somatosensory areas of the brain represent properties common to particular classes of objects.

A third function is regulating the afferent flow of somatosensory information. The peripheral fibers deliver much more information than can be handled at any one moment; the central neural pathways compensate by selecting information for delivery to the mechanisms of perception and memory. Recurrent pathways from higher brain areas modify the ascending information provided by touch receptors, thus fitting the stream of sensory information to previous experience and task goals.

Finally, the touch system provides information necessary for the control and guidance of movement. Interactions between sensory and motor areas of parietal and frontal cortex provide a neural mechanism for predicting the sensory consequences of motor behaviors and for skill learning from repeated experience.