What characteristic of fast nerve fibers permits them to transmit signals perceived as sharp immediate pain?

Merkel nerve endings

These mechanoreceptor nerve endings are responsible for providing information regarding pressure and texture and are classified as slowly adaptive type of mechanoreceptors. These nerve endings also have wide distribution in the human skin. These nerve endings are structurally rigid and are not encapsulated, which causes them to have a sustained response to mechanical deflection of the tissue of less than 1 μm. Due to the sustained response to pressure these nerve endings are classified as slowly adapting. Merkel nerve ending is the most sensitive mechanoreceptor to vibrate at low frequency (within 5–15 Hz) [7].

4.4.2 Mechanoreceptors

Mechanoreceptors are sensory receptors that respond to mechanical deformation of the receptor or surrounding tissue. Mechanoreceptors are involved in hearing, detection of equilibrium, skin tactile sensing, deep tissue sensing, and sensing of arterial pressure.

Hearing or audition involves the transduction of sound waves into neural signals via mechanoreceptors in the inner ear. As shown in Fig. 4.4, sound vibrations from the outer ear are mechanically conducted from the eardrum through a series of tiny bones to the inner ear. Sound vibrations from the middle ear enter the cochlea at the basal end of the cochlea. The cochlea in the inner ear has over 30,000 hair cells that amplify the incoming acoustic signal and encode the frequency content. Outer hair cells primarily amplify sound vibrations, while inner hair cells detect those vibrations and excite the nerve fibers of the cochlear or auditory nerve. The basal end of the cochlea encodes the higher end of the audible frequency range while the apical end of the cochlea encodes the lower end of the frequency range.

The inner ear also hosts vestibular mechanoreceptors that excite the vestibular neurons of the auditory nerve. The vestibular mechanoreceptors communicate a sense of balance and spatial orientation. The vestibular system contains semicircular canals that provide information about rotational motions. Angular accelerations produce relative movements of endolymph in the semicircular canals. The movement of fluid pushes on the cupula in the ear which contains mechanosensitive hairs cells that transduce the mechanical signal into an electrical signal.

Touch or somatic sense receptors are located in the dermis, the bottom layer of skin. There are about 20 different types of nerve endings in the dermis. They can be activated by movement (mechanoreceptor), pressure (mechanoreceptor), chemical (chemoreceptor), and/or temperature (thermoreceptors). Another activation method is from the vibrations generated as a finger moves across a surface and feels these vibrations. The vibration sensing is directly related to function of lamellar (or Pacinian) corpuscles, which are one of the four major types of mechanoreceptors.

Baroreceptors are mechanoreceptors located in the carotid sinus and in the aortic arch that help to regulate arterial blood pressure. Baroreceptors are stimulated by changes in arterial pressure. Increases in pressure stretch or strain the baroreceptors and transmit signals to the central nervous system. The parasympathetic nervous system is activated while the sympathetic nervous system is inhibited in response to baroreceptor activation. Inhibition of the sympathetic system leads to a drop in peripheral resistance while the parasympathetic activation increases the vagal tone on the sinoatrial (SA) node and decreases the heart rate and contractility. The net effect is that blood pressure is decreased. Conversely, blood pressure can be increased with sympathetic activation and parasympathetic inhibition.

Ventricular and Coronary Receptors

Ventricular mechanoreceptors, which are attached to nonmyelinated afferent nerves, are probably of little importance in normal cardiovascular control because of the high pressures required to change their activity. However, it had been suggested, because some receptors respond to increased force of contraction (particularly when the ventricle is underfilled), that they may initiate the so-called ‘vasovagal’ response. This is the abrupt dilatation of blood vessels and cardiac slowing that cause people to faint in response to low cardiac filling or emotional stimuli. More recent work suggests that this is unlikely, because experiments in which ventricular filling was reduced and the cardiac sympathetic nerves were stimulated, did not induce the response (Figure 14). Furthermore, a similar response can occur in patients with heart transplants and therefore ventricles without a nerve supply.

Following an obstruction of a coronary artery (coronary thrombosis), the damaged region of the heart may become distended. This causes the stimulation of ventricular mechanoreceptors in the region, with consequent decreases in heart rate and blood pressure. However, experimental findings have failed to define a physiological, day-to-day role for ventricular mechanoreceptors.

Ventricular chemosensitive afferents may play a role in the responses occurring during coronary artery narrowing. When the blood supply to the heart muscle is inadequate, particularly when its work is increased during exercise, various chemicals accumulate which stimulate chemosensitive afferent nerves. Stimulation of sympathetic afferent nerves mediates cardiac pain (angina pectoris), and may also lead to increases in heart rate and blood pressure. Stimulation of chemosensitive nerves running in the vagus, on the other hand, results in slowing of the heart and a fall in blood pressure. The resultant response seems to depend on the region of the heart that is affected: reduction in flow to the inferolateral wall usually decreases blood pressure; whereas when the anterior wall is affected, the pressure is more likely to increase. Extensively decreased myocardial blood flow, however, generally causes low blood pressure due to inadequate contraction of the left ventricle.

Coronary arterial mechanoreceptors probably function like the aortic and carotid sinus baroreceptors in that they detect changes in arterial blood pressure and bring about responses that tend to restore it to normal. Indeed, the unique characteristics of the coronary baroreceptors, including their resistance to resetting and prolonged central inhibition of sympathetic efferent activity, make them ideally suited as a regulator of mean arterial blood pressure.

6.4.4 Sensory Inputs

Most of the mechanoreceptors in the limbs of mammals are cutaneous or hair follicle receptors. These receptors are sporadically active during the locomotor step cycle. For example, hair follicle receptors and cutaneous receptors in the footpad fire transiently at the moment of ground contact at the onset of the stance phase. Hair follicle receptors covering the limb fire unpredictably in response to surface airflow (Prochazka, 1996). Cutaneous afferents generally do not have strong direct reflex actions on α-motoneurons, but they can initiate more global motor responses such as the corrective response to a trip (Prochazka et al., 1978) and transitions between the stance and swing phases of the locomotor step cycle (Rossignol et al., 2006). Most of the continuous sensory input during movement is provided by the proprioceptive afferents: Golgi tendon organs (TOs) and muscle spindles.

Broadly speaking, TOs signal muscle force and muscle spindles signal muscle length and velocity. Few of the group Ib afferents that arise from TOs are active in a resting muscle (Houk et al., 1971). They start firing when a specific force threshold is reached. This threshold varies widely between individual Ib afferents. They are more sensitive to variations in active force generated by motor units whose muscle fibers insert into the musculotendinous capsule of the receptor (Jansen and Rudjord, 1964). With the recruitment of each such motor unit, there is a stepwise increase in the firing rate of a Ib afferent, but these steps are smoothed out when the firing rates of ensembles of Ib afferents are summed, as they would be in the spinal cord as a result of the convergence of synaptic input from ensembles of Ib afferents onto spinal neurons (Crago et al., 1982). The summed Ib firing rate saturates at high force levels even though most of the contributing Ib afferents respond fairly linearly (Crago et al., 1982). This is because more Ib afferents are recruited at low forces than at high forces, leading to a power-law relationship between ensemble Ib firing rate and force (Lin and Crago, 2002; Mileusnic and Loeb, 2009).

Muscle spindle group Ia and II afferents respond both to length changes of their parent muscle and to the activity of fusimotor efferents emanating from the spinal cord (γ-motoneurons and β-motoneurons, the latter being α-motoneurons which innervate intrafusal as well as extrafusal muscle). The responses of spindle afferents to length changes both in the absence and presence of fusimotor action have been modeled in numerous studies of varying complexity spanning 50 years. The reader is referred to a recent review which provides the mathematical details of the main models, along with a discussion of their pros and cons (Prochazka, 2015). In the simpler spindle models, fusimotor action is represented by two parameters: gain and offset. The more complex models have up to 30 parameters representing fusimotor action, intrafusal mechanics, sensory adaptation, and so on. Because γ-motoneurons are small, very few researchers have managed to record from them in normally behaving animals. This is possible in decerebrate cats (Durbaba et al., 2003; Ellaway et al., 2002; Taylor et al., 2006) but in this case, descending drive is absent or abnormal, and there is strong nociceptive input from open surgical wounds, both of which affect γ-motoneurons. Because there is still no clear consensus on how fusimotor activity is controlled during locomotion and because fusimotor action plays a big role in the more complex models, it is unclear whether these models provide better predictions of muscle spindle afferent activity during normal locomotion than the simpler models.

Muscle spindle models have only been compared head-to-head in a single study of afferent activity recorded during normal locomotion in cats (Prochazka and Gorassini, 1998a, 1998b). Interestingly, in some muscles, more than 80% of the variance in spindle afferent firing was accounted for by the simpler models, presumably because fusimotor action did not vary much in these muscles during locomotion. In other muscles such as the ankle extensors, the predictions of the models were less satisfactory, even after presumed alpha-linked fusimotor action was added.

Because roboticists who design walking machines have been inspired just as much, if not more, by locomotor control in invertebrates versus that in vertebrates, it is worth pointing out that there are remarkable analogs of vertebrate muscle spindles and TOs in invertebrates. Crustacean thoracico-coxal muscle receptor organs have response properties similar to those of mammalian muscle spindles (Bassler, 1983, 1993). They have two types of sensory afferents, T and S, which are analogous to spindle Ia and II afferents (Bush, 1981). They transmit their nonspiking signals to the CNS electrotonically, but in all other respects their responses to length changes and their efferent control by Rml and Rm2 motoneurons (equivalent to γ and β fusimotor neurons) are astonishingly similar to those of muscle spindles. Campaniform sensilla in the external cuticle of many invertebrates show response characteristics comparable to mammalian TOs. Similar response characteristics have also been reported for other classes of invertebrate proprioceptors including locust forewing stretch receptors (Pearson and Ramirez, 1990) and cockroach femoral tactile spines (Pringle and Wilson, 1952).

Mechanoreceptors

There are two groups of mechanoreceptors: (i) encapsulated receptors, including Pacinian corpuscles, Meissner corpuscles, Krause endings and Ruffini endings, which are all innervated by fast conducting myelinated fibres; and (ii) receptors having an organized and distinctive morphology such as the hair follicle receptors and Merkle discs. Iggo (1988) classified each mechanoreceptor with a distinctive range of properties that enable it to receive and respond to a particular parameter of a mechanical stimulus. The Pacinian corpuscles detect and respond to high frequencies of displacement up to about 1500Hz, the Meissner corpuscles and the hair follicles to middle range frequencies (20–200Hz), and the Merkle cells and Ruffini endings to steadily maintained deformation of the skin (DC to 200Hz).

Mechanoreceptors

There are two groups of mechanoreceptors: (1) encapsulated receptors, including the Pacinian corpuscles, Meissner corpuscles, Krause endings and Ruffini endings that are all innervated by fast-conducting myelinated fibers; (2) receptors having an organized and distinctive morphology such as the hair follicle receptors and Merkel discs. Each mechanoreceptor has a distinctive range of properties that enable it to receive and respond to a particular parameter of a mechanical stimulus. The Pacinian corpuscles detect and respond to high frequencies of displacement up to 1500 Hz, the Meissner corpuscles and the hair follicles to middle range frequencies (20–200 Hz), and the Merkel cells and Ruffini endings to steadily maintained deformation of the skin.44 Toma and Nakajima88 investigated the responsiveness of mechanoreceptors in the glabrous skin of the hand. Thirteen single afferent activities were recorded from four kinds of mechanoreceptors. Both fast-adapting (FA) and slow-adapting (SA) units were sensitive to the vibratory stimuli. The relationship between the most sensitive frequency and applied pressure to the skin was analyzed as a tuning curve. FA-type I (FAI) was sensitive to vibratory stimuli at 30–40 Hz and the frequency which entrained one-to-one discharge at lower pressure was between 10 and 80 Hz. FA-type II (FAII) sensitivity was augmented sharply over 60–100 Hz. SA-type I (SAI) and SA-type II (SAII) responsiveness was almost the same, and characteristic sensitivity to the vibratory stimuli was revealed under 15 Hz.

What Is Proprioception and Why Is It Important?

Proprioceptive signals originate from mechanoreceptors within the muscles, tendons, and skin. Through a complex process of multisensory integration, they give rise to at least two important perceptual channels, namely, the sense of limb movement (or kinesthesia) and the sense of joint position; these sensations are crucial for bodily awareness, as well as voluntary motor control, regulation of muscle tone, and postural stability. The term proprioception comes from the Latin words “proprium” and “percipio” indicating the “perception of own self.”

Neuromotor impairments caused by neurological injuries or diseases (such as stroke, spinal cord injury, Parkinson's disease, and multiple sclerosis) are frequently associated with acquired deficits or loss of proprioceptive abilities. Proprioceptive deficits are common also after orthopedic injury and are often the cause of repetitive injury in sport players. These deficits compromise the ability to perform everyday activities since proprioception contributes to the control of posture, motion, and forces. Specifically, it has been demonstrated that the loss of kinesthetic sensation contributes to impaired control of reaching and stabilization behaviors that are vital to an independent life style [1–4]. Although people suffering loss of proprioceptive feedback can move by relying on vision, long processing delays inherent to the visual system (100–200 ms, [5]) yield movements that are typically slow, are poorly coordinated, and require a good deal of attention [6,7]. As a consequence, visually guided corrections may come too late and result in jerky, unstable movements [8]. Often, stroke survivors give up using their contralesional limb because of their sensorimotor deficits [9] even though this reduces the quality of life [2,10]. Proprioceptive deficits or losses might also interfere with motor learning processes [11–13], as well as with the motor outcome of rehabilitative treatment and the recovery after neurological injury [14–16]. Therefore, it is important to evaluate proprioceptive accuracy in a quantitative manner and to determine how proprioceptive deficits impact the ability to perform daily living activities.

Since proprioception results from complex cortical and subcortical processes that integrate sensory inputs from across the different peripheral receptors, measurable indicators of proprioceptive integrity are likely to be affected in the early stages of many neurological diseases, thus providing an opportunity for early diagnosis. Currently, standard neurological examination techniques lack sufficient precision and accuracy to reliably detect early-stage deficits of proprioception. By contrast, robot devices and their resolute measurement technologies are naturally suited for that purpose and for the twin objective of integrating assessment functions with assessment-based rehabilitation interventions.

2.4.4 The Difference Between Hot and Cold Tooth Pain

The mechanical threshold for pulpal nociceptors (mechanoreceptors) is calculated to be about 90 Pa [5]. To verify the model, fluid velocities have been adopted from the literature [1] and the corresponding shear stresses have been calculated, and are listed in Table 2.3.

Table 2.3. Fluid Flow Velocity and the Simulated Shear Stress

Figure 2.7 shows the relationship between the neural discharge rate (over 5 s) and the flow velocity. The experimental observations show that the nociceptive receptors respond in a significantly different manner to the DFF in different flow directions [1]. The neural discharge rate increases progressively as the outward flow velocity increases, above a threshold. In contrast, the associated receptors are much less sensitive to inward flow. The simulated results are in good agreement with the experimental results. The numerical simulations theoretically reveal that the OPD accounts for the difference in the responses of the intradental nerve to the inward and outward flows.

Cold stimuli (0 ∼ 5°C) induce outward flow velocities ranging between 531.2 and 849.9 μm/s [1,39], while the range is approximately 354.1–779.1 μm/s [1] for inward flow responding to hot stimuli (55°C). The inward and outward flow directions and their corresponding fluid velocity magnitudes are inconsistent with those employed as boundary conditions in the present work (Fig. 2.7). Although fluid velocities were employed as the boundary conditions in modeling the TB MSS and the subsequent neural discharge, the difference between hot and cold sensations still can be revealed.

In the case of cold stimuli (0 ∼ 5°C), a short latency (∼1 s) of the neural response can be observed [40,41]. During that stage, the local temperature (where the TB is located) is still far from being able to activate the thermal receptors [41]. Therefore, it seems unlikely that the rapid response to the cold stimulation originates from thermo-sensitive receptors. Note that fluid flow could be detected before a temperature change in DEJ, and the latency of the initiation of the DFF (<1s) [1,12] (induced by either hot or cold stimuli) corresponds to the latency of the neural response. In addition, the flow velocity induced by cold stimuli may easily exceed the threshold [1,39] for activating the mechanoreceptors (Fig. 2.7). Therefore, the initial stage of cold-induced tooth pain (sharp, shooting pain) may involve the activation of mechano-sensitive receptors by the DFF. It should be mentioned that tooth pain (dull, burning pain) after a long latency (∼30 s) of cold stimulation may be attributed to the activation of cold-sensitive nociceptors [43,44].

A neural response can only be detected after 10 s of hot stimuli (55°C) [40,41]. Prior to this, no neural discharge signal can be detected [1,40,41]. This does not contradict with the conclusion that the DFF may evoke the neural response, since hot stimuli cannot initiate the high rate of DFF [1] needed to activate the mechano-sensitive receptors. It is possible that after such a long latency, the temperature around the thermally sensitive receptors reaches the threshold [41], triggers the receptors and causes pain sensation [40,41].

Cutaneous Receptors Include Mechanoreceptors, Thermoreceptors, and Nociceptors

The Surface of the Body Contains a Variety of Mechanoreceptors

The skin contains a variety of mechanoreceptors including Pacinian corpuscles, Meissner’s corpuscles, Ruffini’s corpuscles, Merkel’s disks, and free nerve endings in the skin and surrounding hair follicles. All of these are long receptors, in which the receptor or generator potential is developed within the sensory cell and it fires action potentials based on this receptor potential without an additional connection to a sensory cell. Their axons are all myelinated, so that the CNS is informed quickly whenever any of these receptors detects a stimulus (see Figure 4.3.3).

Figure 4.3.3. Highly diagrammatic representation of the different mechanoreceptors in the skin. The glabrous or nonhairy skin, shown at left, contains a different set of mechanoreceptors from the hairy skin, as shown at the right. Both contain a variety of receptors.

The Pacinian corpuscles consist of a free nerve ending enclosed by a layered capsule, much like an onion. They lie in the subcutaneous layers of the skin, are rapidly adapting, and respond best to vibration. This response is due to the mechanical properties of the capsule, as the frequency response is removed when the capsule is removed.

Meissner’s corpuscles reside in the dermis just below the epidermis. These also rapidly adapt and are thought to respond to fluttering types of stimuli. That is, Meissner’s corpuscles also respond to vibration but at a lower frequency than Pacinian corpuscles.

Merkel’s disks are located in the dermis and have small receptive fields. They are slowly adapting and respond to steady touch-pressure on the skin.

Ruffini’s corpuscles are located in the dermis and are slowly adapting. They have much larger receptive fields and so they may participate both in touch-pressure sense and in proprioceptive sense by detecting the push or pull of skin from one segment of the body on another.

All of the encapsulated nerve endings listed above are present in both the hairy and nonhairy (glabrous) skin. Free nerve endings are present in the epidermis in glabrous skin. They contribute to the tactile sense. They also wrap around hair follicles in the hairy skin, where they detect movement of the hair.

Thermoreceptors Consist of Cold Receptors and Warm Receptors

Cold receptors are free nerve endings with thin myelinated fibers, whereas the warm receptors are free nerve endings with unmyelinated axons with low conduction speeds. They differ from the mechanoreceptors in that they exhibit tonic level of activity at most temperatures. They respond to temperature changes with a phasic component followed by a tonic component that depends on the temperature. This is why when you get into a hot bath it feels hot for a while and then it feels warm. The phasic component indicates the change in temperature upon immersion in the hot water and the tonic component indicates that it is still warm. Much of this perception is due to central processing of the peripheral thermoreceptor input (see Figure 4.3.4).

Figure 4.3.4. Response of thermoreceptors to skin temperature. The skin temperature was held constant at the indicated temperature, while the frequency of action potentials was recorded from fibers representing each of the thermoreceptor types.

5.2.1.2 Perceptual models for human tactile processing

Texture perception is mediated by at least two different mechanoreceptors of the human skin. On the one hand, Pacinian corpuscles are responsible for sensing finer textural surface elements, while on the other hand, coarser surfaces are processed by SA I, i.e., the slowly adapting type I mechanoreceptive afferents [390]. Most empirical research in vibrotactile perception has focused on Pacinian corpuscles, which are sensitive to vibrations in the frequency range from 65 Hz to 400 Hz with the highest sensitivity being around 250 Hz [391]. In a series of experiments, Verrillo and colleagues [392–396] determined the absolute vibrotactile threshold of detectability of sinusoidal stimuli on different locations of the hand. Thereby the minimum stimulus intensity that is consciously perceptible by a human observer serves as the detection threshold. In other words, vibrotactile frequencies below this threshold lack reportability, as they cannot be felt or detected by human observers.

Another important feature of human perception that needs to be considered when developing tactile codecs is the overlay of perceptual effects from other stimuli. Past studies have shown empirical evidence for masking in the perception of vibrotactile signals. The masking effect has been widely studied in sensory perception and describes the phenomenon when the perception of a target stimulus is reduced by the presence of another stimulus, referred to as a mask. This phenomenon occurs in visual, auditory, as well as tactile perception. In vibrotactile perception, on the one hand, masking can occur in the time domain and is referred to as temporal masking. On the other hand, masking effects can also occur in the frequency domain, and is referred to as spectral masking. Furthermore, stimulus competition and stimulus intensity can result in masking that determines the perceptibility of competing tactile stimuli [397–399].

In the context of vibrotactile codec development, stimulus signals consist of various texture patterns, therefore in the following, we will only focus on spectral masking. The authors in [400] observed such masking phenomena for simplified stimuli, where the target consisted of vibrotactile outputs of pure tonal sinusoids in the frequency range of interest ranging from 80 Hz to 380 Hz, whereas narrowband noise (120 Hz, 200 Hz, and 280 Hz) served as the mask. Hence, it can be assumed that masking will also occur for more complex signals, resulting in an increase of perception thresholds around dominant peaks [401]. Thus perceptual thresholds and masking effects can be employed for the design of a vibrotactile codec, as it permits emphasizing perceptible frequency components of a given vibrotactile signal, while penalizing less perceptible ones, or completely filtering out imperceptible information [402]. Rather than spectral stimulus properties, which play a determining role in auditory perception and thus in audio codec design, perceptual processing of tactile information is largely attributable to temporal cues [397,403]. Consequently, temporal cues, such as temporal duration, temporal stimulus delays, and temporal masking effects should be given consideration during the design, development, and evaluation of tactile codecs. The authors in [404] found that Pacinian-mediated texture perception can be predicted by the intensity-based spectral power model, which includes temporal and intensity information, rather than the frequency theory [390]; whereas coarser textural features are coded by spatial variation elicited in SA I afferent firing [390]. Recently, the authors of [405] introduced a time segmented intensity-based model that accounts for relatively slow time-variant vibration patterns. In summary, in the design of vibrotactile codecs (and the subsequent evaluation through quality assessment procedures), these above discussed perceptually relevant features of vibrotactile signal processing, especially temporal and spectral information, are determining factors. Moreover, other factors that affect vibrotactile perception that are relevant especially for codec evaluation with human observers include individual differences between human observers [406,407], body locations [408], cognitive states [397,409], and task situations [400,410].

In the future, we will build such multilayered perceptual models based on psychophysical and neurocognitive experiments in simplified scenes. Additionally, human sensory perception is complex and multimodal. That is, utilizing solely somatosensory perceptual models, as put forth here, might have limitations in comprising effective codecs that represent human perception. We believe multisensory interaction effects, specifically in the context of how another sensory modality, such as vision or audition might influence kinesthetic or tactile perception, merit further investigation. This is of particular importance to provide a truly immersive human experience with promising results for virtual object interaction and teleoperation. Therefore it will be important to investigate and analyze the relationships between physical perceptual stimulus properties and subjective ratings in perceptual quality assessments. Our work will contribute to the design, development, and improvement of the Tactile Internet with Human-in-the-Loop, including both kinesthetic and tactile subsystems.