Encoding of Object Curvature by Tactile Afferents From
Human Fingers
A. W. GOODWIN, 1 V. G. MACEFIELD, 2 AND J. W. BISLEY 1
1
Department of Anatomy and Cell Biology, University of Melbourne, Parkville, Victoria 3052; and 2 Prince of Wales
Medical Research Institute, Randwick, New South Wales 2031, Australia
INTRODUCTION
The opposable thumb and fine motor control that are characteristic of the primate hand allow precise manipulation of
held objects. Sensory feedback from the hand allows features
of the object such as its shape and roughness to be assessed
and this information is in turn important for successful operation of the motor control system. Although receptors in the
muscles and digital joints provide relevant information to
the CNS (Burke et al. 1988; Edin 1990; Edin and Vallbo
1990; Gandevia et al. 1992), specialized receptors in the
glabrous skin of the hand are of paramount importance in
signaling ‘‘small-scale’’ features of an object such as its
surface texture and the fine details of its shape (Goodwin
1997).
There have been many studies of cutaneous afferents innervating the fingers in both humans and monkeys (for review see Darian-Smith 1984; Macefield 1997) but so far
studies related specifically to the encoding of the shape of
handled objects have been restricted to monkeys. LaMotte
and his colleagues have studied the responses of these afferents to a variety of shaped stimuli including shaped steps
indented into and scanned across the fingers (LaMotte and
Srinivasan 1987a,b; Srinivasan and LaMotte 1987), ellipsoids scanned across the skin (LaMotte et al. 1994) and
cylindrical profiles scanned across the skin (LaMotte and
Srinivasan 1996). We have examined the responses of cutaneous afferents in monkeys to spheres contacting the fingerpad (Goodwin et al. 1995; Wheat et al. 1995). Such
studies in monkeys enabled detailed quantitative descriptions
of the responses of the cutaneous mechano-receptive afferents. Comparison with human psychophysical experiments
has pointed to likely neural codes for signaling the stimulus
parameters to the brain (Goodwin et al. 1991). This raises,
once again, the question of how valid it is to compare such
psychophysical measurements in humans with equivalent
neural recordings from monkeys.
In human glabrous skin, four types of specialized mechanoreceptor terminals have been identified histologically.
Merkel cell-neurite complexes and Meissner corpuscles are
located superficially and Ruffini endings and Pacinian corpuscles are located more deeply (Bell et al. 1994; Chouchkov 1973; Halata 1975; Miller et al. 1958). Microelectrode
recordings from the median and ulnar nerves have revealed
four distinct functional classes of low-threshold mechanosensitive afferent fibers. Two classes adapt slowly to a sustained indentation of the skin and are termed slowly adapting
type I or II afferents (SAI or SAII), respectively; two classes
adapt rapidly and are termed fast-adapting type I or II afferents (FAI or FAII), respectively (Vallbo and Johansson
1984). Type I afferents possess small, well-defined receptive
fields; type II afferents have larger fields that are less clearly
defined. A prominent feature of the SAIIs is their capacity
to respond to lateral skin stretch, with many possessing the
property of directional sensitivity and with some exhibiting a
spontaneous regular discharge at rest (Johansson and Vallbo
1983). The relationship between the morphological classes
of receptors and the functional classes of afferents has been
established by a variety of largely indirect observations in
0022-3077/97 $5.00 Copyright q 1997 The American Physiological Society
/ 9K22$$DE09 J458-7
11-07-97 11:14:10
neupa
LP-Neurophys
2881
Downloaded from http://jn.physiology.org/ by 10.220.33.3 on June 17, 2017
Goodwin, A. W., V. G. Macefield, and J. W. Bisley. Encoding
of object curvature by tactile afferents from human fingers. J.
Neurophysiol. 78: 2881–2888, 1997. Isolated responses were recorded from fibers in the median nerves of human subjects by
using microneurography. Mechanoreceptive afferent fibers with receptive fields on the fingerpads were selected. The fingers were
immobilized and spherical stimuli were applied passively to the
receptive field with a contact force of 40-, 60-, or 80-g weight.
The radii of the spheres were 1.92, 2.94, 5.81, or 12.4 mm or `
(flat); the corresponding curvatures, given by the reciprocal of the
radii, were 694, 340, 172, 80.6, or 0 m01 , respectively. When the
spheres were applied to the receptive field center of slowly adapting
type I afferents (SAIs), the response increased as the curvature of
the sphere increased and also increased as the contact force increased. All SAIs behaved in the same way except for a scaling
factor proportional to the sensitivity of the afferent. When a sphere
was located at different positions in the receptive field, the shape
of the resulting response profile reflected the shape of the sphere;
for more curved spheres the profile was higher and narrower (increased peak and decreased width). Slowly adapting type II afferents (SAIIs) showed different response characteristics from the
SAIs when spheres were applied to their receptive field centers.
As the curvature of the stimulus increased from 80.6 to 172 m01 ,
the response increased. However, further increases in curvature did
not result in further increases in response. An increase in contact
force resulted in an increase in the response of SAIIs; this increase
was proportionately greater than it was for SAIs. For SAIIs, the
shape of the receptive field profile did not change when the curvature of the stimulus changed. For fast-adapting type I afferents
(FAIs), the responses were small and did not change systematically
with changes in curvature or contact force. Fast-adapting type II
afferents (FAIIs) did not respond to our stimuli. Human SAIs,
FAIs, and FAIIs behaved like monkey SAIs, FAIs, and FAIIs,
respectively. The response of the SAI population contains accurate
information about the shape of the sphere and its position of contact
on the finger and also indicates contact force. Conversely, whereas
SAIIs possess a greater capacity to encode changes in contact force,
they provide only coarse information on local shape.
2882
A. W. GOODWIN, V. G. MACEFIELD, AND J. W. BISLEY
METHODS
Thirty-seven experiments were performed on 22 healthy human
volunteers (11 female, 11 male; age 18–49 yr). The procedures
were approved by the Committee on Experimental Procedures Involving Human Subjects at the University of New South Wales.
Subjects sat in a comfortable chair with the supinated forearm
supported and the dorsum of the hand embedded in plasticine. The
fingers were held flat by clips affixed to the nails and anchored in
the supporting material. Care was taken to ensure that the fingers
were mechanically stable but that the glabrous skin was not distorted and that the receptive fields of the units were unencumbered
by the plasticene.
Recording procedures
The location of the median nerve at the wrist, Ç1-cm proximal
to the flexor retinaculum, was determined by external electrical
stimulation by using a 1 mm diam probe. Once the optimal site for
eliciting paraesthesia in the digits was found, an insulated tungsten
microelectrode (type TM33B20, World Precision Instruments) was
inserted through the skin. The electrode was directed manually into
a cutaneous fascicle of the median nerve while delivering weak
negative pulses ( õ1 mA, 0.2 ms, 1 Hz) through the microelectrode
/ 9K22$$DE09 J458-7
via an optically isolated constant-current stimulator (MacLab 8/
s & Isolated Stimulator, ADInstruments, Sydney, Australia). An
adjacent subdermal microelectrode, from which 1 mm of insulation
had been removed, served as the reference electrode. After entering
a fascicle of digital nerve origin (paraesthesia elicited at õ0.025
mA), the microelectrode was connected to an amplifier (Gain
10000, band-pass 0.3–5 kHz). A search was then undertaken for
large spikes originating from single cutaneous afferent axons while
the fascicular innervation territory was manipulated and stroked.
Isolated units were studied only if their receptive fields were located on the relatively flat portion of the volar surface of the fingers.
Preference was given to receptive fields on the pulp of the index
finger or middle finger.
An isolated unit was classified as SAI, SAII, FAI, or FAII according to the established criteria of responses to sustained indentation, sensitivity to rapidly changing stimuli, size of receptive field,
and sensitivity to lateral skin stretch. The mechanical threshold
was measured by using a set of calibrated von Frey hairs (SimmsWeinstein aesthesiometers, Stoelting, Chicago, IL) and the receptive field was mapped with a hair of Ç4 times threshold. The
mapping was done as carefully as possible because this procedure
was used to determine the receptive field center. Neural activity
was sampled at 12.8 kHz by using the SC/ZOOM data acquisition
and analysis software (Department of Physiology, University of
Umeå, Umea, Sweden) and stored directly on disk for subsequent
analysis.
Stimulating procedures
The primary stimuli consisted of spherical surfaces made from
delrin with radii of 1.92, 2.94, 5.81, or 12.4 mm or ` (flat); the
corresponding curvatures, given by the reciprocal of the radii, are
694, 340, 172, 80.6, and 0 m01 , respectively. These surfaces were
applied to the skin at the receptive field center by a mechanical
stimulator that has been described previously (Goodwin et al.
1995). Briefly, it consisted of a balanced beam that, once released
by activating a restraining solenoid, applied the surfaces at a constant static force of 40-, 60-, or 80-g weight (0.392, 0.588, or 0.784
N, respectively) determined by adding or removing weights from
the beam. A damper regulated the motion of the beam and set the
initial rate of application at 20 mm/s. At contact, the skin surface
was orthogonal to the direction of application of the surfaces. Stimuli were presented in blocks of trials. For each trial the stimulus
remained in contact with the skin for 1.5 s and was then lifted off
the skin for 3 s before the next trial commenced. A block consisted
of 15 trials. The contact force was successively 40-, 60-, or 80-g
weight and at each force the curvature was successively 80.6, 80.6,
172, 340, and 694 m01 . For each force the first application at 80.6
m01 served as a lead stimulus to minimize interaction from the
previous force and was excluded from the analysis. The complete
block was repeated a further five times to allow assessment of
reproducibility of the responses. For three units, a curvature of 0
m01 was used in place of the curvature of 80.6 m01 .
Some units were sufficiently stable to allow field studies as well.
In these, an x-y micrometer on the stimulator was used to change
the contact point in the receptive field in 1-mm steps, either along
the long axis of the finger or at right angles to it. A block of trials
consisted of the sequence of curvatures 172, 172, 340, and 694
m01 applied at a contact force of 60-g weight to successive points,
separated by 1 mm, from one end of the receptive field to the
other. Again the whole block was repeated five times and at each
point the lead stimulus (the 1st application at 172 m01 ) was excluded from analysis. Alternatively, the field studies were performed by stepping delrin cylinders with the sequence of curvatures
172, 172, 340, and 694 m01 along the long axis of the finger in
0.5-mm steps at a contact force of 60-g weight. The axis of the
cylinder was orthogonal to the long axis of the finger.
11-07-97 11:14:10
neupa
LP-Neurophys
Downloaded from http://jn.physiology.org/ by 10.220.33.3 on June 17, 2017
several species, but it is generally accepted that SAIs, SAIIs,
FAIs, and FAIIs innervate Merkel complexes, Ruffini endings, Meissner corpuscles, and Pacinian corpuscles, respectively (Darian-Smith 1984).
Although Ruffini endings have been found in monkey
hairy skin (Biemesderfer et al. 1978), they have not been
found in the glabrous skin of these species. Corresponding
to this, the functional class of SAIIs has been found in hairy
skin in monkeys but not in their glabrous skin (Harrington
and Merzenich 1970). The remaining three receptor types
and afferent classes that are found in humans are also found
in monkeys. Moreover, the innervation density of SAIs,
FAIs, and FAIIs on the fingers is similar for humans and
monkeys (Darian-Smith and Kenins 1980; Johansson and
Vallbo 1979).
In most of the experiments that provide data that allow
quantitative comparison of the properties of FAIs, FAIIs,
and SAIs in humans and monkeys, the stimulus was a punctate probe. For these stimuli there is a close correspondence
in the properties of the afferents in the different species (for
review see Darian-Smith 1984). A more rigorous comparison has been made with braille dots and embossed dot arrays
scanned across the fingerpad and here too the correspondence was close (Connor et al. 1990; Johnson and Lamb
1981; Phillips et al. 1990, 1992). There have been no recordings from human afferents responding to stimuli with
controlled variation in local shape analogous to the studies
in the monkeys described.
The current study was undertaken principally to address
two specific questions. First, do SAIs, FAIs, and FAIIs on
the central portion of the human fingerpad respond to local
curvature and to contact force in a manner similar to those
in the monkey? Second, how do SAIIs in this region, which
are not present in the monkey, respond to the same stimuli?
Thus we used the same passively applied spherical stimuli
that were employed in the monkey studies, so that direct
comparisons could be made (Goodwin et al. 1995; Wheat
et al. 1995).
RESPONSES OF HUMAN TACTILE AFFERENTS TO CURVED SURFACES
FIG . 1. Schematic finger showing locations of receptive field centers for
41 units studied [15 slowly adapting type I afferents (SAIs), 9 slowly
adapting type II afferents (SAIIs), 10 fast-adapting type I afferents (FAIs),
and 7 fast-adapting type II afferents (FAIIs)]. ●, m, and ., units included
in pooled data; s, units used for corroboration only; m and ., units referred
to specifically in text.
Data analysis
RESULTS
Afferent sample
Isolated responses were recorded from a total of 41 units
consisting of 15 SAIs, 9 SAIIs, 10 FAIs, and 7 FAIIs. Over
one-half of the afferents supplied the index finger, with the
middle and ring fingers being represented by 22 and 24%
of the sample, respectively; only one unit was recorded from
the thumb. A number of compromises had to be made in
this study. Ideally we would have liked to record from a
large population of each class of afferent innervating the
central region of the fingerpad, with stable recording for
long periods. However, because of the difficulties inherent in
human microneurography, the units we studied had receptive
field centers shown on a schematic finger in Fig. 1. Because
SAIs and FAIs have confined receptive fields, the centers
could be determined easily; however, for SAIIs and FAIIs
the fields were more extensive and the centers shown are
more equivocal.
Responses of SAIs to stimuli applied to the receptive field
center
The first experimental protocol was designed to characterize the responses of SAIs to local curvature. Figure 2 shows
experimental records from a typical SAI (Fig. 1, m, receptive
field location). The degree of signal isolation is illustrated
in Fig. 2A by the spike train in response to a sphere of
curvature 80.6 m01 (radius, 12.4 mm) applied to the receptive field center with a contact force of 40-g weight;
the corresponding instantaneous frequency of discharge is
shown in Fig. 2B. In Fig. 2E the response, measured by the
total number of action potentials occurring in the 1st s of
the response, is shown for all four curvatures and all three
/ 9K22$$DE09 J458-7
contact forces. The response increased as the curvature of
the stimulus increased and also increased as the contact force
increased. The standard errors show the variability in responses with the complete stimulus sequence applied six
times over a period of Ç7 min. Responses of single human
afferents were highly consistent.
Increasing the curvature from 80.6 m01 in Fig. 2B to
172 m01 in Fig. 2C resulted in an increase in instantaneous
frequency throughout the time course of the response. For
each of these two responses a normalized instantaneous frequency was calculated by dividing the instantaneous frequency by the mean instantaneous frequency; the normalized
responses are overlaid in Fig. 2D. The two normalized profiles are similar, showing that to a first approximation, the
entire waveform is scaled by the change in the stimulus.
Thus the form of the stimulus-response functions shown in
Fig. 2E for responses measured over 1 s were much the
same for responses measured over 0.2, 0.3 (Fig. 2F), 0.5,
and 0.8 s and over the last 0.5 s.
All SAIs showed the same characteristic increase in response as the curvature of the stimulus increased and as
the contact force increased. However, different fibers had
different sensitivities as evident from their different response
magnitudes to the same stimulus. In Fig. 3, the 10 SAIs
with receptive fields on the fingerpad (Fig. 1, ● ) have been
combined after normalizing to remove the sensitivity factor
(Goodwin et al. 1995). For each afferent the normalizing
factor was its average response to the nine stimuli common
to all the fibers (curvatures 694, 340, and 172 m01 , at forces
40-, 60-, and 80-g weight). The most sensitive fiber was 2.3
times as sensitive as the least sensitive fiber as measured by
the ratio of the normalizing factors. The small standard errors
in Fig. 3 reflect the high degree of consistency of the curvature-response and force-response functions from fiber to fiber. The five SAIs excluded from the pool (because they
were not on the fingerpad or were too close to a crease) also
showed increasing responses with increasing curvature and
with increasing force, but the shapes of the functions were
slightly different, presumably because of the different skin
mechanics in these regions.
The 120 data points forming Fig. 3 fit well to the function
y Å k[a 0 b exp( 0 cx)]. The values of the constants a, b,
and c are 2.04, 1.95, and 0.00236, respectively and for contact forces of 40-, 60- and 80-g weight the values of the
constant k have ratios 1:1.08:1.15, respectively (F Å 451,
P ! 0.001) (see Goodwin et al. 1995 for details of the fitting
procedure).
Responses of FAIs and FAIIs to stimuli applied to the
receptive field center
The FAIs responded with a small number of action potentials during the dynamic phase of the stimulus. Even though
the responses were small, they were highly consistent for
the six repeated applications. Responses for the seven afferents in Fig. 1 ( ● ) are shown in Fig. 4 for all four curvatures
at a contact force of 60-g weight. There was a trend for
responses to increase with increases in curvature but the
changes were small and not systematic from fiber to fiber.
Changing the contact force to 40- or 80-g weight had a
negligible effect on responses.
11-07-97 11:14:10
neupa
LP-Neurophys
Downloaded from http://jn.physiology.org/ by 10.220.33.3 on June 17, 2017
The unitary integrity of each recording was carefully checked
off-line using the spike recognition software incorporated into SC/
ZOOM. For each trial, the time of occurrence of the first action
potential after contact of the surface with skin was determined and
was taken as the commencement of the response. The following
parameters were extracted for further analysis: 1) instantaneous
frequency throughout the response; 2–6) the number of action
potentials occurring in the first 0.2, 0.3, 0.5, 0.8, and 1.0 s of the
response, respectively; and 7) the number of action potentials in
the final 0.5 s of response (a measure of the static response).
2883
2884
A. W. GOODWIN, V. G. MACEFIELD, AND J. W. BISLEY
None of the FAIIs responded at all to our stimuli, presumably
because of the low indentation velocity (see DISCUSSION ).
Receptive field profiles of SAIs
For some units the spatial characteristics of the receptive
field were defined by comparing the responses of the unit
for stimuli applied at several different positions in the field.
FIG . 3. Normalized responses for 10 SAI units (mean { SE). For the
9 stimuli at curvatures 172, 340, and 694 m01 (forces 40-, 60-, and 80-g
weight), there are 10 data points each; for curvature 80.6 m01 there are 9
data points each. For each fiber, normalizing factor was average value at
the 9 points (3 curvatures 1 3 forces) common to all 10 fibers. ●, responses
for single fiber at 0 m01 .
/ 9K22$$DE09 J458-7
In Fig. 5A the responses of a typical SAI (Fig. 1, . ) are
shown for spheres with three different curvatures applied at
points in the receptive field separated by 1 mm along a line
through the receptive field center and parallel to the long
axis of the finger. As the curvature of the surface increased,
the profile became sharper, with an increase in the magnitude
of the peak and a decrease in the width of the profile. Such
profiles were collected for three additional afferents and
showed similar behavior. In addition, profiles for cylinders
FIG . 4. Responses of the 7 FAIs (Fig. 1, ● ). Spheres with 4 different
curvatures (80.6, 172, 340, and 694 m01 ) were applied at a contact force
of 60-g weight. Mean values for 6 applications of stimulus are shown. For
clarity, SEs (n Å 6) are only shown on one line, but their magnitude is
typical of all 7 afferents.
11-07-97 11:14:10
neupa
LP-Neurophys
Downloaded from http://jn.physiology.org/ by 10.220.33.3 on June 17, 2017
FIG . 2. Responses of a typical SAI
(Fig. 1, m ) to spheres of different curvature
applied to receptive field center. A and B:
spike train and instantaneous frequency, respectively for a sphere of curvature 80.6
m01 (radius, 12.4 mm) applied with a contact force of 40-g weight. Time 0 is defined
by the occurrence of 1st spike. C: instantaneous frequency for a sphere of curvature
172 m01 . D: normalized profiles from B
(shaded) and C (open). E: response measured by number of impulses in 1 s
(mean { SE, n Å 6). F: same as E except
that response is number of impulses in 0.3
s. For E and F the 4 curvatures were 80.6,
172, 340, and 694 m01 and the 3 contact
forces were 40-, 60-, and 80-g weight.
RESPONSES OF HUMAN TACTILE AFFERENTS TO CURVED SURFACES
2885
FIG . 5. Receptive field profiles contrasted
for an SAI and an SAII. A: SAI (Fig. 1, . ).
Responses (mean { SE; n Å 6) are shown
at points, separated by 1 mm, along a line
through center of receptive field and parallel
to long axis of finger. More distal stimuli have
greater abscissa values. Contact force was 60g weight. B: SAII (Fig. 1, . ). Responses
(mean { SE; n Å 6) are shown along a line
through receptive field center and orthogonal
to long axis of the finger. Abscissa values are
greater for more ulnar positions. Contact force
was 60-g weight.
Responses of SAIIs
As has been reported by others, we encountered many
SAIIs with clearly isolated action potentials that had receptive fields near the nails, on the edges of the fingers, or
that had poorly defined receptive fields and were presumably
located in deeper tissues. However, in this study we wished
to record only from those SAIIs that had receptive fields
located on the volar surfaces of the fingers, preferably on
the fingerpads. These afferents responded vigorously to our
stimuli.
When spheres with different curvatures were applied at a
number of contact forces to the estimated center of the receptive field, there were two striking differences in the responses of SAIIs compared with the responses of SAIs described above. These are illustrated in Fig. 6A by the responses of the SAII (Fig. 1, m ), which is typical of our
sample. Like the other classes of afferents described, the
responses of the SAIIs were highly consistent with repeated
stimuli. Comparing Figs. 6A and 2E, the first obvious difference is the effect of the curvature of the stimulus. For SAIIs,
the response increased when the curvature increased from
80.6 to 172 m01 , but further increases in curvature had a
negligible effect on the responses.
The second difference is the effect of force. For SAIIs,
increases in force resulted in relatively larger increases in
response than was the case for SAIs. For example, at a
curvature of 694 m01 , increasing the contact force from 40to 80-g weight resulted in a 40% increase in response for
the SAII in Fig. 6A but only a 14% increase for the SAI in
Fig. 2E. Responses of the five afferents (Fig. 1, ● ) were
combined in Fig. 6C after normalizing to remove differences
in scale because of differences in sensitivity among the afferents. The small standard errors confirm that all five afferents
have similar stimulus-response functions. As was the case
for the SAIs, if a change in the stimulus caused a change in
the response of an SAII, then the change was seen throughout
the time course of the response. Thus the stimulus-response
functions, illustrated in Fig. 6A for the number of impulses in
FIG . 6. Response characteristics of SAII afferents. A and B: spheres applied to the receptive field center of the SAII (Fig.
1, m ). Response measure is the number of impulses in the 1st s ( A) or 1st 0.3 s (B) of the response (mean { SE; n Å 6).
C: combined normalized responses (mean { SE; n Å 5) for spheres applied to receptive field centers of the 5 units shown
by solid symbols in Fig. 1. For each fiber, normalizing factor was average value at 9 points (curvatures 694, 340, and 172
m01 1 forces of 40-, 60-, and 80-g weight).
/ 9K22$$DE09 J458-7
11-07-97 11:14:10
neupa
LP-Neurophys
Downloaded from http://jn.physiology.org/ by 10.220.33.3 on June 17, 2017
with different curvatures were measured for six of the SAIs
and showed the same behavior of an increase in peak response and a decrease in profile width with an increase in
curvature.
2886
A. W. GOODWIN, V. G. MACEFIELD, AND J. W. BISLEY
DISCUSSION
Microneurography has been used to study the responses
of isolated single units in human peripheral nerves since
1967 (Hagbarth and Vallbo 1967). The major advantage of
recording from humans is that the data can be related directly
to human behavior, circumventing the issues of species differences that are usually encountered. However, experiments
that require long periods of stable recording with immobilized fingers, or that seek to define the properties of a specific
subset of peripheral nerves, can be done more satisfactorily
in monkeys than in humans. Thus a vast amount of our
knowledge about the tactile system rests on comparisons of
psychophysical experiments in humans and neural recordings from monkeys. Direct comparison of some of the
properties of SAIs, FAIs, and FAIIs in humans and monkeys
is possible from experiments where the stimulus was an
indenting or vibrating probe (for review see Darian-Smith
1984), braille scanned across the fingertips (Johnson and
Lamb 1981; Phillips et al. 1990), or patterns of embossed
dots scanned across the fingertips (Connor et al. 1990; Phillips et al. 1992). In all cases the properties of human and
monkey afferents were remarkably similar.
Experiments in monkeys have highlighted the importance
of the local curvature of a stimulus in determining the responses of cutaneous afferents (Goodwin et al. 1995; LaMotte and Srinivasan 1987a,b). Corresponding experiments
on human afferents described in this paper show similar
properties for the SAIs, FAIs, and FAIIs that occur in both
humans and monkeys. The most responsive afferents to our
stimuli were the SAIs. Their responses increased when the
curvature of a sphere applied to the receptive field center
increased and also when the contact force increased. When
the stimuli were located at different positions in the receptive
field, the shape of the receptive field profiles reflected the
shape of the stimulus; for a more curved surface the peak
of the profile increased and its width decreased. The FAIs
responded with a small number of action potentials and their
responses did not change in a systematic way with changes
in the curvature or contact force of a sphere applied to the
receptive field center. The shapes of their receptive field
response profiles did not reflect the curvature of the stimulus.
FAIIs did not respond to the stimuli.
/ 9K22$$DE09 J458-7
Comparison of SAI responses in humans and monkeys
It was shown previously that SAIs in both humans and
monkeys are sensitive to edges, which are regions of high
curvature (Johansson et al. 1982; Phillips and Johnson
1981a). The experiments reported here show that the entire
curvature-response function is similar for human and monkey SAIs. For spheres applied to the receptive field center
of human SAIs, the response was of the form k[2.04 0
1.95 exp( 00.00236x)] with the constant k having ratios
1:1.08:1.15 for contact forces of 40-, 60-, and 80-g weight
respectively. For monkeys, the corresponding function is
k[1.91 – 1.62 exp( 00.00243x)], with the constant k having
ratios 1:1.28:1.53 for contact forces of 10-, 15-, and 20-g
weight, respectively (Goodwin et al. 1995). The shape of the
exponential function is almost identical in the two samples,
indicating a close similarity in the way local shape is signaled by the SAI populations in humans and monkeys. The
smaller force factor in humans than in monkeys suggests
that SAIs on the center of the fingerpad are not that crucial
for indicating contact force in humans where SAIIs may
play a role. For both primates there are other sources of
information on contact force. The receptor mechanisms operating in Merkel complexes and the nature of the skin mechanics must be similar in the different primate species.
These observations boost our confidence in using nonhuman
primate data for fundamental mechanisms such as transduction (Looft 1994) and skin mechanics (Srinivasan 1989)
when explaining human tactile performance. There are, however, some qualifications. The most obvious relates to the
size of the finger that is much larger in humans than in
monkeys.
There are at least four ways in which finger size may play
an important role. First, larger fingers have larger pads so
that contact areas are generally larger and corresponding
forces need to be higher. We used a range of 40- to 80-g
weight in humans and a range of 10- to 20-g weight in
monkeys; the resulting magnitudes of the responses of the
afferents in humans and in monkeys were similar. For patterns of dots scanned across the fingers, Connor et al. (1990)
argued that a contact force of 100-g weight in the human
was equivalent to a contact force of 30-g weight in the
monkey. Phillips et al. (1990) scanned braille characters
across the skin at a contact force of 60-g weight in humans;
Johnson and Lamb (1981) used contact forces of 20-g
weight and 60-g weight with monkeys. Second, when objects
do contact the more curved parts of the fingers, the greater
curvatures in the smaller monkey fingers may lead to a
greater sensitivity to features of the stimulus like local gaps
or edges. Third, the afferents innervating the curved edges
of the fingers may show quantitative differences because
the curvatures of the fingers depend on their size. Fourth,
particularly with higher forces (for example as used in lifting
objects that weigh several hundred g), the mechanics of the
finger as a whole may be different. In the experiments here
and in our curvature experiments in the monkey we have
confined our studies to the population of afferents innervating the relatively flat portion of the fingerpad; here the relative size difference is less of an issue.
A critical reason for establishing the similarity of human
and monkey afferent responses to spheres was to determine
11-07-97 11:14:10
neupa
LP-Neurophys
Downloaded from http://jn.physiology.org/ by 10.220.33.3 on June 17, 2017
the first second of response, were similar when the response
measure was the number of impulses occurring in the first
0.2, 0.3 (illustrated in Fig. 6B), 0.5, or 0.8 s or in the
last 0.5 s.
Despite the fact that SAII receptive fields are more diffuse
than those of the SAIs, the receptive field profiles were
distinctly defined and highly repeatable. Fig. 5B shows the
profile for the unit (Fig. 1, . ) mapped with spheres positioned along a line at right angles to the long axis of the
finger. Responses decreased as the stimulus moved away
from the receptive field center but, unlike the SAIs, the shape
of the profile was essentially independent of the curvature
of the sphere. These characteristics were common to the four
afferents mapped with spheres or cylinders. We do not have
sufficient data to make a meaningful comparison of the
widths of the profiles in the two classes of slowly adapting
afferents.
RESPONSES OF HUMAN TACTILE AFFERENTS TO CURVED SURFACES
the validity of our population response reconstructions,
based on monkey data, for explaining human tactile performance (Goodwin 1997). The data here support our hypothesis that the shape of the sphere is represented in the shape
of the SAI population response. Its position on the skin is
represented in the position of the activity profile within the
SAI population and, to a lesser degree, the contact force is
represented in the overall magnitude of activity in the SAI
population.
Responses of SAIIs
/ 9K22$$DE09 J458-7
SAIIs, cannot do? Although the answers to these questions
are not known at present, the properties of SAIIs, including
their sensitivity to lateral stretch, suggest a major role in
sensing tangential loads while lifting and manipulating objects. If monkeys were able to perform as well as humans
in such tasks they would have to be using different neural
processing strategies based only on the SAIs, FAIs, and
FAIIs. However it is more likely that the evolution of hand
function has reached a greater degree of sophistication in
humans than in monkeys. Most of the data on monkeys result
from simple behavioral observations and, although it is clear
that they have considerable manual dexterity, it is not clear
that they can do everything that humans can do (Christel
1993). For example, could a monkey execute tasks equivalent to writing with a pen or performing delicate surgical
maneuvers?
Responses of FAIs and FAIIs
As was the case in the monkey, the FAIs in the human
responded to the spheres but with small responses that did
not change systematically with changes in curvature or contact force. They are unlikely to play a major role in signaling
these stimulus parameters to the brain. The FAIIs did not
respond at all to our stimuli, nor did the FAIIs examined in
the monkey with the same stimuli (Goodwin et al. 1995).
This may seem surprising because the initial contact with
the skin was at a velocity of 20 mm/s and it has been shown
that FAIIs respond at even lower indentation velocities in
both humans and monkeys (Knibestol 1973; Lindblom and
Lund 1966). However, our spherical stimuli are larger than
most of the punctate probes commonly used and do not have
changes in curvature that occur at the edges of flat probes.
More importantly, the forces used by us were set by gravity
so that, although the initial velocity was 20 mm/s, contact
with the skin would have resulted in a progressive reduction
in velocity during the indentation phase. The velocity thresholds quoted lower were measured with stimulators that produced constant velocities throughout the indentation.
Limitations of our data
The restrictions of human microneurography imposed a
number of limitations on our data. It is not possible to search
for more than Ç4 h to find an afferent with a desired receptive field location, nor is it possible to record from the
afferents for many hours routinely. In addition, the human
hand and fingers cannot be immobilized for long periods
with the same stability as in anesthetized monkeys. Thus
our sample of afferents is, of necessity, a compromise. For
the SAIs and FAIs the size of the pool on the fingerpads is
large enough for quantitative estimates of population responses to spheres with different curvatures applied to the
receptive field center. A detailed analysis of the receptive
field profiles of the SAIs (as was done in the monkey) is
not possible. The data (Fig. 5A) suggest a strong similarity
for profiles in humans and monkeys but we are not justified
in making quantitative comparisons of factors like the widths
of the receptive field profiles. Because our aim with the
SAIIs was to study only those afferents with receptive fields
clearly on the fingerpad, our population is not large. Never-
11-07-97 11:14:10
neupa
LP-Neurophys
Downloaded from http://jn.physiology.org/ by 10.220.33.3 on June 17, 2017
SAIIs have not been found in the monkey finger but they
occur in large numbers in the human finger (Johansson and
Vallbo 1979). Their large receptive fields suggest that sensitivity to local curvature would be minimal; however, this
was not certain particularly for afferents supplying the central portion of the fingerpad. The SAIIs we recorded showed
a small increase in response when the curvature of the sphere
changed from 80.6 to 172 m01 (radius, 12.4–5.81 mm) but
further increases in curvature had no effect on the response.
This contrasts with the properties of SAIs and adds further
information relating to the differences in receptor mechanisms of Merkel complexes and Ruffini organs (Khalsa et
al. 1996; Phillips and Johnson 1981b). In terms of human
sensation, our data reinforce the notion that SAII responses
are unlikely to signal information to the brain about the local
shape of an object in contact with the skin.
The SAIIs show a systematic increase in response with
increases in contact force and this increase is proportionately
larger than for SAIs. Thus the SAII population may signal
information to the brain about forces between the fingerpads
and contacted objects. Because the responses depend on
force but not on local shape, the signals could be easily
interpreted directly. In contrast with the local shape of an
object and its position of contact on the fingerpad, which
can only be signaled by the cutaneous afferents, there are a
number of sources that can signal contact force in addition
to the SAIs and SAIIs on the fingerpad. These include responses from the large number of afferents innervating the
remaining portion of the digits and, in addition, during active
manipulation, responses of Golgi tendon organs and efference copy (Gandevia et al. 1992).
We have intentionally studied the subpopulation of SAIIs
located on the fingerpad. There are a large number of SAIIs
located at some distance from a sphere contacting the fingerpad and these may respond because they are sensitive to
lateral stretch (Johansson and Vallbo 1983). The quantitative details of the responses of this subpopulation and the
possible role of direction sensitivity are unknown. It is interesting to note that microstimulation of single SAIIs does not
evoke any sensation in awake humans, whereas stimulation
of single SAIs does (Macefield et al. 1990; Torebjörk et al.
1987). The relevance of this observation to the role of whole
population responses is not clear. However, it is known that
microstimulation of a single muscle spindle afferent does
not generate any perceptual response, whereas activation of
many such afferents generates a clear illusion of movement
(Macefield et al. 1990).
Two intriguing questions remain. What is the role of SAIIs
in humans and what can humans do that monkeys, lacking
2887
2888
A. W. GOODWIN, V. G. MACEFIELD, AND J. W. BISLEY
theless the data are consistent and valuable in suggesting
possible roles for these afferents that have not been found
in monkeys. Because of the difficulty of finding receptive
fields on the center of a fingerpad, a small number of afferents on the middle or proximal phalanxes or close to an
interphalangeal crease were studied. These afferents were of
some value in that their properties were consistent with those
on the fingerpads; however, they were not used in any pooled
data analysis.
This investigation was supported by grants from the National Health and
Medical Research Council of Australia.
Address reprint requests to A. W. Goodwin.
Received 3 June 1997; accepted in final form 12 August 1997.
BELL, J., BOLANOWSKI, S., AND HOLMES, M. H. The structure and function
of pacinian corpuscle: a review. Prog. Neurobiol. 42: 79–128, 1994.
BIEMESDERFER, D., MUNGER, B. L., BINCK, J., AND DUBNER, R. The piloruffini complex: a non-sinus hair and associated slowly-adapting mechanoreceptor in primate facial skin. Brain Res. 142: 197–222, 1978.
BURKE, D., GANDEVIA, S. C., AND MACEFIELD, G. Responses to passive
movement of receptors in joint, skin and muscle of the human hand. J.
Physiol. (Lond.) 402: 347–361, 1988.
CHOUCHKOV, C. N. The fine structure of small encapsulated receptors in
human digital glabrous skin. J. Anat. 114: 25–33, 1973.
CHRISTEL, M. Grasping techniques and hand preferences in Hominoidea.
In: Hands of Primates, edited by H. Preuschoft and D. J. Chivers. Vienna:
Springer-Verlag, 1993, p. 91–108.
CONNOR, C. E., HSIAO, S. S., PHILLIPS, J. R., AND JOHNSON, K. O. Tactile
roughness: neural codes that account for psychophysical magnitude estimates. J. Neurosci. 10: 3823–3836, 1990.
DARIAN-SMITH, I. The sense of touch: performance and peripheral neural
processes. In: Handbook of Physiology. The Nervous System. Sensory
Processes. Bethesda, MD: Am. Physiol. Soc., 1984, sect. 1, vol. III, p.
739–788.
DARIAN-SMITH, I. AND KENINS, P. Innervation density of mechanoreceptive
fibres supplying glabrous skin of the monkey’s index finger. J. Physiol.
(Lond.) 309: 147–155, 1980.
EDIN, B. B. Finger joint movement sensitivity of non-cutaneous mechanoreceptor afferents in the human radial nerve. Exp. Brain Res. 82: 417–422,
1990.
EDIN, B. B. AND VALLBO, Å. B. Muscle afferent responses to isometric
contractions and relaxations in humans. J. Neurophysiol. 63: 1307–1313,
1990.
GANDEVIA, S. C., MCCLOSKEY, D. I., AND BURKE, D. Kinaesthetic signals
and muscle contraction. Trends Neurosci. 15: 62–65, 1992.
GOODWIN, A. W. Extracting the shape of an object from the responses of
peripheral nerve fibers. In: Neural Aspects of Tactile Sensation, edited
by J. W. Morley. Amsterdam: Elsevier. In press.
GOODWIN, A. W., BROWNING, A. S., AND WHEAT, H. E. Representation of
curved surfaces in responses of mechanoreceptive afferent fibers innervating the monkey’s fingerpad. J. Neurosci. 15: 798–810, 1995.
GOODWIN, A. W., JOHN, K. T., AND MARCEGLIA, A. H. Tactile discrimination of curvature by humans using only cutaneous information from the
fingerpads. Exp. Brain Res. 86: 663–672, 1991.
HAGBARTH, K.-E. AND VALLBO, Å. B. Mechanoreceptor activity recorded
percutaneously with semi-microelectrodes in human peripheral nerves.
Acta Physiol. Scand. 69: 121–122, 1967.
HALATA, Z. The mechanoreceptors of the mammalian skin. Ultrastructure
and morphological classification. Adv. Anat. Embryol. Cell Biol. 50: 1–
77, 1975.
HARRINGTON, T. AND MERZENICH, M. M. Neural coding in the sense of
touch: human sensations of skin indentation compared with the responses
/ 9K22$$DE09 J458-7
11-07-97 11:14:10
neupa
LP-Neurophys
Downloaded from http://jn.physiology.org/ by 10.220.33.3 on June 17, 2017
REFERENCES
of slowly adapting mechanoreceptive afferents innervating the hairy skin
of monkeys. Exp. Brain Res. 10: 251–264, 1970.
JOHANSSON, R. S., LANDSTROM, U., AND LUNDSTROM, R. Sensitivity to
edges of mechanoreceptive afferent units innervating the glabrous skin
of the human hand. Brain Res. 244: 27–32, 1982.
JOHANSSON, R. S. AND VALLBO, Å. B. Tactile sensibility in the human hand:
relative and absolute densities of four types of mechanoreceptive units
in glabrous skin. J. Physiol. (Lond.) 286: 283–300, 1979.
JOHANSSON, R. S. AND VALLBO, Å. B. Tactile sensory coding in the glabrous
skin of the human hand. Trends Neurosci. 6: 27–32, 1983.
JOHNSON, K. O. AND LAMB, G. D. Neural mechanisms of spatial tactile
discrimination: neural patterns evoked by Braille-like dot patterns in the
monkey. J. Physiol. (Lond.) 310: 117–144, 1981.
KHALSA, P. S., HOFFMAN, A. H., AND GRIGG, P. Mechanical states encoded
by stretch-sensitive neurons in feline joint capsules. J. Neurophysiol. 76:
175–187, 1996.
KNIBESTOL, M. Stimulus-response functions of rapidly adapting mechanoreceptors in the human glabrous skin area. J. Physiol. (Lond.) 232: 427–
452, 1973.
LAMOTTE, R. H. AND SRINIVASAN, M. A. Tactile discrimination of shape:
responses of slowly adapting mechanoreceptive afferents to a step stroked
across the monkey fingerpad. J. Neurosci. 7: 1655–1671, 1987a.
LAMOTTE, R. H. AND SRINIVASAN, M. A. Tactile discrimination of shape:
Responses of rapidly adapting mechanoreceptive afferents to a step
stroked across the monkey fingerpad. J. Neurosci. 7: 1672–1681, 1987b.
LAMOTTE, R. H. AND SRINIVASAN, M. A. Neural encoding of shape: responses of cutaneous mechanoreceptors to a wavy surface stroked across
the monkey fingerpad. J. Neurophysiol. 76: 3787–3797, 1996.
LAMOTTE, R. H., SRINIVASAN, M. A., LU, C., AND KLUSCH-PETERSEN, A.
Cutaneous neural codes for shape. Can. J. Physiol. Pharmacol. 72: 498–
505, 1994.
LINDBLOM, U. AND LUND, L. The discharge from vibration-sensitive receptors in the monkey foot. Exp. Neurol. 15: 401–417, 1966.
LOOFT, F. J. Response of monkey glabrous skin mechanoreceptors to random-noise sequences: I. Temporal response characteristics. Somatosens.
Mot. Res. 11: 327–344, 1994.
MACEFIELD, V. G. Signalling touch, movement and manipulation by mechanoreceptors in human skin. In: Neural Aspects of Tactile Sensation, edited
by J. W. Morley. Amsterdam: Elsevier. In press.
MACEFIELD, G., GANDEVIA, S. C., AND BURKE, D. Perceptual responses to
microstimulation of single afferents innervating joints, muscles and skin
of the human hand. J. Physiol. (Lond.) 429: 113–129, 1990.
MILLER, M. R., RALSTON, H. J., AND KASAHARA, M. The pattern of cutaneous innervation of the human hand. Am. J. Anat. 102: 183–197, 1958.
PHILLIPS, J. R. AND JOHNSON, K. O. Tactile spatial resolution. II. Neural
representation of bars, edges, and gratings in monkey primary afferents.
J. Neurophysiol. 46: 1192–1203, 1981a.
PHILLIPS, J. R. AND JOHNSON, K. O. Tactile spatial resolution. III. A continuum mechanics model of skin predicting mechanoreceptor responses to
bars, edges, and gratings. J. Neurophysiol. 46: 1204–1225, 1981b.
PHILLIPS, J. R., JOHANSSON, R. S., AND JOHNSON, K. O. Representation of
Braille characters in human nerve fibres. Exp. Brain Res. 81: 589–592,
1990.
PHILLIPS, J. R., JOHANSSON, R. S., AND JOHNSON, K. O. Responses of human
mechanoreceptive afferents to embossed dot arrays scanned across fingerpad skin. J. Neurosci. 12: 827–839, 1992.
SRINIVASAN, M. A. Surface deflection of primate fingertip under line load.
J. Biomech. 22: 343–349, 1989.
SRINIVASAN, M. A. AND LAMOTTE, R. H. Tactile discrimination of shape:
responses of slowly and rapidly adapting mechanoreceptive afferents to
a step indented into the monkey fingerpad. J. Neurosci. 7: 1682–1697,
1987.
TOREBJÖRK, H. E., VALLBO, Å. B., AND OCHOA, J. L. Intraneural microstimulation in man. Its relation to specificity of tactile sensations. Brain 110:
1509–1529, 1987.
VALLBO, Å. B. AND JOHANSSON, R. S. Properties of cutaneous mechanoreceptors in the human hand related to touch sensation. Hum. Neurobiol.
3: 3–14, 1984.
WHEAT, H. E., GOODWIN, A. W., AND BROWNING, A. S. Tactile resolution:
peripheral neural mechanisms underlying the human capacity to determine positions of objects contacting the fingerpad. J. Neurosci. 15: 5582–
5595, 1995.