BEHAVIOURAL

BRAIN
RESEARCH
ELSEVIER

Behavioural Brain Research 84 (1997) 81-97

Research report

Functional architecture of the mystacial vibrissae

Michael Brecht a,,, Bruno Preilowski a, Michael M. Merzenich b
a Tabingen University, Tabingen, Germany
b W.M. Keck Center for Integrative Neuroscience, University of California, San Francisco, USA
Received 8 February 1996; revised 10 June 1996; accepted 10 June 1996

Abstract
We investigated the transduction operation and function of the mystacial vibrissae, using a comparative morphological analysis
and behavioral experiments in rats. Vibrissal architecture was documented in a series of mammals to identify evolutionary
conserved features of vibrissal organization. As a result of this analysis, we distinguish between a frontal microvibrissal system and
macrovibrissal system of the mystacial pad. The latter was invariably comprised of whiskers aligned in regular rows. In each row,
whiskers were oriented perpendicular to the animal's rostrocaudal axis; all shared a specific dorsoventral orientation. In all species,
progressing from rostral to caudal in any vibrissal row, there was a precisely exponential increase in whisker length. Each whisker
appeared to act as a lever-like transducer, providing information as to whether or not - but not where - an individual vibrissa
had been deflected. The rat's frontal microvibrissae system was found to have a vibrissa tip density that was about 40 times higher
than that of the mystacial macrovibrissae. In behavioral studies spatial tasks and object recognition tasks were used to investigate
(a) search behaviors; (b) single whisker movements; (c) object recognition ability; and (d) effects of selective macro- or microvibrissae
removal on task performances. A clear distinction between the functional roles of macro- and microvibrissae was demonstrated in
these studies. Mystacial macrovibrissae were critically involved in spatial tasks, but were not essential for object recognition.
Microvibrissae were critically involved in object recognition tasks, but were not essential for spatial tasks. A synthesis of these
morphological and behavioral data led to the following functional concept: The mystacial macrovibrissae row is a distance decoder.
Its function is to derive head centered obstacle/opening contours at the various dorsoventral angles represented by vibrissal rows.
This distance detector model is functionally very different from traditional concepts of whisker function, in which the mystacial
whiskers were hypothesized to form a fine grain skin-like object-recognizing tactile surface.
Keywords: Barrel; Neuroethology; Somatosensory system;Transduction operation of the vibrissa apparatus; Vibrissa function; Vibrissae morphology; Whisker

1. Introduction
A comprehension of the basic operations of transduction is crucial for understanding any sensory system.
The transduction operations of the vibrissal apparatus
have not yet been described in an analytical way. In
these comparative morphological and behavioral studies
on the vibrissae we ask: H o w are the first order structural
features and functional operations of the vibrissal apparatus related?
The vibrissae are a m o n g the youngest of the major
m a m m a l i a n sensory systems. They are lacking in the
m o n o t r e m a t a [22], emerging first in the theria about
* Corresponding author. MPI for Brain Research, Postfach 710662,
D-60496 Frankfurt/Main, Germany.
E-mail: brecht@mpih-frankfurt.mpg.d400.de

120 million years ago. Facial vibrissae were originally

organized into five subfields, and m a n y mammals still
exhibit this primitive pattern [ 1,22]. In the vast majority
of mammals, the facial vibrissae mystaciales are the most
prominent division of the vibrissae apparatus. In this
analysis we focus on the function of this orderly array
of facial whiskers. For the rat's mystacial vibrissae, it is
useful to distinguish between the long, laterally oriented
'macrovibrissae' and the shorter, more numerous and
more frontal 'microvibrissae'.
Despite a long history of morphological and behavioral work beginning with Vincent (1912) [31] and
intense, more recent neurobiological interest beginning
with Woolsey and van der Loos (1970) [36], few
researchers have attempted to functionally characterize
this important sensory system [12]. Moreover, while
much research effort has been directed towards studying

0166-4328/97/$17.00 Copyright 1997 Elsevier Science B.V. All rights reserved

PII S0166-4328 (96)00136-2

82

M. Brechtet al./BehaviouralBrainResearch84 (1997) 81-97

the macrovibrissal physiology at the single whisker level,

relatively little attention has been paid to the functional
architecture of the vibrissae apparatus or to its operations, considered as a whole. Most commonly, the mystacial macrovibrissae have been described as an extension
of the general touch system.
The idea that the whiskers operate in a touch-like
manner was explicitly stated by Simons 1-27] to apply
to the vibrissae mystaciales - the 'macrovibrissae'
according to our terminology. This view stressed the
functional continuity of neighboring vibrissae, suggested
that the vibrissae move in a flexible fluid like way, and
hypothesized that they allowed for high resolution perceptions, comparable to those derived from primate
finger tips. This general hypothesis has often been
expressed with the assumptions that a major function of
the whiskers is to expand the search space of the
animal[34], and that the whisker array forms a kind of
'long range skin' [27]. Although rarely explicitly stated,
many researchers appear to presuppose that the different
whiskers of the array are functionally equivalent, and
that the whisker array forms a diffuse spatial sensor.
In contradistinction to these earlier studies and analyses it is here hypothesized that the mystacial macrovibrissae form a special distance detector sense organ.
A description of whisker function has to explain how
vibrissal afferents act cooperatively to generate useful
information about stimuli for brain analysis. Such a
description must be based on the relative functional
contributions of the individual whiskers in the mystacial
pad. A sense organ by its collective operation accounts
for more than any single input element, or the simple
sum of input elements [9]. One important prediction of
a sense organ concept is that the input elements constituting the vibrissal organ cannot have evolved independently without altering the organ's essential function. If
the macrovibrissae comprise a special sense organ, relatively constant whisker arrangements should be seen in
many different evolutionary divergent mammals that
possess it.
Generally, sense organs derive a highly specific and
useful stimulus representation. To find out what could
be the specific stimulus representation that is derived by
the macrovibrissal system, it was first asked: What is
the basic architecture of the vibrissae apparatus?
Towards that end, conserved species-invariant features
of the vibrissae array were defined, in contradistinction
to earlier studies, which concentrated on interspecies
differences [2,10,17,22,29].
The evolutionarily conserved spatial arrangement of
whiskers bears implications for the sampling and coding
of spatial information. We investigated this aspect quantitatively and asked: What is coded by a single whisker,
as compared with a whisker row or the entire vibrissal
array?
In the behavioral studies we asked the question: What

roles do different whiskers of the vibrissae apparatus

play? Specifically, how does the vibrissal system account
for object recognition and for spatial localization? How
do the rat's macro- and microvibrissae contribute to
these different behaviors?
Finally, on the basis of morphological and behavioral
data, we present a model of the functional architecture
of the vibrissae apparatus.

2. Materials and methods

2.1. Morphological analyses

2.1.1. Animals and species selection
Facial vibrissae arrays were studied quantitatively in
ten mammalian species, including:
Methatheria:
Marsupialia:
American
opossum
(Didelphis sp.); Shrew like opossum (Monodelphis domestica); Phalanger (Trichosurus vulpecula).
Eutheria: Insectivora: Tenrec (Tenrec ecaudatus);
Rodentia (rodents): Albino rat, n = 15, 2-20 months old,
males and females, (Rattus norvegicus (Sprague Dawley));
Mouse, n = 8, 10-20 months old, males and females (Mus
musculus); Pinnipedia (seals): Common seal (Phoca vitulina); Gray seal (Halichoerus grypus); Sea lion (Zalophus
californianus). Primates: Fat-tailed dwarf lemur
(Cheirogalens medius).
Additionally the vibrissa apparati of cats, ferrets,
hamsters and rabbits were studied qualitatively.
The rationale of this species selection was based on
two premises. First, the rodent and seal species are
established 'vibrissae experts' [17,29], and functional
features of vibrissal architecture should be particularly
pronounced in these species. Other species are distantly
related and were selected to identify general, evolutionarily conserved features of vibrissal architecture.
The measurements on rodents were performed on
freshly killed specimen which came from electrophysiological experiments. Except for the seals, the other
measurements were performed on specimen from the
animal collection of the Department of Zoology of
Tiibingen University (Ttibingen, Germany). These specimen had been preserved in alcohol, which is the preservation method of choice for hair and skin structures. The
vibrissae of the pinnipedia were measured from preserved
scalps of the animal collection of the Naturkunde
Museum in MOnster (Germany).
For rats and mice measurements were taken from
several specimen, for the other species only single representatives were used; the selection being based on which
specimen had the best preserved whiskers.
2.1.2. Definition of macro- vs. microvibrissae and the
studied parameters
For our quantitative comparison between the rat's
macro- and microvibrissae subsystems, we regarded the

M. Brecht et aL /Behavioural Brain Research 84 (1997) 81-97

83

and the nomenclature). There are about 30 of these

laterally oriented macrovibrissae.
The 'microvibrissae' are: (1) The most rostral whiskers

whiskers of rows A, B and the more caudal whiskers of

rows C, D, as 'macrovibrissae' (see Fig. 1 for an illustration of the general organization of the rat's vibrissae

a
Mystacial
brissae-field

"~
~

~ - -

External mystacial
Microvibrissae-field

Lower Jaw
Microvibrissae-field

A.

B,* %o
"

\
\

** :2 .

ooo o~oO

Straddler

Macrovibrissa

Microvibrissa

..(3,t~

g. " " : t . . J,~f.-.

Eo'..o.:" I--C-~

- -

h~'*

\ Microvibrissae-only

\'M

/"

\'

'llJ//

Macrovibrissae-only
Fig. 1. Mystacia! macro- and micro~brissae in the rat. (a) Side view of the mystacia] whisker fields. The mystacial microvibrissae-fie]d continues
on the inner side of the upper lip. (b) Magnified schematic frontal view of the mystacial microvibfissae. On one side only the skin positions of the
follicles of micro- and macrovibrissae are indicated. The lower jaw microvibfissae are not shown. (c) Schematic frontal view of the mystacia!
macrovibrissae. On one side only the skin positions and the cauda|most whisker of each row of macrovibdssae as well as the straddlers (dashed)
are drawn. Microvibrissae are not shown.

84

M. Brechtet al./BehaviouralBrain Research84 (1997) 81-97

of row C, D and E; these are 5-10 whiskers that are

oriented and positioned frontally, and thus positioned
around the corner from the more laterally directed
mystacial macrovibrissae. (2) The 35-60 whiskers of
rows F, G, H, I and J (the 15-25 more caudal whiskers
of these rows are in organization and length intermediate
to macro- and microvibrissae). The 20-35 more rostral
whiskers of these rows are typical microvibrissae, a
majority of which points downwards (ventrally).
For all studied animals, the layout of the vibrissal
apparatus, i.e., relative whisker orientation, spacing,
position of the vibrissae origins, and vibrissal lengths
distribution were defined. After a catalog of speciesinvariant features had been established, additional animals were checked for possible deviations of this scheme.
The length distributions of whiskers were measured
in all studied mammals. In investigating the length
distribution we were interested in the 'functional' or
'effective length' of the whiskers defined as the distance
from the skin surface to the vibrissa tip, i.e., maximal
distance at which the vibrissa touches an obstacle. In
the rodents we simplified this measurement by cutting
the whisker just above the skin. Regression analysis was
used to evaluate the relationship between functional
whisker length and position.
We also estimated whisker density for quantitative
comparisons between the rat's lateral macrovibrissal and
frontal microvibrissal subsystems. In the vibrissat system,
the distance between neighboring whiskers is a morphological indicator of the spatial sampling density. Whisker
density equals the number of whiskers per unit area
within the sensory plane that they form. The most
appropriate and simplest definition of the sensory plane
formed by the vibrissae is the plane formed by the
vibrissae tips. This plane was approximated as a set of
four trapezoid shapes formed by the whisker tips of the
five whisker rows (the straddlers were assumed to be
standing in line with the rest of the row).
2.2. Behavioral analyses
2.2.1. Animals
Studies were performed in four adult, blind, female
RCS-rats (1-2 years old, 150-250g weight), with an
inherited retinal defect, and in five adult male albino
rats (5-6 months old, 200-300g weight, SpragueDawley, Interfauna Tuttlingen). In their home cages, the
animals had access to water and food ad libitum
throughout the study and the experiments. Room lights
were on a regular 12-h light/12-h dark schedule.
2.2.2. Recording of whisker movements
During the object discrimination paradigm described
below, the movements of single whiskers were recorded
by means of a special recording technique: Single whiskers of the rat were marked with a miniature patch reflex

foil. Measurements taken of whiskers with and without

foil indicated that this foil patch did not affect whisker
movements. A light source was placed laterally and its
light reflected onto the scene by a 45 inclined halfsilvered mirror. The video recordings were made from
above through the mirror. The light gain could be
maximized by adjusting the focal length of the camera
to match the virtual position of the light source in the
mirror. An RCA-camera with a sampling rate of 50 Hz
was used to record movement sequences. The shutter of
the RCA-camera was open for the complete frame
duration (20 ms), and the movement of the labelled
vibrissal spot during this time appeared as a variable
length white line in the frame. Some sessions were taped
in darkness using infrared light.
2.2.3. Object discrimination task: Psychophysics of
object recognition
Search time measurements for the identification of a
target object in the presence of a number of distractor
objects have been a major paradigm for studying visual
object recognition in humans [30,18]. This task was
adapted to study vibrissal object recognition in albino
and blind rats. Rats were presented with cookies of
different sizes and shapes that tasted sweet or bitter
depending on size or shape.
Small cookies were made from a batter that contained
sugar, butter and flour (1:2:6), and a few drops of water.
Non-target cookies were embittered with caffeine, which
is odorless. Cookies were flat ( 1-2.5 mm thick) geometric
forms. Rats had to discriminate between sweet small
right triangles (6 mm side length) as targets and bitter
small squares ( 6 m m side length) and/or bitter large
triangles (8 mm side length) as distractors.
Catch trials were run with well trained rats; in such
trials embittered target-shape cookies or sweet, nonembittered distractor-shape cookies were presented to
each animal (5-8 such tests per animal). Several sessions
of the albino animals (including catch trials) were performed under infrared illumination or in total darkness.
The rats were trained to perform the discrimination,
with the discriminanda presented on an elevated stage
(see Fig. 4a). Two types of experiments were performed.
In the first, animals were presented with one or no
targets (a sweet small triangle) together with a variable
number of distractor cookies. This arrangement was
designed to determine the interdependence of object
number and search time.
In a second experiment the target cookie (a small
sweet triangle) was presented to the rats together with
15 distractor cookies (small squares) in a regular 4 x 4
array. This test situation is shown in Fig. 4a. The target
appeared at random positions in the array on different
trials. In between, trials were run in which single targets
or distractors without additional distractors were presented at random positions of the array. An objective of

M. Brechtet al./BehaviouralBrainResearch84 (1997) 81-97

this experiment was to determine the relationships
between target positions and search times. These tasks
were also used to assess the rat's object recognition
abilities, using search time measurements before and
after whisker removal. All trials were videotaped. Search
time was determined in frame by frame analyses.
2.2.4. Spatial task
The locating of a single edible cookie in a 75 80 cm
field was chosen as a spatial task for the four blind
RCS-rats. Two metal cylinders (10 cm diameter and a
platform (20 20 x 7 cm) that appeared at random locations in the open field served to enhance the spatial
complexity of the cookie locating task. The cookie was
randomly positioned in this field.
2.2.5. Selective whisker removal
In the four blind animals, facial vibrissae were
removed in two steps involving the two facial whisker
subdivisions of mystacial macrovibrissae and microvibrissae. Before and after each removal, rats were tested
on object recognition and spatial (cookie locating) task.
Trials were analyzed for search time, as well as for hits
and misses. A trial was scored a 'hit' if the animal picked
up the target cookie during the first search attempt. A
trial was scored as a 'miss' if the rat turned its head
away from the cookies, stopped head movements and
ended the search.
For whisker removal, animals were held in a plastic
bag with a snout opening. Pseudo-shaving control procedures in which the animals were handled in the same
way as for whisker removal were interspersed with actual
removals. For these control procedures, animals were
immobilized in the bag for several minutes, and the skin
was palpated with a pair of scissors or a shaver, as in
the removal procedures.
In two animals, the mystacial macrovibrissae were
removed first with a pair of scissors, with the external
microvibrissae (mystacial and lower jaw) trimmed
second, using an electric shaver. In the other two animals,
the vibrissal removal sequence was reversed.
All test trials within the whisker removal experiment
were run within 5-6 days, with 35 object recognition
tasks and 50 cookie location tasks/animal/condition
(Pretest, pseudo shaving, microvibrissae or macrovibrissae removal, micro- + macrovibrissae removal).

3. Results

3.1. Morphology of the vibrissal system

3.1.1. The dist&ction between lateral macrovibrissae and
frontal microvibrissae
In the course of a simple analysis of vibrissae arrays,
it became clear that the longer, more lateral mystacial

85

vibrissae and the shorter, more frontal mystacial vibrissae have different functional characteristics. Since it was
also found that whisker length was a key parameter in
the vibrissal system, it appeared appropriate to classify
whiskers with respect to their sizes. The macrovibrissae
are the long, more posterior, 'classic' mystacial whiskers;
the microvibrissae are the shorter, more anterior whiskers. ~ For macrovibrissae, whisker length was a major
parameter determined by the position of the whisker tip
relative to the surrounding whisker tips and the geometry
of the sensory plane formed by the whisker tips. For the
microvibrissae, whisker length was only a secondary
parameter with respect to the relative position of the
whisker tip.
The transition between macro- and microvibrissae is
not abrupt. Nevertheless, the following results justify
distinguishing them.
There are 40-70 microvibrissae (see also Fig. 1). The
quantitative morphology of microvibrissae is complicated by the short length of these whiskers and the
difficulty of differentiating them from fur hair. Moreover,
unlike macrovibrissae, missing microvibrissae are not
easily detected.
The rat has two further microvibrissal fields not
included in our mystacial macro-/microvibrissae comparison. One internal microvibrissae field is on the inner
side of the upper lip; this field in the rat's mouth is
continuous with the above-described external field and
consists of 20-40 whiskers in a high-density arrangement. Another microvibrissae field is found on the rostral
aspect of the lower jaw (ca. 30 whiskers).
As documented in Table 1 the organization of the rat's
mystacial macrovibrissae and the external microvibrissae
is quite different.
Table 2 tabulates some of the quantitative differences
between mystacial macro- and microvibrissae. The most
obvious differences are found in the length of the whiskers and the density of the sampling array. The latter is
>40 to > 100 times higher for of the microvibrissal
system than for the macrovibrissal system.
3.1.2. Species-invariant architecture of the mystacial
macrovibrissae
Qualitative invariances of the mystacial macrovibrissae.
In many mammals, including all species studied here,
the mystacial whiskers or macrovibrissae are the most
prominent or the only vibrissae division [10,22]. In a
majority of mammals, they consist of a set of whiskers
that are highly ordered with respect to their location,
lengths and orientations.
A central observation is that rows are the basic units
of the mystacial pad: (a) Rows are found in all species.
(b) Rows are always straight and approximately parallel.

1We thank H.U. Schnitzlerfor suggestingthis terminology.

86

M. Brecht et al./Behavioural Brain Research 84 (1997) 81-97

Table 1
Qualitative comparison between the macrovibrissae the microvibrissae

Morphological
characteristics

Single whisker
Mystacial array

Morphological
location

Major field
Other occurrences

Movement

Whisking movement
Head movements

Table 2
Quantitative comparison between the
microvibrissae

Macro-vibrissae

Micro-vibrissae

Lateral orientation (on the

average)
Sensory plane almost
perpendicular to skin and
rostrocaudal axis
A-, B- row and the non frontal
parts of C-, D-, and E-row
Many additional subfields
with few or single whiskers
Major role in target contact
Minor role in target contact

Diverse orientations
(ventral and frontal are most common)
Sensory plane more parallel to skin

macrovibrissae and

Macrovibrissae

Microvibrissae

Length
Number

> 4 mm
30-35 (mystacial)

Estimated extent of
the sensory plane
Average whisker
density
Minimum whisker
density
Maximum whisker
density

16.4 cm2(10-25 cm2)

< 7 mm
40-70(external,
mystacial)
0.69cm2(0.3-1.5 cmz)

2/cm 2

87/cm 2

0.35/cm2

10/cmz

25/cm2(10-50/cm2)

281/cm2(200-600/cm2)

Note. Our observations relate to the mystacial macro- and microvibris-

sae division and not to other vibrissae subfields. The term sensory
plane refers to the plane formed by the vibrissae tips. The quantitative
measurements on number and sensory plane refer to one side of the
head only.

(c) Whiskers in a row are close together, often only a

minimal positional distance is observed. (d) There is a
relatively great distance between neighboring whisker
rows; often a m a x i m u m distance spacing of rows is seen.
(e) The whiskers within a row share one dorsoventral
orientation, which is maintained during whisking movements. (f) Whiskers of a row sample highly overlapping
spatial information.
As m a n y as nine organizational constraints have been
found to be invariant in the 14 species studied. These
organizational principles of the mystacial macrovibrissae
are summarized in Fig. 2.
There is a strict whisker-specific multiparametric order
in the array. Thus, with knowledge of the relative
location of a whisker, its relative spatial orientation and
its relative length are predictable. According to their
orientation and bending, the whiskers point away from
each other, i.e., they diverge. It is important to note that
the vibrissae do not form a cloud-like distribution in
space. Moreover, vibrissae origins do not follow an

The most frontal parts of C-, D-, E- row and F-, G-, H-, I-, J-row
Only two other subfields on the lower jaw and inside the mouth.
No single or few whisker subfields
Unknown
Major role in target contact

equidistant (maximum spacing) distribution on the pad.

Rows very strictly confine locations and specific dorsoventral orientations of their whiskers. The positions of
vibrissae origins are generally crowded towards the front
of the animal. However, in all examined species, the
macrovibrissae were oriented laterally, and thus do not
form a forerunning sensory array.
Quantitative invariant length tuning - organ-pipe row
architecture. In all studied species, effective whisker

length was found to increase exponentially along rows,

proceeding in the rostral-to-caudal direction. Whisker
length was plotted against whisker position on a logarithmic scale in Fig. 3a for the A-row, and in Fig. 3b for
the D-row. These straight line functions reveal an exponential effective whisker length versus position scale.
The effective length of the next more caudal row
neighbor of any whisker was 1.2 to 1.6 times longer, a
constant value depending on the species and the
specific row.
The intraspecies variance of whisker length tuning
along the row was examined more closely in rats and
mice. As shown in Fig. 3c and 3d the same pattern as in
inter-specific comparisons was obtained. While the absolute whisker lengths differed between individual rats and
mice, the relative whisker lengths (the exponential
lengths versus positions distribution) applied to all examined individuals.
Another observation from the rat and mouse data
concerns the so-called 'straddlers'. The straddlers are the
four caudal-most whiskers of the mystacial pad, and are
found only in rodents. They are laterally displaced
relative to the whisker rows. The straddlers deviated
(towards shorter lengths) from an exponential lengths
versus positions scaling.
Finally, regression analyses were used to evaluate how
well an exponential function described the relationship
between whisker position and length. For the D-row
whisker-position relationship, an exponential function
explained most of the variance; for the D-row of the
various examined species the correlation coefficients
between exponential fits and data were between

M. Brecht et aL /Behavioural Brain Research 84 (1997) 81-97

87

Whisker position
constraints:
1. A majority of the animal's
macrovibrissae are located on the
mystacial pad: otherwise random
positions.

2. Whiskers are organized in rows.

3. Rostrocaudal parallel rows.

Whisker orientation
constraints:
4. Maximal spacing between rows. 7. Whisker orientation
perpendicular to rostrocaudal
5. Minimal spacing of neighboring
axis.
whiskers within the row.

Whisker length
constraints:
6. Maximally rostral position of the 8. Divergence of whisker vectors
within the dorsoventral plane
whisker row.
of the row.

9. Exponential length tuning

along the whisker row.

Fig. 2. Principles of mystacial pad architecture. The first column describes constraints on position of the whisker origin. Starting from the top, the
rules which lead to a species-invariant distribution of whisker positions (bottom) are listed and illustrated. The second column describes constraints
on whisker orientation given the species invariant whisker positions. The third column illustrates the rule which leads to the species-invariant
distribution of whisker lengths. All species studied showed the vibrissae organization illustrated on the bottom right.

r2=0.942 and 0.994 (species average r2=0.978).

Similarly, high correlation coefficients between exponential fits and length position data were observed in
intraspecific comparisons for the D-Row of rats (range:
r2=91.6-100, average r2=96.8) and mice (range:
r2=92.2-98.7, average r2=96.9). The excellent fit
between effective whisker length and sensory surface
position by an exponential function is shown in Fig. 3e.
For the average laboratory rat, the deviation from the

predicted exponential lengths was less than 2 mm per

whisker. Many specimens showed deviations of less than
1 mm per whisker from an exponential length scaling.
All other species studied showed an equivalently precise
length versus position relationship.
While the effective length distributions along rows
were species invariant, the length distributions along the
arc (i.e., over columns) of the mystacial pad did not
follow any strict species invariant rule (Fig. 3f).

88

M. Brecht et aL /Behavioural Brain Research 84 (1997) 81-97

a
A-Row

[]
Tenrec
[] Rat
o Grey seal
O Phalanger
A Mouse
IA Shrewl. opossum

100- -

,~
1000

Sea lion
u Grey seal
Common seal
0 A. opossum
Rat
Z~ Phalanger
Shrewl. opossum
o Mouse

D-Row

~" I O 0 -

E
E

10-

t-t. lemur

e-

e'-

~, l O - ~

--I

4
Position

5
6
Position

A-Row in Rats and Mice

1O0

~.

Rat 5
Rat 6
Rat 7
Rat 8
--4--- Rat 9
& Rat 10
0 Rat 11
- - l l - - Mouse 1
Mouse 2
A Mouse 5
+
Mouse 6
A Mouse 7
X Mouse 8

E
E

t.--i

4==

10

e-

--I

Position
Fig. 3. Whisker length distribution along the mystacial pad in a variety of species. The effective vibrissae length (distance from skin to whisker tip,
disregarding curvature) is plotted in a logarithmic scaling vs. position. Position 1 represents the caudalmost whisker. (a) A-row length distribution
for species with a complete A-row. In rodents the ~t-straddler was not included. (b) Length distribution of the D-row, a central mystacial macrovibrissae row. In the rodents the k-straddler was not included. (c) Intraspecific pattern of A-row (including the ~t-straddler) length distribution in rats
and mice. (d) Intraspecific pattern of D-row (including the k-straddler) length distribution in rats and mice. (e) Whisker length data for the D-row
fitted with exponential functions. (f) The whisker length distribution along the first arc (column) of the mystacial pad. The medium rows tend to
be somewhat longer, but, in contrast to the A- and D-rows, no species-invariant scaling is seen.

M. Brecht et al./Behavioural Brain Research 84 (1997) 81-97

89

D-Row in Rats and Mice

1t
0
O
A
A

Rat 5
Rat 6
Rat 7
Rat 8
+
Rat 9
+Rat
10
---I--- Rat 1 1
X Mouse 1
0 Mouse 2
[] Mouse 3
--@-- Mouse 4
A Mouse 5
@ Mouse 6
- - I - - Mouse7

E
E
f-

10
t/

Position

Exponential Fits

Arc

80.
I=,
rl

I0

60\

E
E

\'-k

\\

Grey s e a l
Common seal
Phalanger
Mouse ur

120 t
seal

100

Grey s e a l
Phalanger
Rat
Shrewl. opossum

EE 80

t-

40t-

60

Fat-t. lemur

.J

20-

20
i

01

4s6e

;;o

0
A1

,
B1

i
C1

,
D1

,
E1

i )F1

Position

3.2. Behavioral observations on vibrissal perception in the

rat
3.2.1. Object recognition
Albino- and blind RCS-rats achieved high levels of
performance in cookie shape and size discrimination

tasks. Task acquisition was fast (within a day or two),

and errors as measured by false positives (when nontarget cookies are picked up) were few (1% or less).
The results from the catch trials with both albinos
and blind animals indicated that the animals discriminated the cookies primarily or exclusively by shape:

90

M. Brechtet al./BehaviouralBrainResearch84 (1997) 81-97

More than 90% of the embittered target-shapes (19 of

21 such cookies) were taken up by the animals, and less
than % 5 of the non-embittered distractor-shapes (3 of
70 such cookies) were taken up by the animals. Albinos
performed the discrimination in darkness in a shape
dependent manner (normal trials as well as catch trials).
Thus both blind animals and albinos appeared to be
able to use tactile cues for the discrimination.
Search behavior. Some species such as rats and hamsters whisk, that is, they move their whiskers in rapid
back-forth successions, generally within the plane of a
macrovibrissae row [4, 33, 34]. When blind rats or albino
rats searched an area for cookies they whisked intensely.
If small object numbers had to be discriminated the
behavioral strategies of albino rats and blind animals
were very similar. It was often seen that a rat passing in
the vicinity of the cookies detected them with its long
macrovibrissae. After detection, a specific sequence of
behaviors followed. Invariably, the rat turned its head
towards the cookies. During a discrimination phase that
varied in length as a function of the number of cookies
to be discriminated (0.2-5 s), the rat oriented frontally
toward the cookies and ultimately stationed its head
directly over the target cookie. At that time, the microvibrissae between nose and mouth were in direct contact
with the target cookie. The retrieval of the target cookie
followed.
Role of identified whiskers in the task. Videotaped
movements of single mystacial macrovibrissae during
the object discrimination task showed that these whiskers maintained a perpendicular orientation with regard
to the animal's rostrocaudal axis throughout the task.
The whiskers of the C-row had a lateral orientation, the
A-row a more dorsal orientation and the E-row a more
ventral orientation. The caudal macrovibrissae deviated
slightly backwards from the average lateral orientation,
and the anterior mystacial whiskers had a more frontal
orientation. In the context of the search behavior, this
lateral orientation implies that the macrovibrissae contact the target cookie during search, before the animals
turns toward the cookies, i.e., before the beginning of the
discrimination. The long mystacial macrovibrissae were
not in contact with the cookies during the discrimination
phase. This pattern of whisking movements and target
approach with a centering of the microvibrissae relative
to the discriminanda was very similar or indistinguishable between blind animals and albinos.
Quantitative aspects of vibrissal search. Blind rats
investigated a cookie cluster by moving their heads from
one cookie to another. This one-by-one inspection strategy was also indicated by the quantitative analysis of
the dependence between search time and object number.
A linear increase in search time with display size was
recorded. This increase in search time was twice as steep
in the target-absent condition as in the target-present
condition. These observations demonstrate that vibrissal

search occurs as a serial, self-terminating process. This

conclusion applies only to blind animals. In one of the
recognition tasks the four blind rats were presented with
an ordered 4 x 4 array of cookies (see Fig. 4a); the side
length of the array was 3.2 cm, and therefore considerably smaller than the maximum diameter of the whisker
fan (which is about 11 cm). In all animals, there was a
great increase in search time from the frontal row of the
array to the posterior rows of the cookie display
(Fig. 4b). Moreover, on average, targets in middle positions of the rows were found faster.
When single targets or distractors without the distractor objects were presented at random positions of
the array the search times were very short (shorter than
the average search time in the first row) and largely
uniform cross the array positions. This indicates that
the search time increase across array positions in our
paradigm is due to the perceptual load induced by our
distractor objects. All of these observations are consistent
with the conclusion that the animals searched the display
serially, starting from the middle position of the first
row and moving on to the side and the posterior rows.
Effects of selective whisker removal. Figs. 5 and 6
illustrate the effects of selective whisker removal in the
genetically blind RCS-rats on object recognition and the
spatial cookie search task in the open field.
Macrovibrissae removal did not result in a significant
search deficit in the object recognition task (Fig. 5), but
led to a dramatic search time increase in the spatial
cookie location task (Fig. 6). On the other hand, microvibrissae removal had no visible effect on the cookie
locating task (Fig. 6), but resulted in a profound deficit
in object recognition ability (Fig. 5).
Thus, in these trained RCS-rats, a double dissociation
between microvibrissae accounting for object recognition
and macrovibrissae accounting for spatial object detection was recorded.

4. Discussion
From these results, three interconnected hypotheses
are derived.
(1) The long mystacial macrovibrissae and the short
frontally directed microvibrissae comprise two functional subdivisions of the rat's vibrissal system.
(2) The microvibrissae can be viewed as an object
recognizing sense organ. They provide the sampling
density that is crucial for tactile object recognition.
They play a negligible role in sampling spatial
information.
(3) In contrast, the mystacial macrovibrissae can be
viewed as a distance detecting/object locating sense
organ. They have an evolutionarily highly conserved
architecture. We suggest that they function as a

M. Brecht et aL /Behavioural Brain Research 84 (1997) 81-97

91

The experimental set up

camera

C~

Animal

Animal l_

25002000

400(

.~ 1500

E 300(
.=_
~) 200

.~ 1000

._E

50C
(

100

.T=

o')

Animal 3

Animal

400040003000

E 3000-

2000

2000
E

1000

100

(
"-

"F

~)

"E_

ou
i..

front

13}

Fig. 4. (a) Blind RCS-rats were presented with an ordered 4 x 4 cookie array consisting of 15 bitter square distractor cookies and one sweet triangular
target cookie at different positions. Note that the cookie array was much smaller (< 10 cm 2) than the animal's whisker fan, which spans > 30
cm 2. (b) For four blind RCS-rats, the averaged time until pickup of the target is displayed for each of the 16 positions.

92

M. Brecht et aL/Behavioural Brain Research 84 (1997) 81-97

(a)

Macrovibrissae removal first

Animal 1

._~

Animal

1O0

100

90

90

80

80

70

70

60

,~

so

-r-

4o

60
50

P_
,-r-

40

3o

30

2O

20

lO

10

==
._

+J2m
o

"o

(b)

Microvibrissae

removal first
Animal 4

Animal 3
k ,/r
I O0

100
ffl

._=

90

90

(n

80

70

70
60

60
-r-

80

50

.-r-

50

40

40

30

30

::'0

20

10

lO

==
g

.=
'~
.ci

==

=.

":"
.~m

'~

.=
=-

==
=.

"-

~.

==
o.

~e

"~o

~,E
~7.=

Fig. 5. Effects of selective macro- and microvibrissae removal in blind RCS-rats on hits and misses in the object recognition task (illustrated in
Fig. 4a). (a) Effects in the two animals where macrovibrissae were removed first. (b) Effects in the two animals where microvibrissae were removed
first. A Cochran-test showed that there are highly significant (P<0.001) differences between the different conditions in each animal; df= 3 in (a) and
df=2 in (b). The distribution of hits and misses in the columns marked with **differs highly significantly from the respective pretest condition (left
column) (P < 0.001 McNemar Chi2-test, df= 1). Only microvibrissae removal leads to significant deficits in the object recognition task.

distance detector array providing

spatial information.

head-centered

4.1. The distinction of a fovea-like microvibrissae system

and a spatial macrovibrissae system
Our object discrimination experiment and control
procedures - namely the use of blind animals, task
performance of albino animals in darkness, catch trials,

and the selective whisker removal - show that both

albinos and blind animals are capable of tactile object
discrimination. Our video recordings indicate that both
groups appear to use a similar behavioral strategy. The
behavioral data point to a different role for macroversus
microvibrissae in an object recognition task. This difference is manifested (a) by the rat's search behavior; (b)
by the observation of whisker movements; and (c) by
behavioral changes - or lack of changes - following

93

M. Brecht et al./Behavioural Brain Research 84 (1997) 81 97

(a)

Macrovibrissae
Animal

.E_

180

.E

160

03

4-1

=0

Animal

200

t-

removal first

160

I--

140

140

120

120

I O0

100

80

80

60

60

"6

40

40

20

20

o
O)
r,_
>
.~

m
In
u)
T"

'~

==

(b)

t.

+ ' Q'--- ~

0
"0

>

~ E

0 >
~- 0

removal first

Animal 3

Animal

870 I

._.

40

30

60

._E

6o

uuE

~.~

Microvibrissae

t'n

Il-

50

40

In

30

20

P
~'o

10~
o.

0~-10

10 t
-10
-20

"~

m
._

>~

+~_

o
~ 0

0 ~ 0

~E

,o>

Fig. 6. Effects of selective macro- and microvibrissae removal in blind RCS-rats on search time in the spatial localization task. (a) Effects in the two
animals where macrovibrissae were removed first. (b) Effects in the two animals where microvibrissae were removed first. The stars indicate a
significant (*P<0.05, two-tailed t-test, df=49) or a highly significant (**P<0.001) difference from the pretest condition. Only macrovibrissae
removal leads to significant deficits in the spatial localization task.

94

M. Brecht et al./Behavioural Brain Research 84 (1997) 81-97

selective macrovibrissae removal. At least in blind rats

the mystacial macrovibrissae play only a limited role and are not sufficient by themselves - for object recognition. The findings from the vibrissal search experiments,
the object array experiment and the rat's search behavior
indicate that the animals searched the object displays
serially. The large search time differences across the
positions of the object display indicate that the rat is
not able to 'view' the entire array at once on a level
adequate for object recognition, but had only a small
effective tactile window for sampling objects, smaller
than the 3.2 3.2 cm of the display, and far smaller than
the sensory plane of the macrovibrissae (> 30 cm2). From
our removal results we argue that object recognition
occurs primarily in a relatively small tactile window
corresponding to the microvibrissae.
This general distinction between the functional roles
of vibrissal subsystems is supported by observations on
the morphological differences between macro- and
microvibrissae. Whisker density in the microvibrissae
system is about 40-100 times higher than for the macrovibrissae system. Based on receptor density, one might
think of the microvibrissae as a fovea-like system,
brought into action after stimuli to be explored are first
located by the macrovibrissae. Behaviorally both albinos
and blind animals appear use a kind of 'foveation'
strategy which brings the microvibrissae onto the
discriminanda.
Two studies on vibrissal object recognition, one on
the walrus by Kastelein and van Gaalen 1-16] and one
on the sea lion by Dehnhardt I-5] report qualitatively
similar findings. According to these studies, neither of
these pinnipeds use their long more caudal whiskers for
object recognition, but engage their short frontal vibrissae for such tasks.
Our behavioral observations and the macrovibrissae
removal experiments demonstrate a major role of the
macrovibrissae in spatial, object-locating tasks. This role
has been recognized since the first behavioral study on
whiskers by Vincent [31], who described deficits of
dewhiskered rats on maze tasks. A long list of related
observation has been published (e.g. 1-23,24]). A particularly interesting effect results from one-sided vibrissae
removal (hemi-vibrissotomy), which causes the animal
to move along a wall with the vibrissae-intact body side
facing the wall 1-20].

4.2. Functional architecture of the macrovibrissae:

Distance detector theory
4.2.1. Distance detector theory: Computational tasks of
the mystacial macrovibrissae
How does the macrovibrissae architecture function in
sampling spatial information?
Straight, generally densely spaced whisker rows
appear to be fundamental units of the macrovibrissal

pad. The macrovibrissae are not arranged for a homogenous, equidistant sampling of space. As the rows confine
whisker positions and orientations, whiskers of a row
occupy - and are whisked through - overlapping segments of space. What is different about the whiskers of
a row is their length. Indeed, whisker length varies highly
systematically along the rows (and not along arcs, i.e.,
columns) in all studied species.
From these facts it is concluded the whisker row is a
distance decoder comprised of binary distance detectors;
the longest (the most caudal) untouched whisker codes
the minimal distance to an obstacle. The corresponding
computational task of the mystacial pad is to derive
obstacle/opening contours, the dorsoventral angles of
which are coded by the rows in a sensory plane at the
top of the animal perpendicular to its rostrocaudal axis.
This distance contour information is generated in a
robust binary form, and coded in head-centered coordinates. Fig. 7 schematically displays the properties of such
a hypothesized distance detector array.
The idea that the row performs an orderly extraction
of distance information provides a functional explanation for the fact that the mystacial pad is made from
row-modules with their organ-pipe architecture. The
observed precision of effective whisker length tuning is
quite extraordinary if one considers the permanent
random mechanical degradation (breakage, wear, bending) of the vibrissae due to their exposed body position.
This precise exponential length tuning is requisite, if the
row is providing a fine scale for distance, and is the
basis for breakdown of spatial information into a coordi-

Fig. 7. Hypothetical transduction operation of the macrovibrissae

according to the distance detector theory. The sensory plane is oriented
perpendicularly to the animal's rostrocaudal axis and different rows of
the mystacial macrovibrissae code for different angles. A single row
functions as a distance decoder, composed of a set of whiskers of
different length, each acting as binarily (touched/untouched) coding
distance detector. Short distances are over-represented. An obstacle
(gray bar) deflects all whiskers (black) longer than the minimal obstacle
distance. The system defines contours between longer touched whiskers
and shorter untouched whiskers. Moreover, the sign of the contour is
fixed: Only contours of obstacles (outside) and openings (inner side)
can be detected.

M. Brecht et al./Behavioural Brain Research 84 (1997) 81-97

nate system formed by labeled lines. Moreover, it is a

scaling scheme that overly represents the short and
presumably more important distances and might optimize the sensory contribution of each distance detector
(whisker) by keeping the relative length difference
between neighboring whiskers constant. While whisker
origins are closely spaced, there is some divergence of
the whiskers of a row. This might be due to the need of
reducing mechanical crosstalk between neighboring
whiskers.
The lateral orientation of the mystacial whiskers maximizes the search space for distance information. The
lateral orientation and the minimal rostrocaudal extent
of the pad indicates that the rostrocaudal dimension is
secondary and that the functional sensory plane formed
by whisker tips is oriented laterally, perpendicular to
the rostrocaudal axis of the animal. According to this
view, the rostrocaudal whisking movement might be
considered a distance scanning behavior, as it allows a
row of whiskers - all of different length - to move
through an overlapping segment of space.
While we suggest that the set of different length
whiskers of a row provides an efficient mechanism for
distance estimation, distance information is not easily
available from single whiskers. The vibrissa hair itself is
a dead structure, and the deflections sensed by the
receptors around the whisker shaft [3,21] reveal no
direct non-ambiguous information about the length (=
distance) at which the whisker is deflected.
The functional properties of a distance detector array
might be outlined by contrasting them with an alternative view of mystacial whisker function, which postulates
a skin-like operational mode of the mystacial array
[27]. In a 'skin-model' the vibrissae tips are hypothesized to form a sensory array comparable to a fingertip.
Table 3 lists tentative differences of these models. A
distance detector array seems to be suited to provide
coarsely coded spatial information for orientation and
object detection, as opposed to high-resolution information for object recognition, which a skin-like sensory
array is more suited for. In a distance detector array the

95

stimulation of a shorter whisker automatically implies

that the longer whiskers are stimulated as well.
Therefore, the type of transduced information represents
obstacle/opening contours between sets of engaged
(longer) and non-engaged (shorter) whiskers, as compared to pixel-like images of surfaces sensed by the skin.
Additionally, the sensory surface of a distance detector
array is anisotropic, because the sign of the contour is
not interchangeable. The obstacle is always outside, and
the opening is always inside. This differs from the
isotropic surface of a skin-like array. A distance detector
array forms a discontinuous sensory surface with perpendicularly oriented whiskers and a rigid vibrissa apparatus. The sensory surface of a skin-like array is more
continuous, with flexible whisker orientations in a fluidlike vibrissae apparatus [4]. Thus, the coding of spatial
information in both models is grossly different.
We argue that the architecture of the macrovibrissae
supports best detector functions, but this does not imply
that binary coded obstacle/opening contours are the
only information available from the macrovibrissae. The
single whisker can provide more elaborate inputs, such
as successive deflection (timing = surface roughness)
and deflection-directional information. At the same time
our behavioral results do not support the idea that this
single whisker information is generally used to deduce
the fine geometry of a contacted obstacle.
Like other sense organs the macrovibrissae certainly
serve different and multiple functions depending on
species and behavioral context. Thus, while the distance
detector model tries to capture what we consider to be
basic principles of macrovibrissae operation, it cannot
constitute an universal or exclusive theory of the
macrovibrissae.

4.2.2. Morphological and behavioral evidence

As demonstrated by the way that the distance detector
model was derived, a distance detector array hypothesis
requires and explains the morphological invariances of
the mystacial pad. At the same time, several of these

Table 3
Functional comparison between the distance detector model and a skin-like mode of operation of the macrovibrissae

Behavioral function
Type of transduced information
Properties of the sensory surface

Spatial coding

Distance detector array

Skin-like array [27]

spatial orientation and monitoring

obstacle/opening contours
functionally non-continuous
anisotropic
perpendicular orientation
necessarily rigid
maximal distance information
minimal rostrocaudal information
representation in spatially fixed head-centered coordinates
binary, coarse, touched/untouched coding across the
input array

object recognition and spatial orientation

pixel-like image of surfaces
functionally continuous
isotropic
flexible orientation
fluid, mosaic-like
minimal distance information
rostrocaudal information preserved
flexible spatial relationships
fine coding of relationships and patterns across
the input array

96

M. Brechtet al./BehaviouralBrain Research 84 (1997) 81-97

morphological features appear to be inconsistent with a

skin-like operation of the macrovibrissae. Thus:
(1) The spatial resolution of the long mystacial whiskers
is limited, because of the low number of only 30
whiskers which erect a huge whisker fan.
(2) The neighborhood relationships of vibrissae tips (the
major information communicated by a skin-like
array) play only a secondary role in the mystacial
pad. The vibrissae lengths are different. The peripheral whisker rows are generally shorter than the
central whisker rows, and the vibrissa tips are therefore not lying in a plane. The rotatory whisking
movements distort the proportional relationships of
the whisker tips.
(3) The mystacial whiskers are oriented laterally and
not frontally.
Experimental results not fully consistent with our
hypothesis are recorded in reports on extremely fine
texture discriminations of two grids by the mystacial
macrovibrissae [4,11]. However, while these studies
provide evidence for vibrissal texture discrimination,
both studies were not designed to differentially test the
involvement of macro- and microvibrissae in this task.
Therefore, at present it can not be judged how macroand microvibrissae compare in their texture discrimination abilities and which whiskers the rats preferentially
use under such circumstance.
The distance detector theory might be tested by its
behavioral predictions. We predict that animals are
massively superior in estimating distances rather than
extents with their vibrissal system. In a Y-maze forced
choice paradigm in which the rats have to chose either
the wider of two gates or have to select the wider of two
walls, the animals are expected to perform substantially
better in the gate-choice-task. Similarly, we predict that
the distance-estimation-abilities of the animals should
quantitatively covary with the vibrissal architecture. This
could be tested in the gate-choice paradigm with an
animal whose whiskers have been clipped except for the
C-row. If the maximal distance which the longest whiskers can touch is taken as a reference and progressively
smaller distances are to be discriminated by the animal,
an abrupt increase in the animal's distance-estimationability (measured by error rate or reaction time) should
be seen when the distance becomes small enough for the
next shorter whisker to touch. Such a 'quantal' course
of the discrimination curve would also demonstrate the
detector nature of the whisker signal.
4.2.3. Distance detector theory and existing
neurobiological data o f the vibrissal system
How does a distance detector theory fit into the rich
data on the neurobiology of the barrel cortex?
(1) While more recent electrophysiological and imaging
studies demonstrate multiwhisker RFs, there is a

(2)

(3)

(4)

(5)

consensus that strong response preferences for single

whiskers are a major characteristic of barrel cortex
neurons [14,25,26,28,32]. Together with the fact
that multi-whisker stimulation almost invariably
leads to inter-whisker inhibition, we conclude that a
binary touched/untouched signal is the major information format in the barrel cortex, as opposed to
selectivity for complex multi-whisker stimulation
patterns.
There is an unexplained anisotropy of inhibitory
interaction in barrel cortex [19]. Stimulation of
caudal vibrissae inhibits responses to stimulation of
more rostral vibrissae more strongly than rostral
stimulation inhibits responses to caudal whiskers.
The stronger inhibition from long to short whiskers
selectively enforces obstacle (outside)/opening
(inside) contours, and thus, this anisotropy appropriately relates to the anisotropy of the mystacial pad
according to a distance detector theory.
Lesions in the barrel cortex do not lead to a significant change in the behaviorally measured temporal
or behavioral detection thresholds for macrovibrissa
deflection. On the other hand, lesions to the appropriate barrels impair the rat's ability to use the
macrovibrissa for estimating distance in a gap jumping task [ 13]. These results are expected if the barrel
cortex does not generally extract high resolved temporal and deflection information from macrovibrissae inputs but is concerned with extracting spatial
distance information from macrovibrissa inputs.
In rat and mouse macrovibrissae barrel cortex
physiological [25,28] and anatomical evidence has
accumulated for a preferential connectivity between
the barrels of a row, as compared to the connectivity
of barrels between neighboring rows. This type of
connectivity is consistent with the claim that the
rows are functional units of the macrovibrissae array.
The tectal representation of the vibrissae has been
studied in the mouse [6-8] and more recently in
the rat [ 15]. In these neural representations, macrovibrissae are strongly emphasized in areal extent and
are aligned with the overlaying map of the visual field.
The mapping of vibrissal space onto visual space
matches the layout of a distance detector system (a
caudal to rostral coding of eccentricity, with the
rows coding for different angles). The existence of a
precise, very reproducible mapping of macrovibrissae inputs onto visual space is strongly consistent
with the idea that the macrovibrissae code distance
in fixed, head-centered coordinates.

4.3. Conclusion

From morphological and behavioral observations, it

is hypothesized that the mystacial macrovibrissae organ

M. Brecht et al./Behavioural Brain Research 84 (1997) 81-97

functions as distance detector array that derives distance

contours. This concept is very different from traditional
touch-like views of macrovibrissae function. In contrast
to the macrovibrissae, the microvibrissae appear to form
a high resolution tactile sensor and operate in a more
touch-like manner.
It is obvious, that the operational mode of the vibrissal
array cannot be resolved by a single study, and we do
not claim that our data definitely prove the distance
detector theory. We also realize that some may consider
our theoretical conclusions to go beyond the presently
existing data. This, however, we think is acceptable,
because it provides a frame for further studies to
differentially test the competing models of whisker
function.

[ 11 ]

[12]

[13]

[14]
[15]

[16]

[ 17]

Acknowledgement
We like to thank Manfred Ade, Dean Buonomano,
Dan Goldreich, Josh Gordon, Renate Ruhl, Professor
Schnitzler and the Naturkunde Museum MOnster. The
RCS animals were kindly provided by Professor LaVail's
laboratory. Bruno Preilowski was supported by a
Fellowship of the McDonnell-Pew Trusts.

[18]
[19]
[20]

[21]
[22]
[23]

References
[24]
[ 1]Ade, M., Diplomarbeit, Fakult~it for Biologie der Universit~t Tiibingen, 1993.
[2] Ahl, A.S., Relationship of vibrissal length and habits in the sciuridae, J. Mammal., 68 (1987) 848-853.
[3] Andres, K.H., Ober die Feinstruktur der Rezeptoren an Sinneshaaren, Z. Zellforsch., 75 (1966) 339-365.
[4] Carvell, G.E. and Simons, J.D., Biometric analysis of vibrissal
tactile discrimination in the rat, J. Neurosci., 10 (1990) 2638-2648.
[5] Dehnhardt, G., Preliminary results from psychophysical studies
on the tactile sensitivity in marine mammals. In J. Thomas and
R. Kastelein (Eds.), Sensory Abilities of Cetaceans, Plenum Press,
New York, 1990, pp. 435-446.
[6] Dr~ger, U.C. and Hubel, D.H., Physiology of visual cells in mouse
superior colliculus and correlation with somatosensory and auditory input, Nature, 253 (1975) 203-204.
[7] Dr~iger, U.C. and Hubel, D.H., Responses to visual stimulation
and relationship between visual, auditory and somatosensory
inputs in mouse superior colliculus. J. Neurophysiol., 38 (1975)
690 713.
[8] Drager, U.C. and Hubel, D.H., Topography of visual somatosensory projection in mouse superior colliculus, J. Neurophysiol., 39
(1976) 91-101.
[9] Gibson, J.J., The Senses Considered as Perceptual Systems.
Houghton Mifflin Company, Boston, 1966.
[ I0] Goldschmid-Lange, U., Ober die morphologischen Unterschiede

[25]
[26]
[27]

[28]
[29]

[30]
[31]
[32]
[33]
[34]
[36]

97

im Aufbau der facialen Vibrissen einiger S~iugetierarten, Zool.

Anz., 196 (1976) 417-427.
Guic-Robles, E., Valdivieso, C. and Guajardo, G., Rats can learn
a roughness discrimination using only their vibrissal system,
Behav. Brain Res., 31 (1989) 285-289.
Gustafson, J.W. and Felbain-Keramidas, S.L., Behavioral and
neural approaches to the function of the mystacial vibrissae,
Psychol. Bull., 84 (1977) 477-488.
Hutson, K.A. and Masterton, R.B., The sensory contribution of
a single vibrissa's cortical barrel, J. Neurophysiol., 56 (1986)
1196-1222.
Ito, M., Some quantitative aspects of vibrissa-driven neuronal
responses in rat neocortex, J. Neurophysiol., 46 (1981) 705-715.
Kao, C.-Q., McHaffie, J.G. and Stein, B.E., Response properties
and somatotopy of vibrissa-activated neurons in rat superior colliculus, Soc. Neurosci. Abstr., 16 (1989) 223.
Kastelein, R.A. and van Gaalen, M.A., The sensitivity of the
vibrissae of a pacific walrus (Odobenus rosmarus). Part I, Aquatic
Mammals, 14 (1988) 123-133.
Ling, J.K., Vibrissae of marine mammals. In R.J. Harrison (Ed.),
Functional Anatomy of Marine Mammals, Academic Press, New
York, 1977, pp. 387 415.
Luck, S.J., Cognitive and neural mechanisms of visual search,
Curt. Opin. Neurobiol., 4 (1994) 183-188.
McCashland, J.S. et al., Functional asymetries in the rodent barrel
cortex, Somatosens. Motor Res., 8 (1991) 111 116.
Milani, H., Steiner, H. and Huston, J.P., Analysis of recovery
from behavioral asymetries induced by unilateral removal of
vibrissae in the rat, Behav. Neurosci., 103 (1989) 1067-1074.
Patrizi, G. and Munger, B.L., The ultrastructure and innervation
of rat vibrissae, J. Comp. Neurol., 126 (1966) 423--423.
Pocock, R.I., On the facial vibrissae in the mammalia, Proc. Zool.
Soc. Lond., (1914) 889-912.
Schiffman, H.R. et al., Role of vibrissae for depth perception in
the rat (Rattus norvegicus), Animal Behav., 18 (1970) 290-292.
Schmidberger, G., Ober die Bedeutung der Schnurrhaare bei
Katzen, Z. Vergleichende Physiol., 17 (1932) 387-407.
Simons, D.J., Response properties of vibrissa units in the rat SI
somatosensory neocortex, J. Neurophysiol., 41 (1978) 798-820.
Simons, D.J., Temporal and spatial integration in the rat SI
vibrissa cortex, J. Neurophysiol., 54 (1985) 615-635.
Simons, D.J., Neuronal integration in the somatosensory whisker/
barrel cortex. In Diamond and Hall (Eds.), The Barrel Cortex of
Rodents, Plenum Press, New York, 1995.
Simons, D.J. and Woolsey, T.A., Functional organization in
mouse barrel cortex, Brain Res., 165 (1979) 327-332.
Sokolov, V.E. and Kulikov, V.F., The structure and function of
the vibrissal apparatus in some rodents, Mammalia, 51 (1987)
125-138.
Treisman, A. and Gelade, G., A feature integration theory of
attention, Cogn. Psychol., 12 (1980) 97-136.
Vincent, S.B., The function of vibrissae in the behavior of the
white rat, Behavior Monogr., 1 (1912) 1-81.
Welker, C., Receptive fields of barrels in the somatosensory neocortex of the rat, J. Comp. Neurol., 166 (1976) 173 190.
Welker, W.I., Analysis of sniffing of the albino rat, Behaviour, 22
(1964) 223-244.
Wineski, LE., Movements of the cranial vibrissae in the golden
hamster (Mesocritus auratus), J. Zool. Lond., 200 (1983) 261 280.
Woolsey, T.A. and van der Loos, H., The structural organization
of layer IV in the somatosensory region (SI) of mouse cerebral
cortex, Brain Res., 17 (1970) 205-242.