Acta Physiol Scand 1990, 139, 371 385
Change in the pattern of behavioural specialization of
neurons in the motor cortex of the rabbit following
lesion of the visual cortex
YU. I. ALEXANDROV, YU. V. GRI NCHENKO and T. JARVILEHTO*
Institute of Psychology, Laboratory of Neurophysiological Basis of Mental Activity, USSR
Academy of Sciences, Moscow, USSR, and *Department of Behavioural Sciences, University of
Oulu, Finland
ALEXANDROV, YU. I., GRINCHENKO, YU. V. & JARVILEHTO, T. 1990. Change in the
pattern of behavioural specialization of neurons in the motor cortex of the rabbit
following lesion of the visual cortex. Actµa Physiol Scand 138, 371-385. Received 4
January 1989, accepted 29 November 1989. ISSN 0001-6772. Institute of Psychology,
Laboratory of Neurophysiological Basis of Mental Activity, USSR Academy of
Sciences, Moscow, USSR, and Department of Behavioural Sciences, University of
Oulu, Finland.
In order to find out whether damage of the visual cortex (area 17) of the brain results
in a functional reorganization of the motor cortex, experiments were carried out with
freely moving rabbits performing a food acquisition task in an experimental cage. Two
rabbits served as controls, while in three rabbits the visual cortex was bilaterally
damaged. Analysis of the activity of 575 neurons in the control and operated rabbits
after the recovery of the original instrumental food acquisition behaviour revealed a
marked difference in the behavioural specialization of the neurons in the motor cortex
of two operated rabbits compared with the control animals. Although the same types of
units as in the control rabbits could be found in the operated rabbits (M neurons
activated in relation to body and limb movements, S neurons activated in relation to
food seizure and L neurons activated in relation to learned food acquisition task), the
number of S units was about half of that in the controls and the number of L units about
double. The relative number of activations of the neurons in the operated rabbits was
significantly less frequent during the food seizure and more frequent during the learned
behaviour. This difference indicates a change in the pattern of behavioural specialization
of the neurons in the motor cortex due to the damage of the visual cortex. In this
reorganization, the motor cortex became more like (but not identical to) visual and
limbic cortices that normally contain noticeably more L neurons than the motor cortex.
The number of neurons activated in relation to the behaviour in the operated rabbits,
as compared with the control animals, was smaller in the upper and larger in the lower
layers of the motor cortex. This may indicate recruitment of new neurons from the lower
cortical layers.
Key words: damage of the visual cortex, motor cortex neurons, single-unit activity in
behaviour, rabbits, recovery from brain damage.
Investigations of the effects of local lesions of
various brain structures are widely used for
increasing our understanding (1) of the role of a
Correspondence: Yu. I. Alexandrov, Institute of Psychology, USSR Academy of Sciences, Yaroslavskaya
13, 129366 Moscow, USSR.
given brain structure in the normal behaviour of
the organism and (2) of recovery mechanisms
after brain damage. These goals are interde
pendent; in order to understand normal func tioning of the brain we must use knowledge
about recovery processes in the experiments
with brain damage and, on the other hand, our
371
372
Yu. I. Alexandrov et al.
understanding of recovery mechanisms is itself
determined by the prevailing views on the normal
functioning of the brain (Laurence & Stein
1978).
From the point of view of the functional
systems theory (Anokhin 1974), recovery from
brain damage means regaining the ability to
achieve useful behavioural results in the organism—environment interaction which is realized by coordinated activity of neural elements
belonging to different brain structures. This
implies that local brain damage is followed by a
general reorganization of neural activity that
embraces not only the structures closely connected, both morphologically and functionally,
with the lesioned structure, but also distant
structures belonging to other 'morphofunctional' systems. This reorganization enables recovery of the ability to achieve the same adaptive
effect as that before the lesion.
In contrast to the above position, it is held
that the brain (of an adult at least) is incapable of
functional reorganization; recovery of the behaviour altered by brain damage is thought to be
based on neuronal processes that were spared in
the damage (Spear & Baumann 1979, Le Vere
1980).
The general task of the present work was to
determine whether or not functional reorganization follows brain damage. From our
point of view, the role played by a certain
structure in behavioural acts is determined by
the behavioural specialization pattern of its
neurons (Alexandrov 1989), i.e. by the particular
set of the types of the neurons, and the relative
number of the neurons belonging to these types,
whose activity is related in a specific manner to
certain behaviour. The activity of these neurons
reflects realization of particular sets of functional
systems necessary for given behaviour (Shvyrkov
1986).
The aim of the present study was to find out
what, if any, changes occur in the pattern of
behavioural specialization of the neurons of such
a ' distant' structure as motor cortex after damage
of the visual cortex of the freely moving rabbit
carrying out a food acquisition task. The present
work is the first one in which changes in the
functional characteristics of neurons were studied
after brain damage in freely moving animals.
MATERIALS AND METHODS
Subjects. Five experimentally naive adult rabbits of
both sexes (weight 23 kg) were used in the study and
kept in separate cages in the vivarium with a 12-h
light-dark cycle.
Experimental procedure. Freely moving animals were
taught to acquire food by pressing one of the two
pedals in the experimental cage (60 x 60 cm, see Fig.
1). Pressing a pedal activated a feeder at the same wall
of the cage. The training procedure (2 weeks) ended
when a rabbit's behaviour at both the front and the
rear walls of the cage had become repeatedly cyclic
with the following behavioural phases: the rabbit
pressed a pedal, turned to the corresponding feeder,
lowered the head, seized the food, lifted the head,
turned to the pedal, approached the pedal, pressed
it, etc. Only one pedal was effective at any instant.
After completion of the training the striate cortex
was bilaterally lesioned (see later) in three rabbits (N
1, 2 and 3). Two intact rabbits (N 4 and 5) served as
controls. After the operated rabbits had recovered
(4-5 days after the operation), i.e. they regained the
ability to obtain food, and the duration and variability
of duration of successive phases of the behavioural
cycle returned to the preoperative level, unit activity
was recorded (see later) for 4 days in the motor cortex
during the recovered food acquisition behaviour. The
data yielded by the operated rabbits were compared
with those obtained from the same region of the
control animals during the same behaviour.
The choice of the two brain areas was based on the
following considerations:
(1) These areas of the rabbit may be clearly
distinguished and with a high degree of certainty
correlated with Brodman's areas 4 and 17 (Kappers
et al. 1936, Blinkov et al. 1973).
(2) In our earlier studies we gathered many data on
characteristics of the neuronal activity in these areas
(especially in the motor cortex) in a similar ex
perimental situation as in the present study (Alex
androv et al. 1984, Shevshenko et al. 1986, Alexandrov
1989).
(3) The motor cortex may be considered 'distant'
with respect to the visual cortex not only in the spatial
sense, but also in relation to the anatomical connections
between these two cortical areas; ablations of the
striate cortex do not result in any significant de
generation in the motor cortex (Shkolnik-Yaros 1965,
Montero & Murphy 1976, Towns et al. 1977).
Cortical lesions. All surgery was performed under
sterile conditions. The rabbits were anaesthetized
with Nembutal (40 mg kg l). Cortical tissue was
aspirated from the striate cortex region VI, determined
according to Rose & Rose (1933) and Moore &
Murphy (1976), with a surgical aspirator after
retraction of the overlying meninges.
Recording techniques. Activity of the neurons was
recorded from the anterolateral part of the motor
cortex (A 4.0; L 4.0) during a microelectrode
penetration. Stimulation of this region resulted in
marked lower jaw movements. Glass microelectrodes
with 2.5м КС1, tips of 1-3 µm diameter and im-
Neural reorganization and brain damage
373
Fig. i. Experimental cage and actographic marks of food acquisition behaviour. (A) Rabbit in the
behavioural cycle at the rear wall; (B) scheme of the cage: i, 4, pedal and feeder of the front wall;
2, 3, pedal and feeder of the rear wall; 5, 8, light indicators of head lowering into a feeder and pedal
pressing respectively; 6, 7, digital displays of the timer and neuronal discharge counter
respectively. (C) Diagram of the relation of the segments of an actogram to phases of the
behavioural cycle (for abbreviations, see text).
pedance of 15 MQ. at 1.5 kHz were used and driven
by a micromanipulator.
Electrical activity of m. masseter pars profundus
was recorded bipolarly by wire electrodes implanted in
the muscle. Pedals and feeders were equipped with
photoelectric devices. The changes of their potential
served as actographic marks of the phases of the
behaviour (see Fig. 1).
Unit discharges, EMG and actographic marks of
the behaviour at the front and the rear walls (two
separate channels) were tape-recorded. Rabbits' behaviour was simultaneously video-recorded (the audio
channel of the videotape recorder was used to record
simultaneously neuronal spike activity). In the front
wall of the cage (Fig. 1) were light indicators of the
pedal pressing (8) and head lowering (5), and two
counters, one (7) triggered by the spikes from the
recorded unit and the other (6) by a 50-Hz pulse
generator (timer). Timer impulses were also taperecorded, which enabled synchronization of the unit
activity with the video recording.
Behavioural and neural analysis. The following
behavioural phases were defined in the food acquisition
cycle (see Fig. 1): approach to the pedal (AP), pedal
pressing (PP), approach to the feeder (AF), lowering
the head into the feeder (LIF), food seizure (FS), food
grinding (FG), lifting the head from the feeder (LFF)
and regular chewing (RC).
Behavioural deficits after the lesions were evaluated
on the basis of the ability to obtain food from the
feeder by pressing the pedal and on the basis of the
changes (compared with the preoperative period) in
the duration and variability of duration of the
behavioural phases. Differences in the latter two
indexes were estimated by t- and F-tests, respectively,
and were considered significant if P < 0.05.
The number of neurons whose activity could be
detected was monitored in each penetration. Unit
activity was analysed by plotting rasters and histograms with reference to different behavioural phases
determined by the actographic marks and EMG. In
addition, we calculated, by means of the readings of
the timer and impulse counter in the stop-frame mode
of the video recorder, the number of impulses during
each of the successive 20-ms intervals with reference
to the different phases of the behaviour of an animal.
The data obtained from the successive realizations of
the behavioural cycles were then summed, thus
yielding histograms plotted in relation to the behavioural events which were not indicated by the
actographic marks.
The vertical localization of the recorded neurons
was determined within one of the three parts (А, В, С)
of a track that could be related to the higher,
intermediate and lower layers of the cortex respectively. Because of the variability of the vertical structure
in the cortex we treated only A and С parts of a track
as definitely different in respect of their relation to the
cortical layers: the superficial and deeper layers.
For each unit the background frequency and the
374
Yu. I. Alexandrov et al.
ratio of the background frequency to the frequency of
the discharge during an activation was calculated.
Differences in the distribution of these parameters
were tested by the Kolmogorov Smirnov test. Activations were estimated during the different phases of
the behaviour by relating increases of the firing
frequency to the background discharge. Activation
was considered to be present if a discharge appeared
(in a unit without background activity) or there was a
marked increase in the firing frequency (in the units
with background activity) during a behavioural phase
in all repeated realizations of the behaviour. A change
in the firing frequency was considered to be an
activation when it deviated by more than two standard
deviations from the mean background frequency of a
given unit.
The first step in the classification of the units was
based on the relation of their activation to the
behavioural phases: units related and non-related to
the different phases of the behavioural cycle. The first
group was further divided into three subgroups which
were activated in relation to the different phases of the
behavioural cycle (see Results). This classification was
possible on the basis of the two cycles of the food
acquisition behaviour and by the use of some
additional tests. The two cycles of the food acquisition
behaviour enabled a comparison of the unit activity
related to the achievement of a given goal (e.g.
'contact with the pedal', 'contact with the food') in
the different environments (at the front vs rear wall)
and by means of opponent movements (turns to the
right vs to the left). On the other hand, behavioural
acts in the different behavioural cycles that pursued
different goals could be characterized by similar
movements (e.g. a rabbit's movement to the left in the
cycle at the front wall during its approach to the
pedal). The following additional tests were used: a
passive displacement of the rabbit's body, movements
subsumed under the defence behaviour, orienting and
searching, exploring the empty feeders and taking the
food out of the experimenter's hand from above (to
take the food the animal had to lift its head) at
different positions in the cage.
The significance of the differences in the number of
units belonging to a classification group and the
significance of the differences in the number of units
belonging to different groups was estimated by the /\'2test. Differences were considered significant if P <
0.05.
Morphological control. Upon the completion of the
experimental series, the rabbits were killed with an
overdose of Nembutal and perfused with 0.9% saline
and 9% formalin. Forty-micrometre serial frozen
sections were stained according to Nissl and sketched
with the help of a microscope. The degree and borders
of the lesions were determined according to the atlases
of Rose & Rose (1933), Blinkov et al. (1973) and
McBride & Klemm (1968).
RESULTS
Extent of lesions
The morphological control demonstrated substantial lesions of the striate cortex in all three
rabbits (see Fig. 2, right). The area peristriata
and the occipital cortex (in rabbit N 3, right
hemisphere) were also partly damaged. The
depth and all the other characteristics of the
lesions were similar in all operated rabbits. Some
representative sections of the brain of the rabbit
N 1 are presented in Fig. 2 (left column). All the
layers of the cortex appeared to be lesioned, the
damage affecting even the central white matter.
The lesion produced a cavity which remained
open in the lower cortical layers, compensatory
gliosis scar tissue forming only on the surface.
Recovery of the food acquisition behaviour
There were no statistically significant differences
in the preoperative behaviour (duration and
variability of the duration of the different
behavioural phases) between the operated and
control rabbits. In spite of some variations in the
amount of the damage, behavioural disorders
and the dynamics of the disappearance of the
behavioural deficits were similar in all the
operated animals.
Throughout the first post-operative day, no
instrumental food acquisition behaviour was
observed in the operated rabbits. On day 2 all
the rabbits started to perform the food acquisition
task, but the duration of the phases of the
behavioural cycle and the variability of the
duration of these phases significantly exceeded
the preoperative level. Also, the mean duration
and the variability of the duration of the approach
to a pedal, pressing a pedal and the approach to
the feeder significantly exceeded the preoperative
level during day 3, but on day 4 both returned to
the preoperative level (data not shown). After
the recovery the sequence of movements was the
same as before the operation.
Neurophysiological data
Activity of 575 neurons was analysed (274 in the
control and 301 in the operated animals). The
mean number of the neurons that could be
detected in a microelectrode track was similar in
both groups: 27.9 ± 7.1 and 24.1±8.0 respectively.
Neural reorganization and brain damage
Fig. 2. The extent of the cortical lesion in operated rabbits. Right: borders of the damaged zones
in rabbits N i, N 2 and N 3, reconstructed on the basis of serial sections plotted on the schematic
drawing of the rabbit cerebral cortex (dorsal view). Cytoarchitectonic fields are shown according
to Rose & Rose (1933): Rsg, regio retrosplenialis granularis: Pstr, area peristriata; Str, area striata;
375
376
Yu. I. Alexandrov et al.
The relation between the number of behaviourally related and non-related neurons was
not different in the operated animals as compared
with the control rabbits. The latter had 130 units
(47%) ш tne behaviourally related group and
the rest 144 (53%) in the behaviourally nonrelated group. In the operated rabbits these
groups numbered 138 (46%) and 163 (54%)
cells, respectively.
Histograms of the distribution of the background frequency and of the ratios of the
discharge frequency during activation to the
background frequency in the operated rabbits
did not differ significantly (χ²-test) from those
obtained from the control animals.
Types of behavioural specialization of neurons in
the control and operated rabbits
The first subgroup of the behaviourally related
units consisted of neurons displaying activation
in relation to a particular movement of the body,
head or lower jaw, irrespective of the result
sought or of the environment of the behavioural
act involving these movements (M neurons).
These neurons were found in both the control
and operated animals (Fig. 3). Some units showed
activations related to horizontal or vertical
movements of the body and/or the head, others
could fire bursts unselectively in relation to any
movement. The M subgroup also included those
neurons that were activated in relation to
rhythmic lower jaw chewing movements. These
neurons were observed only in the control
rabbits.
The second subgroup (S neurons) consisted of
neurons which were activated in relation to
seizing and/or grinding the food irrespective of
the movements and the environment, but showed
no activation related to chewing movements of
the lower jaw or to head movements other than
those involved in the seizure of the food. Also,
these neurons were found in both the control
and operated rabbits (Fig. 4). The unit presented
above in Fig. 4 displayed activation during the
food seizure in the behavioural cycle at the front
wall (A), at the rear wall (B) and during the food
seizure from the experimenter's hand (C). No
other lower jaw movements were accompanied
by an activation. The unit presented below was
activated in relation to the food seizure and food
grinding up to the regular chewing in the food
acquisition cycle at the front wall (A), at the rear
wall (B) and with the food seizure from above,
from the experimenter's hand (C).
The third subgroup (L) consisted of neurons
which displayed activations only in the behavioural phases formed in the learning process
in the experimental cage. These acts included
the approach to and/or pressing of the pedal and
the approach to a given feeder. The activations
were independent of the means of the realization
(e.g. the act could involve opposing movements;
its environment could vary, etc.). Moreover,
these units did not display any activation related
to the same movements when these were
observed during any other behaviour. Figure 5
presents examples of the subgroup L neurons of
the control and operated rabbits, which were
activated in relation to the approach towards and
pressing of a pedal in the both behavioural
cycles. Figure 6 shows a neuron which displayed
activation during one behavioural cycle only.
Differences in the pattern of behavioural
specialization of neurons in the control and
operated rabbits
The most characteristic feature of behavioural
specialization pattern observed in the motor
cortex of both control rabbits was the great
prevalence of S units. For the operated rabbits N
1 and 2, the distribution of the number of
behaviourally related neurons was different from
that in the control animals. The number of S
units in the operated rabbits N 1 and 2 was
about half of that seen in the control animals,
while the number of L units was about double
(Fig. 7).
Number of activations related to the different
behavioural phases in the control and operated
rabbits
Differences in the behavioural specialization
pattern of the neurons of the control and operated
rabbits were reflected also in the number of
activations associated with the successive phases
of the food acquisition behaviour (see Fig. 1).
Figure 8 presents the 'overall picture' of the
activations of the neurons during the behavioural
cycle, which is based on the calculated percentage
of the activations occurring during each of the
successive behavioural phases in relation to the
total number of activations observed in the
control and operated rabbits. The relative number
Fig. 3. Examples of the activity of M subgroup neurons, activated during leftward body movement. Neuronal activations related to the leftward
movement in the front-wall (A) and rear-wall (C) cycles, during the transition from the front-wall to the rear-wall cycle (B) and from the rear-wall to
the front-wall cycle (D-CONTR). D-OPER, activation accompanying the animal's shift to the left by experimenter (marked by oblique arrows). On
the histograms, plotted by means of videotape analysis (see Methods): abscissa, time in seconds; ordinate, number of impulses; bin, 80 ms. Arrows mark
the onset of the leftward body movement that served as a reference point for histogram plotting. Here and in other figures: CONTR, neuron of control
rabbit; OPER, neuron of operated rabbit; EMG, electromyogram of m.masseter; ACTOGR. FRONT, actogram of behaviour in the front-wall cycle;
ACTOGR. REAR, in the rear-wall cycle; IAN, impulse activity of neuron. Actogram symbols as in Fig. 1.
Fig. 4. Examples of the activity of S subgroup neurons that were activated in relation to the act of food taking. Above: neuron of control rabbit that
was activated only in relation to actual food seizure. (A) Activity in the front-wall cycle followed by probing an empty feeder. (B) Activity in the rearwall cycle. (C) Activity during the food seizure from experimenter's hand. Below: neuron of control rabbit that was activated in relation to food seizure
and grinding. (A) Activity in the front-wall cycle. (B) Activity in the rear-wall cycle followed by probing an empty feeder. (C) Activity during the food
seizure from experimenter's hand. All rasters below unit activity (each line of points shows one trial) are plotted with reference to the onset of the EMG
burst in the m.masseter that corresponded to the first mouth closing during the food seizure.
Fig. 5. Examples of the activity of L subgroup neurons in control (above) and operated rabbits (below). Units were activated during an approach to
and pressing of the pedal in the front-wall (A) and rear-wall (B) cycles. Rasters are plotted with reference to the instant of releasing the pedal.
Fig. 6. Examples of the activity of L subgroup neurons that were activated in only one behavioural cycle. Above: neuron of a control rabbit that was
activated during an approach to and pressing of the pedal in the front wall cycle; below: neuron of an operated rabbit that was activated during an
approach to and head lowering into the feeder in the front wall cycle. (A) Activity in the front wall cycle; activity in the rear-wall cycle. C-CONTR, orienting
behaviour (both pedals are inoperative), crossing from the rear-wall cycle to the front wall of the cage and, after pedal pressing, food seizure from
experimenter's hand.
Neural reorganization and brain damage
381
DISCUSSION
Fig. 7. Relative number of the M, S and L neurons in
control and operated rabbits. Ordinate: percentage of
the units belonging to the given specialization type,
expressed as a percentage of all units analysed in
control animals N 4 and 5 (n = 274), in operated
rabbits N 1 and 2 (n = 183) and in operated rabbit N
3 (n = 118). Significance of the difference in relative
number of units of a given type in operated rabbits
compared to that in the control animals: * P < 0.05,
**P< 0.001.
of the activations during the food seizure and
food-grinding phases was significantly smaller in
the operated rabbits N 1 and 2 as compared with
the control rabbits. Conversely, the number of
activations at the learned phases of the behavioural cycle was significantly larger. Modifications of the specialization pattern in the
operated rabbit N 3 were limited to a smaller
number of units that were activated during the
chewing. The activity observed in the rabbits N
1 and 2 at this phase is due to tonic activations
of S neurons during the chewing.
Distribution of behaviourally related and nonrelated neurons along the vertical axis of the
cortex
The control rabbits showed no significant
difference between the number of behaviourally
related and non-related units in the parts of the
microelectrode track or in the lower and upper
layers of the cortex (about 30% neurons in each
part). The operated rabbits displayed a significant difference between the number of behaviourally related and non-related neurons in
both the upper (A; P < 0.01) and lower (С; P <
0.05) cortical layers. This difference was mostly
due to a smaller number of behaviourally related
neurons in the A segment of a track (about 20%,
on average) and to a larger number of behaviourally related neurons in the С segment
(about 45 %, on average). The difference in the
number of the behaviourally related units in
these layers was statistically significant (P <
0.001).
Significant differences in the pattern of behavioural specialization of neurons in the motor
cortex were observed in the operated rabbits N
i and 2 as compared with the control rabbits.
These differences in the specialization pattern
cannot be accounted for by changes in the motor
structure of behaviour, because it was not
changed after the operation (as compared to the
controls). Neither can they be explained by
changes of any parameters of movements or
environmental information available. Our earlier
results show that the behaviourally related
neurons of the motor cortex are activated at a
given phase of the behaviour in spite of the
variability of its actual motor realization that is
observed even in the simplest acts and in spite of
the marked differences in the available environmental information (Alexandrov & Alexandrov 1982, Alexandrov et al. 1984, Alexandrov
1989). The operated rabbit N 3 showed no
significant difference to the control rabbits in the
number of units belonging to M, S and L
subgroups.
An analysis of the data obtained in the earlier
studies on the neuronal activity of the central
and peripheral structures correlated with the
history of formation of the behaviour indicates
that the neurons participate in the realization of
functional systems that have been formed at
successive stages of the phylo- and ontogeny: the
realization of a set of systems of different ' ages'
that subserve the achievement of the useful
behavioural results (Alexandrov & Alexandrov
1982, Shvyrkov 1986, Alexandrov 1987, 1989).
According to the criteria developed and described
in the latter articles for the selection of units
featuring different behavioural specialization,
units can be classified as belonging to functional
systems of different 'ages'. The neurons of the L
subgroup belong to the new systems, those of the
S subgroup to the older systems (formed at early
stages of the individual development) and the
neurons of the M subgroup to the oldest systems.
In comparison with the control rabbits, in the
operated rabbits a smaller number of S units and
a larger number of L neurons was found. In
spite of a certain variability of the functional
morphology of the cerebral cortex, unit activity
recordings from the anterolateral part of the
motor cortex during various forms of food
acquisition behaviour always yielded in our
382
Yu. I. Alexandrov et al.
Front —w a l l
c y c l e
R e a r
w a ll
c y c l e
Fig. 8. 'Overall picture' of activations of units in control and operated rabbits in the front- and
rear-wall cycles. Ordinate: number of activations expressed as a percentage of the total number
of activations observed in behaviourally related neurons in a given behavioural cycle in control
rabbits N 4 and 5 (front-wall cycle n = 202, rear-wall cycle n = 203), in operated rabbits N 1 and
2 (front-wall cycle n = 101, rear-wall cycle n = 97), in operated rabbit N 3 (front-wall cycle n =
103, rear-wall cycle n = 104). Significance of the difference in the relative number of activations
related to a given behavioural stage in operated rabbits as compared with control rabbits: *P<
0.05, **P < 0.01, ***P < 0.005.
earlier studies a similar pattern of behavioural
specialization that was characterized by the same
quantitative dominance of S neurons (Alexandrov 1989). The same number of neurons
detected in the microelectrode track, the same
frequency of the background activity and the
constant frequency observed within activation
bursts demonstrated in the operated rabbits,
together with the literature showing insignificant
morphological changes in the motor cortex after
ablation of the visual cortex (see Methods),
would suggest the following hypothesis: The
differences found in the pattern of behavioural
specialization of the neurons of the motor cortex
between the control and operated rabbits are not
due to a massive cell death or degenerative
supersensitivity in the operated rabbits. Rather
they are due to a functional reorganization of the
pool of the neurons studied,
Since the quantitative relation between the
behaviourally related and non-related neurons
was the same, while the number of behaviourally
Neural reorganization and brain damage
related neurons was smaller in the upper and
larger in the lower layers of the cortex, we may
assume that the differences observed in the
specialization pattern were not due simply to
selective exclusion of the behaviourally related
neurons within this group. The difference was
probably due to the fact that new cells obtained
behavioural specialization and mostly in the
lower cortical layers.
The neurons of the motor and visual cortex
belong to the different, but partly overlapping,
sets of functional systems. Hierarchically coordinated action of these systems subserves the
achievement of useful behavioural results (Alexandrov 1989). After the damage of the visual
cortex the same result as before the damage may
be achieved through a reorganization of the
hierarchy of functional systems. This reorganization manifests itself as a change of the pattern
of the behavioural specialization of the neurons
in the motor cortex which lost a considerable
portion of its specificity — the marked prevalence
of S neurons in the normal rabbits above all
other specialization types - and became more
like (but not identical to) other cortical areas,
such as the visual and limbic cortex, that
normally contain significantly more neurons
belonging to the new systems than the motor
cortex (Alexandrov 1989, Shevskenko et al.
1986).
Recovery processes following brain damage
can be compared with those occurring during
learning (Laurence & Stein 1978). The present
results indicate certain neuronal aspects which
are similar between learning (Gorkin 1987) and
recovery: in both the behavioural recovery and
learning, changes in the pattern of the initial
specialization entail a relative decrease in the
number of S subgroup neurons, as well as the
recruitment of new nerve cells. Where recruited
new neurons come from is not clear. Are they
' silent' neurons brought up from the reserve (as
is the case with learning a new act), behaviourally
non-related neurons or neurons having some
other behavioural specialization but not activated
during the food acquisition behaviour in the
normal animals? All that can be stated now is
that this reserve of new behaviourally related
neurons may be situated in the lower layers of
the motor cortex.
The fact that the involvement of the motor
cortex neurons in the subserving behaviour
actually change is clearly demonstrated by the
383
differences in the 'overall picture' of the
activations of the neurons, which shows an
increase of the activity during the newly acquired
behavioural acts and a decrease during the food
seizure in the operated as compared with the
control animals. The change in the role played
by a distant structure - the motor cortex - in the
subserving behaviour may be considered to
represent at least one of the possible bases of the
recovery, i.e. recovery of the initial behavioural
pattern observed after the lesion of the striate
cortex. This change is based on the reorganization of the unit activity, including a change in
the behavioural specialization pattern of the
neurons, as well as of the localization of the
neurons involved in the behaviour.
This conclusion would appear to contradict
the results obtained by Spear & Baumann (1979),
who demonstrated that the receptive fields of the
neurons in the suprasylvian visual area were
modified immediately after the ablation of fields
17-19 and did not exhibit any other change
during the recovery of the visual discrimination
behaviour. From these results Spear & Baumann
(1979) concluded that functional reorganization
plays at most a negligible role in the recovery
after the lesion of the visual cortex of the cat.
The authors noted, however, that they could not
rule out entirely the possibility that changes
might have occurred in some parameters of unit
activity that were not investigated.
We believe that, as far as the behavioural
recovery is concerned, these as yet uninvestigated
parameters must include the behavioural relation
of the activity of the neurons examined. Our
belief is based on the established fact that the
specific features of the behavioural relation of
this activity cannot be predicted on the basis of
the properties of their receptive fields alone
(Chapin & Woodward 1982, Alexandrov &
Grinchenko 1984, Alexandrov 1989).
Furthermore, even if we restrict ourselves to
the criterion of the unchanged receptive fields,
Spear & Baumann's (1979) conclusion is not true
for all types of lesions, structures and behavioural
patterns tested. Several studies have revealed
gradual changes in the receptive fields of the
neurons in the intact structures, changes that
underlie the recovery of specific behavioural
patterns after a brain damage varying in location
and in the method of the lesion (Senatorov &
Silakov 1985).
Neuropsychologists and physiologists are well
384
Yu. I. Alexandrov el al.
aware of the broad interindividual variability of
the recovery mechanisms after one brain area has
been damaged. The principal factor determining
interindividual variability of the effects of brain
damage is thought to be the specificity of the
individual development (Luria 1973, Stein et al.
1983), which determines premorbid characteristics of the neuronal specialization.
Individual specificity may account for the
noticeably fewer changes in the pattern of
behavioural specialization observed in the operated rabbit N 3, as compared to the operated
rabbits N 1 and 2, though the dynamics of the
behavioural recovery (according to the analysed
indices) was similar in all three operated rabbits,
and the extent of the lesion in the rabbit N 3 was
at least not smaller than in the rabbits N 1 and
2. While the results of the present study clearly
point to the unique reorganization processes in
the rabbit N 3, there is not enough evidence to
account for this anomaly. Three explanations, no
one of which excludes the others, may be
suggested:
(1) More marked changes in other structures
occurred in the operated rabbit N 3 than in the
rabbits N 1 and 2.
(2) Some functional differences existed in the
preoperative state of the motor cortex studied.
(3) The change in the behavioural role of the
motor cortex units in the operated rabbit N 3
was different in some other parameters not
investigated in our study.
In any case the fact that changes in the
behavioural specialization of the units in the
motor cortex were found in two rabbits shows
that functional reorganization of the brain is
possible after cortical damage in freely moving
rabbits.
The authors are grateful to Dr V. N. Matz for her
valuable assistance in morphological control. The help
of Dr L. I. Alexandrov in the preparation of the
English version of the manuscript is also gratefully
acknowledged.
REFERENCES
ALEXANDROV, YU.I. 1987. Systemic specialization of
the neurons of various cortical areas and its
phylogenetic changes. In: V. J. A. Novak & J.
Mlikovsky (eds.) Towards a New Synthesis in
Evolutionary Biology, pp. 266—267. Czeckoslovak
Academy of Science, Praha.
ALEXANDROV, YU.I. 1989. Psychophysiologkal Significance of the Activity of Central and Peripheral
Neurons in Behaviour. Science Press, Moscow (in
Russian).
ALEXANDROV, YU.I. & ALEXANDROV I.O. 1982. Specificity of visual and motor cortex neurons activity in
behavior. Ada Neurobiol Exp 42, 457-468.
ALEXANDROV, YU.I. & GRINCHENKO, YU.V. 1984.
Activity of somatosensory and visual cortex neurons
related to the testing of their receptive fields and to
the realization of food acquisition behaviour.
Neurophysiology 16, 254-268 (in Russian).
ALEXANDROV, YU.I., GRINCHENKO, YU.V., SHVYRKOV,
V.B., SAMS, M. & JARVILEHTO, T. 1984. Behavioral
specificity of motor cortex units in freely moving
rabbits. Ada Psychol Fenn X, 3-15.
ANOKHIN, P.K. 1974. Biology and Neurophysiology of
Conditioned Reflex and its Role in Adaptive Behaviour. Pergamon Press, Oxford.
BLINKOV, S.M., BRAZOVSKAYA, F.A. & PUCILLO, M.V.
1973. Atlas of the Rabbit Brain. Medicine Press,
Moscow (in Russian).
CHAPIN, J.K. & WOODWARD, D.S. 1982. Somatosensory transmission to the cortex during movement : gating of single cell responses to touch. Exp
Neurol 78, 654-669.
GORKIN, A.G. 1987. Behavioral specialization of
cortical neurons at different stages of learning. In:
V. B. Shvyrkov, V. M. Rusalov & D. G. Shevshenko (eds.) EEG and Neuronal Activity in
Psychophysiologkal Studies, pp. 78-81. Science
Press, Moscow (in Russian).
KAPPERS, C.U.R., HUBER, G.C. & CROSBY, E.C. 1936.
The Comparative Anatomy of the Nervous System of
Vertebrates Including Man. vol. 11. Macmillan, New
York.
LAURENCE, S. & STEIN, D.G. 1978. Recovery after
brain damage and the concept of localization of
function. In: S. Finger (ed.) Recovery from Brain
Damage: Research and Theory, pp. 369-407. Plenum
Press, New York.
LE VERE, Т.Е. 1980. Recovery of function after brain
damage: a theory of the behavioral deficit. Physiol
Psychol 8, 297-308.
LURIA, A.R. 1973. Basics of Neuropsychology. Moscow
University Press, Moscow (in Russian).
MCBRIDE, R.L. & KLEMM, W.R. 1968. Stereotaxic
atlas of the rabbit brain based on the rapid method
of photography of frozen, unstained sections.
Commun in Behav Biol Part A 2, 179-215.
MONTERO, V.M. & MURPHY, E.N. 1976. Corticocortical connections from the striate cortex in the
rabbit. Anat Rec 184, 486-493.
MOORE, D.T. & MURPHY, E.N. 1976. Differential
effects of two visual cortical lesions in the rabbit.
Exp Neurol 53, 21-30.
ROSE, M. & ROSE, S. 1933. Die Topographie der
architectonischen Felder der Grosshirnrinde am
Kaninchenschaddel. J Psychol Neurol 45, 264-276.
SENATOROV, V.V. & SILAKOV, V.L. 1985. Long-lasting
Neural reorganization and brain damage
reorganization of neuronal activity in neocortical
field 7 after unilateral section of the visual radiation
in the cat. Neurophysiology 17, 50—57 (in Russian).
SHEVSHENKO, D.G., ALEXANDROV, YU.I., GAVRILOV,
V.V., GORKIN, A.G. & GRINCHENKO, YU.V. 1986.
Comparison of unit activity in different cortical
areas in behavior. In: V. B. Shvyrkov (ed.) Neurons
in Behavior: Systemic Aspects pp. 23-25. Science
Press, Moscow (in Russian).
SHKOLNIK-YAROS, E.G. 1965. Neurons and Interneuronal Connections. Visual Analyser. Medicine
Press, Leningrad (in Russian).
SHVYRKOV, V.B. 1986. Behavioral specialization of
neurons and the system-selection hypothesis of
learning. In: Human Memory and Cognitive Capa-
385
bilities, pp. 599-611. Elsevier Publishers, Amsterdam.
SPEAR, P.D. & BAUMANN, T.P. 1979. Neurophysiological mechanisms of recovery from visual cortex
damage in cats: properties of lateral suprasylvian
visual area neurons following behavioral recovery.
Exp Brain Res 35, 177-192.
STEIN, D.G., FINGER, S. & HART, T. 1983. Brain
damage and recovery: problems and perspectives.
Behav Neural Biol 37, 185-222.
TOWNS, L.C., GIOLLI, R.A. & HASTE, D.A. 1977.
Cortico-cortical fiber connections in the rabbit
visual cortex: a fiber degeneration study, J. Comp
Neurol 173, 560-573.