MANIPULATORY FORELIMB MOVEMENTS IN NORMAL AND PYRAMIDOTOMIZED DOGS

ACTA NEUROBIOL. EXF. 1980, 40 : 965-984 MANIPULATORY FORELIMB MOVEMENTS IN NORMAL AND PYRAMIDOTOMIZED DOGS Teresa G~RSKA and Maria ZALEWSKA-WALKOWSKA Department of Neurophysiology, Nencki Institute of Experimental Biology Warsaw, Poland Abstract. Normal and unilaterally pyramidotomized dogs were trained to retrieve food from various test devices to allow an estimation of the importance of the pyramidal tract for forelimb movements. The effects of pyramidotomy were relatively weak on movements which to a least extent required use of distal muscles. They were strongest on movements in which wrist flexion and toes flexion and adduction were necessary to get hold of the food. The latter tasks, however, were also among the most difficult ones for normal dogs. The deficits of movements performed after pyramidotomy were similar as previously found in cats if allowance is made for generally weaker components of distal limb muscles in the dog than in the cat. INTRODUCTION In recent literature an increasing attention is paid to comparison of function of the pyramidal tract in various species (see 31). Sectioning the pyramids in monkeys results in an abolition of independent finger movements, which was related to direct corticomotoneuronal connections in these species (1, 24, 25, 35). However, pyramidal lesions impair movements of distal muscles also in subprimates (4, 13, 20). The pyramidotomized animals are not able to grasp small pieces of food and in some species e.g., cats this deficit is very pronounced and persists for a long time (13). The aim of the present study was to analyze the role of the pyrami-

dal tract in control of movements of distal muscles in the dog. Dogs compared to cats have much less developed digital dexterity, as evidenced by their inability to grasp. It seemed, therefore, interesting to investigate what kind of manipulatory movements could be trained in normal dogs and to compare their performance with that of pyramidotomized animals. Previous experiments with sectioning the pyramidal tract in gags have dealt only with postural reflexes (27) or movements performed mainly in proximal joints (8, 9, 17-19, 33). Some preliminary results have already been published (16). MATERIAL AND METHODS Subjects. The experiments were performed on 10 mongrel dogs of 'both sexes, which at the beginning of the training were 4 mo old. In 5 of them the left medullary pyramid had been transected 1 mo before 1. The animals were trained to retrieve standard sized (1 X 1 X 1 cm) pieces of boiled meat from various test devices with the right paw. Tasks. Seven sets of tests consisting of 38 motor tasks altogether, presented in a sequence which required more and more extensive involvement of distal limb muscles were used in this study. The tasks could be divided into the following 3 classes, 10-16 tasks in each class. A. Easy tasks, requiring no or only small involvement of movements of distal muscles. They consisted of two sets of tests, with 5 tasks in each set. In the first test the animals had to retrieve food from behind vertical bars (Fig. 1, I). Three different spaces between bars were used 2, the most narrow space requiring usually supination of the forearm. Morsels of food were placed at a distance of 5 and 10 cm behind the bars in case of the two wider spaces, and 5 cm away in case of the most narrow space. In the second test, the animals had to retrieve food from between two vertical walls (Fig. 1, 11). The spaces between walls were similar as in test I and the height of walls were 5 and 10 cm in case of two 1 According to our previous experiments (12) the motor skill in dogs does not improve after they reach the age of 3 mo, which is compatible with the data showing that at this age the process of maturation of the motor cortex (10, 11) and of the pyramidal tract (3, 21) is almost completed. 2 In the majority of dogs the width of their paw, measured at the metacarpus ranged between 5.5 and 6.5 cm and for such paws spaces equal 6, 5 and 4 cm were used. For paws 5 cm wide, spaces of 5, 4 and 3 cm were aapplied. The minimum and maximum width of animals' paws at the beginning and the end of the experimental period (8 mo) was 5.0-6.5 and 5.5-6.5 cm respectively, except for one dog (P-11) whose paw grew up to 7.0 cm.

wider spaces and 5 cm in case of the most narrow space. Food was placed 10 cm from the front edge of the device, except for the task of narrow spaced low walls, where this distance was 5 cm. Fig. 1. Test devices used to investigate the manipulatory abilities in dogs. For detailed explanation see text. B. Tasks of moderate difficulty requiring some involvement of movements of distal muscles. They consisted of 12 tasks, in which the animals had to retrieve food out of troughs with a front wall inclined at different angles (Fig. 1, 111). The troughs were of four different depths: 2, 3, 4 and 5 cm, and for each depth 3 different inclinations of the front wall were used: at 45O, 67' and in a form of a quarter of a cylinder (r=d). The animals were first trained to get hold of food put in the most shallow (2 cm) trough with the front wall inclined successively at 45O, cylindrically, and at 67" and then from deeper troughs presented in the same order. C. Difficult tasks requiring most extensive use of distal muscles. They consisted of 4 sets of tests in which the animals had to retrieve food from: (i) holes shaped like inverted truncated cones with walls incllned at 67' (Fig. 1, IV). The cones differed in depth: 2, 3, 4 and 5 cm

and in width: 2 and 4 cm diameter at the bottom, which made 8 tasks in this set altogether. The animals were first trained to get hold of food from larger cones, used in order of increasing depth, and then from narrow cones presented in the same order; (ii) a horizontal tube, 7 cm in diameter and 6 cm long, placed at the height of the animals' shoulder; (iii) oblique tubes of similar diameter 3, 5 and 6 cm deep, placed at the level of the floor, and (iv) similar vertical tubes, 3, 4, 5 and 6 cm deep (Fig. 1, V-VII). In the latter two sets of tests the tubes were presented in an order of increasing depth. The test devices I1 through VII were placed behind vertical bars of a cage as shown in Fig. 5. Experimental procedure. The preliminary training consisted of letting the animals to explore the experimental room and to eat morsels of meat scattered over the floor. Dogs which after a few preliminary sessions still showed signs of strong fear or lack of strong alimentary drive were excluded from further training. During the proper training the experimenter put successively morsels of meat in the experimental devices and the anilmals were allowed to retrieve them with the right paw and eat them. Movements of the left limb were prevented by putting on the paw a shoe-like cuff. If the dog succeeded to take out a piece of meat, such a trial was considered successful, as opposed to unsuccessful trials in which the animal failed to get hold of meat, despite several attempts. The latter trials were also reinforced by food given by the experimenter, in order to prevent extinction of the trained movements (13). However, the number of movements performed in such unsuccessful trials would always somewhat exceed the nulmber of movements performed in successful trials in a given task. In each motor ta& the training was continued to a criterion of 8 successful trials out of 10 consecutive ones or, in tasks in which the animals failed to reach this criterion, despite a relatively long training, to an arbitrarily set total number of movements performed during the training of this task. Their number varied according to the difficulty of tasks and were 50 for all tasks in test I, 100 in test 11, 150 in tests 111, V and VI, and 250 in tests IV and VII. All motor tasks of all sets of tests were trained, independently whether the animals solved the preceding easier tasks, with the exception of the last most difficult set (test VII), where the training with the use of deeper tubes was given up if the animals failed to solve the preceding task with a more shallow tube. Each experimental session consisted of 40-80 trials with one to three se,ts of tests and the anilmals were trained four times a week. The training involving all the 38 tasks lasted 2-6 wk, depending on the animal's dexterity, and in each dog it was repeated 4 times. In the ope-

rated animals the first series of training begun 1 month after the surgery i.e. at the age of 4 mo, and the next ones 3, 5 and 7 mo postoperatively, when dogs had reached the age of 6, 8 and 10 months respectively. Normal animals were trained at the same ages. Indices of motor performance. The animals' motor performance was estimated both quantitatively and qualitatively. The following quantitative indices were used: (i) the percentage of tasks solved i.e, those in which the criterion of 8 out of 10 successful trials was reached; (ii) the lenght of training per task; this index was expressed in the number of movements and was calculated from both solved and unsolved tasks. In the former the number of movements required to reach the first criterion trial was taken into account, in the latter the total number of movements performed before the training of a given task was discontinued (see above); and (iii) the number of movements per trial in the last 10 trials of training of a given task, irrespectively whether this task was solved or unsolved. The use of the number of movements instead of the number of trials in calculating the lenght of training seemed to be a better index of motor performance, since various dogs performed different numbers of movements per trial and, moreover, these numbers depended on the stage of training. At the beginning of the training the animals usually had a tendency to stop to work after a few unreinforced attempts, while later on they gradually learned to perform up to 7-8 movements even if not earlier reinforced. The qualitative analysis of motor performance concerned the accuracy of movements. A movement was considered well-aimed or accurate if the dog hit at once with its paw the piece of meat in the test device. Other movements in which the animals put first the paw too far, to near or aside the morsel of food or did not insert it into the test device were classified as dysmetric. They were divided, in turn, into two categories: corrected and uncorrected dysmecric movements. In the former category the animals, after missing the spot where meat was placed, corrected the error i.e. managed to hit the piece of meat with the paw when, for example, they retracted the limb; in the latter category they failed to do it. In order to evaluate the accuracy of movements for each dog the total percentage of misdirected movements and the percentages of corrected and uncorrected movements were calculated for the last 10 trials of training of each task. The mixed design analysis of variance (26) was used in order to evaluate the overall effects of the pyramidotomy and of the training in the animals' performance. Data expressed in percentages were analyzed after arcsin ~6 transformation. The differences in the results obtained in individual training series were analyzed with the Duncan test (7).

Surgery. The surgery was done in astetic conditions, under Nembutal (Sodium pentobarbital) anesthesia (30 mglkg of body weight) with Fenactil (1 mglkg) and Atropine sulfate (0.005 mglkg) given 30 min. before. The parapharyngeal approach was used, as first described by Starlinger (33), and the left pyramid was transected under a dissection microscope with watchmaker's forceps and a blunt probe bent at a proper angle (9). The level of the trapezoid body was preferably chosen, whenever convenient, in order to reduce the risk of additional damage of the medial lemniscus. Histology. The histological verification of lesions was done on 20 pm thick, paraffin embedded, frontal sections from the medulla and pons. Three successive sections out of each 20 were stained with Nissl, Kliiver-Barrera and Welcke methods. The extent of destruction of the pyramid was specified by measuring the cross sectional area of pyram~dal fibres spared relative to the total area of the opposite intact pyramid. RESULTS Lesions Figure 2 shows the extent and the level at which the pyramidal lesions were made in the 5 operated dogs. The left pyramid was lesioned completely in two dogs (P-I11 and P-IX) and nearly completely in three dogs (in 87 and 83OIo in dogs P-I and P-VIII respectively and in 94O/o in dog P-11). The lesions were with practically no additional damage of the medial lemniscus except in dog P-VIII in which about 115-114 of the left medial lemniscus showed degeneration. In the dog P-I11 the lesion extended too far laterally destroying probably part of the left spinothalamic tract (34). In addition, in two dogs the right pyramid was slightly damaged (in llo/o in dog P-I and g0/a in dog P-111). In the latter dog the reticular formation on the right side also showed degeneration at the level of the lesion and somewhat rostrally to it probably due to a concomitant damaging of blood vessels, when sectioning the pyramid. A comparison of motor performance in individual pyramidotomized dogs will be presented in the end section of the Results. Comparison of motor performance in normal and pyramidotomized dogs The motor performance of both normal and pyramidotomized dogs appeared to be greatly influenced by the difficulty of the tasks. Figure 3 gives a comparison of the animals' performance in the three classes of tasks (A, B, C) of increasing difficulty. For each class of tasks the values obtained in the four indices of performance, described in the Method,

Fig. 2. Extent of medullary lesio<(black) in individual operated dogs. L, left; EL, right side of the medulla. Dots in the diagram for dog P-I11 indicate degeneration of reticular formation. Abbreviations: TB, trapezoid body; SO, superior olive; nv, nvii, nucleus of the V and VII nerve respectively.

Fig. 3. Comparison of motor performance in normal (continuous line) and pyramidotomized (dashed line) dogs in successive training series (I-IV) in 3 classes of tasks (A, B, C) of increasing difficulty. A, tests I and 11; B, test 111; C, tests IV-VII (see Method and Fig. 1). ad, indices of motor performance: a, mean percentage of solved tasks; b, mean number of movements performed in the training of single task; c. mean number of movements per trial, perf'ormed in the last 10 trials of training of each task; d, mean percentage of dysmetric movements performed in the last 10 trials of training of each task. Arrows denote training

and the four successive training series (I-IV) are shown.. The results of the analysis of variance of data presented in Fig. 3 are shown in Table I. A. Easy tasks. The performance of tasks, consisting of retrieving food from behind vertical bars or from between vertical walls (Test I and 11) was very little affected by the pyramidotomy (Fig. 3A). Both normal and operated dogs solved, on the average, 98O/o of these tasks already in the first training series (Fig. 3Aa); they reached the criterion almost immediately (Fig. 3Ab) and required almost identical number of move- Values of F statistics for different sources of variation in the performance of 3 classes of tasks of increasing difficulty (A, B, C) Index of performance Source of variation Easy tasks (A) Tasks of moderate difficulty (B) % of solved tasks 1 Surgery (a) Training Surgery x Training -- - Length of training / Surgery per tasks 1 Training (b) Surgery x I Training ::: Number of movements Surgery 1, 8 0.05 1 5.60 ' 8.50 I --- -- per trial Training 16.86 "3.4-7.41 -- -- - (c) Surgery x 1.04! Training 1 1 1 1.-I I - - % of dysrnetric ) Surgery 11,81 23.53 1 5.72 17.65 movements Training, 3, 24 I (dl I Surgery x j 3, 24 I Underlines denote statistically significant differences: - - -, P < 0.05; - -, P < 0.025; -, P < 0.01; =, P < 0.005; =, P < 0.001. series in which the differences between normal and pyramidotomized dogs were statistically significant, when tested by analysis of variance and Duncan test; thin arrow, P < 0.05; thick arrow, P < 0.01. Asterisks and circles indicate significant differences between the scores obtained by normal (symbols near the continuous lines) and pyramidotomized (symbols near the dashed lines) dogs in two succ6ssive or two nearest training series: circles, P < 0.05; asterisks P < 0.01.

ments per trial to retrieve food (Fig. 3Ac). The solely index of motor performance which showed a statistically significant overall effect of surgery was the total percentage of dysmetric movements (Table I); the differences between the normal and operated animals were limited to the first two training series (Fig. 3Ad). Since tasks included into this group were very easy, the animals' performance was only little influenced by the training. An overall effect of training was observed only with respect to the number of movements per trial and the percentage of misdirected movements (Table I). In the former index a significant decrease in the number of movements per trial was obtained in the second training series both in normal and operated dogs and next training series did not improve the performance any further (Fig. 3Ac). In the latter index (Fig. 3Ad) significant differences between successive training series were obtained in the first two Fig. 4. Mean percentages of two kinds of dysmetric movements performed by normal (N) and pyramidotomized (Op) dogs in the last 10 trials of training of 3 classes (A, B, C) of tasks of increasing difficulty. Black and hatched, uncorrected and corrected dysmetric movements respectively (see Method); white, wellaimed movements. A, tests I and 11; B, test 111; C, tests IV-VII. I-IV, successive training series. Triangles and squares denote training series in which the percentage of corrected or uncorrected dysmetric movements in operated dogs was significantly increased as compared to normal animals when tested by the analysis of variance and Duncan test. Triangle, P < 0.05, square, P < 0.01.

VII Fig. 5. Photographs of movements performed by (a) one of the best normal dog (N-I) and ( b ) one of the worst pyramidotomized dog (P-VIII) in tests I-VII. The photographs were taken after the end of the fourth training series, i.e. in the pyramidotomized dog 8 mo after the surgery. For other explanations see text.

log - N-I 90 - N- V - N-II ' 0 80-70 - 70 60 - GO - 50-50- P-I / - pq&../, f4.., i' /a. PG 4 I i ~$1 v -* / I 11 m IT I II,y N Fig. 6. Comparison of motor performance in individual normal (left) and pyramidotomized (right) dogs. Abcissa, successive training series (I-IV); ordinate, percentage of tasks solved calculated cut of the total 38 tasks used in thz present experiment. Roman numbers denote individual dogs: N, normal; P, operated.

' I training series in normal animals and between the second and the third training series in operated animals. The decreased accuracy of movements observed in pyramidotomized dogs in the first two training series was mainly due to an increase of corrected dysmetric movements, the percentages of uncorrected dysmetric movements, being similar in both normal and operated dogs (Fig. 4A and Table 11). This latter type of movements significantly decreased in the second training series in both groups of dogs and remained thereafter on an essentially unchanged level (Duncan test). On the other hand, the corrected dysmetric movements showed a training effect only in operated dogs: their percentages significantly decreased in the third training series and the fourth series did not influence the animals' performance any further. In normal dogs the percentages of corrected dysmetric movements was very low beginning from the first training series. Values of F statistics for different sources of variation in the performance of two categories of dysmetric movements, corrected and uncorrected, in tasks of increasing difficulty (A, B, C) Source of variation I ''' -- - - - -. - -... -.- Easy tasks (A) Tasks of moderate d~fficulty (B) 1 D~ficult taslcs (C) Corrected 1 UncOr- Coirecicd UnCOr- Corrected Uncorrected, rected 1 rected Surgery Training Surgery x Training -- Underlines denote statistically significant differences: - - -, P < 0.05; - -, P < 0.025; -, P < 0.01;, : - - P < 0.905; =, P < 0.001. Movements perfo'me'd in this group of tasks are illustrated in Fig. 5, I-II. Bo'th normal (a) and operated (b) dogs retrieved food from test' devices by raking it back over the floor. However, operated dogs more often than normal ones overstretched 'the limb (see Fig. 5, Ia and b) or put the paw partly over one of the walls (see Fig. 5, IIa and b). They also hold the paw somewhat more stiff, without any noticeable wrist or toes flexion (Fig. 5, I-II). These deficits were best seen shortly after the surgery, while at the end of the experiments they became almost imperceptible in the majority of dogs even for an experienced observer. 5 - Acta Neurobiol. Exp. 6/80

B. Tasks of moderate difficulty. The performance of tasks, consisting of retrieving food from troughs of different depth with a front wall inclined at a different angle (Test 111), was more influenced by the pyramidotomy. An overall significant effect of surgery was obtained in all the four indices of motor performance (Table I). The operated animals solved less tasks (Fig. 3Ba), despite a longer training per tasks (Fig. 3Bb) and more movements performed in a single trial (Fig. 3Bc). Their movements were also less accurate than in normal animals (Fig. 3Bd). In the first two indices significant differences between the results obtained by normal and pyramidotomized dogs were limited to the first training series (Fig 3Ba and b), in the third index to the fourth series (Fig. 3Bc) and in the fourth index to the second and third training series (Fig. 3Bd). The performance in this group of tasks was also more influenced by the training. An overall effect of training was found to be significant in all indices of motor performance, with no significant interaction between Surgery X Training except for the percentage of solved tasks (Table I). In normal animals the effects of training were limited to the second training series in all the indices except for the percentage of solved tasks, where the animals reached a very high level of performance already in the first training series (Fig. 3Ba-d). In operated animals statistically significant improvement in the performance in the second training series was obtained with respect to the percentage of solved tasks and the lenght of training per task (Fig. 3a and b). In other indices the improvement of performance as a result of training was much less accentuated; no significant differences between the data obtained in successive training series were observed with respect to the percentage of dysmetric movements (Fig. 3Bd), while in the number of movements per trial such differences were obtained only between the first and the fourth (as well as the third) training series (Fig. 3Bc). Likewise in easy tasks the decreased accuracy of movements in pyramidotomized dogs in test I11 was mainly due to an increase in the percentage of corrected dysmetric movements (Fig. 4B and Table 11). The operated dogs performed about twice as much of these movements than normal animals and these differences were significant in all the training series except the first (Duncan test). No effect of training was observed with respect to these movements either in normal or operated dogs. On the other hand, the uncorrected dysmetric movements were not significantly increased in pyramidotomized dogs and their number diminished as a function of training (Table 11). In normal dogs a significant decrease in the percentages of these move-

ments occurred in the second training series, whereas in operated animals in the third and fourth training series. Movements performed in test I11 are illustrated in Fig. 5, IIIa-b. Both normal and operated dogs retrieved food from troughs by raking it up over the ascending front wall. However, the operated dogs more often than normal ones overstretched the limb and/or put it on the side-wall of the trough (Fig. 5, IIIb). They also raked food up without any visible flexion or adduction of foretoes (see Fig. 5, IIIa and b). The most difficult tasks in this test consisted of retrieving food from 4 and 5 cm deep troughs with a front wall inclined at 67" or shaped like a quarter of a cylinder (r=d). The operated animals failed as a rule to solve these tasks in the first training series. On the other hand, they were able from the very beginning to retrieve food from troughs of similar depth but with the front wall inclined at 45", since this latter tasks required less use of distal limb muscles. C. Difficult tasks. The animals' performance in this class of tasks appeared to be much worse than in the preceding ones. Normal animals during the first training series solved on the average 66O/01 of these tasks and reached a level of 90 /a of solved tasks during the fourth training series (Fig. 3Ca). They also required a much longer training per task (Fig. 3Cb) and performed more movements per trial (Fig. 3Cc) than in tasks of moderate difficulty. The percentage of dysmetric movements was likewise the highest in this class (Fig. 3 Cd). The performance of pyramidotomized dogs was inferior to that of normal animals. During the first training series they solved on the average only 45O/o of tasks and reached a level of 71 /o of solved tasks during the fourth training series (Fig. 3Ca). Other indices of motor performance were also worse in pyramidotomized animals (Fig. 3C5-d). The overall effect of surgery was significant in all the indices of motor performance (Table I). The differences between the normal and pyramidotomized animals were most marked in the percentage of dysmetric movements (see F values in Table I) and in this index the performance of operated dogs was significantly inferior in all the four training series (Fig. 3Cd). In other indices of performance statistically significant differences between data of normal and operated dogs were limited to two (Fig. 3Cc), or only one, (Fig. 3Ca and b) training series. An overall effect of training was also significant for all the indices of motor performance, with no significant interaction between Surgery X Training (Table I). The training effect was, however, in genera1 weaker than in tasks of moderate difficulty (see F values for tasks B and C in Table I). In normal dogs a significant improvement in the performance as a result of training was found in two indices only

(Fig. 3Ca and d) and in both of them only data obtained during the first and last training series were statistically different. In operated dogs, a significant improvement was found in three indices (Fig. 3a-c), the last two training series being, in general, the most effective in ameliorating the animals' performance. The differences in the performance of the normal and the operated dogs were particularly marked in those tasks which required the most extensive use of distal muscles. This is illustrated in Table I11 which Mean percentages of tasks so1v.d by normal (N) and pyramidotomized (Op) dogs in successive trainings series in different subsets of tasks classified as difficult (class C) I I,, Successive training series -- Subset of tasks Dogs I I1 1 I11 IV Large cones (4 tasks) Narrow cones (4 tasks) Horizontal and oblique tubes (4 tasks) Vertical tubes (4 tasks) shows the mean percentages of tasks solved by normal and pyramidotomized dogs in successive training series for various subsets of tasks classified as difficult. Tasks consisting of retrieving food from large cones and from horizontal and oblique tubes appeared to be relatively easy and hence the differences between normal and operated animals were relatively small. Normal animals solved 100 /o of these task already during the first training series. The operated dogs solved in the first training series only 75O/oi of these tasks, but their performance showed a tendency for improvement, so that in the fourth training series they achieved a level of 85-100 1o of solved tasks. Retrieving food from the set of narrow cones presented a more difficult task for both normal and operated dogs. Normal animals solved only 65O/o of these tasks during the first training series and in the fourth series they still failed to reach the level of 100 /o of solved tasks. The performance of operated dogs was in the first training series much worse but in the fourth series they similarly reached

a 90 /o level of performance. The difficulty in solving these tasks increased as a function of the depth of cones, the cone 5 cm deep being the most difficult. Retrieving food from vertical tubes was the most difficult set of tasks for normal animals and an almost unsolvable problem for operated ones. None of the normal dogs during the first training series was able to retrieve food from the most shallow (3 cm deep) tube, but in the fourth training they reached the level of 70 /o of solved tasks; three out 5 dogs were then able to retrieve food from tubes of all the four depths (3, 4, 5 and 6 cm) and one dog from the depth of 3 and 4 cm. On the other hand, the pyramidotomized dogs did not succeed to solve any of these tasks during the first three training series. In the last series they reached a level of 10 /o of solved tasks i.e., 2 out of 5 dogs learnt to retrieve food from the most shallow cylinder. The inferior performance of pyramidotomized dogs in this class of tasks and particularly in tasks consisting of retrieving food from vertical tubes resulted both from a decreased accuracy of movements and a deficit of movements of distal muscles. Since in all tasks classified as difficult the animals had to insert the paw into relatively small openings, the percentage of dysmetric movements was much higher in this class, both in normal and operated animals, than in tasks classified as easy or of moderate difficulty (see Fig. 4A-c). Also contrary to latter tasks, the decreased accuracy of movements in pyramidotomized dogs was mainly due to an increase of uncorrected dysmetric movements and not of the corrected ones (Table 11). No training effects were observed with respect to either category of dysmetric movements in operated dogs, whereas in normal animals only uncorrected dysmetric movements showed a significant decrease in the fourth training series (Duncan test). In test of cones pyramidotomized dogs usually overstretched the limb (Fig. 5, IVb) and only when they pulled it back the paw could fall into the opening. In tasks consisting of retrieving food out of the tubes the animals often scraped the external walls of the tubes (Fig. 5, Vb) or their edges (Fig. 5, VI and VIb). One of the factors which additionally interfered with the animals' ability to insert the paw into the tubes was their tendency to spread, instead of adduct, the toes at the final stage of the movement. The functional deficit of distal muscles rendered movements of operated animals also less effective, particularly in task which required the most extensive use of distal muscles. Normal dogs usually retrieved food by raking it up over the front wall of the devices, the toes remaining flexed and adducted (Fig. 5, IV-VIIa). On the other hand, the pyramidotomized dogs, even when they succeeded to insert

the paw into the test devices, usually only displaced the meat toward the front wall, but were not able to rake it up due of impairment of toes ventriflexion. Comparison of motor performance in individual animals The motor performance of operated dogs was not correlated with the completeness of the pyramidotomy, but rather with the extent of damage of other structures. Figure 6 shows the percentages of the 38 tasks solved by individual normal and operated dogs in successive training series. In the dog P-IX with complete pyramidotomy the level of performance was even slightly better, particularly in the first training series, than in dogs with uncomplete lesions, sparing 6-17OIo of pyramidal fibers (P-I, P-11, and P-VIII, see Fig. 2). On the other hand, in the dog P-I11 in which the lesion extended too far laterally and, moreover, was associated with some degeneration within the reticular formation (see Fig. 2) the performance was much inferior, particularly shortly after surgery. The percentage of dysmetric movements was also very high in this animal. However, in further training series its performance considerably improved, so that at the final stage of training it reached a similar level that the majority of other operated animals. In the dog P-VIII with uncomplete pyramidal lesion and partial degeneration of the medial lemniscus, the percentage of dysmetric movements was also relatively high. However, it is doubtful whether damage of the medial lemniscus was the main reason of the animal's poor performance, particularly in the second to fourth training series. Rather it might have been due to same undefined individual factors, since the dog P-IX with the lesion restricted to the pyramid and the normal dog N-I1 did not improve their performance either in successive training series. DISCUSSION The results of the present study indicate that in the dog, as in other species (1, 4, 13, 20, 25, 35), the pyramidal tract controls movements of both distal and proximal muscles but it is particularly important for distal muscles. Tasks in which digital movements were not necessary were only little affected by pyramidotomy, whereas tasks requiring extensive use of distal muscles were considerably affected and the differences between the normal and operated animals persisted till the end of the experiments, i.e., for 8 mo after surgery. These

latter tasks were, however, very difficult for normal animals too, due to a little elaborated digital dexterity in the dog. The behavioral deficits observed on the present study are in agreement with the results of our previous experiments (15) in which the motor cortex was stimulated in unilaterally pyramidotomized dogs. Stimulation of the hemisphere with the transected pyramid failed tc yield movements of foretoes flexion during the first five months after the surgery. Movements of wrist ventriflexions were also considerably affected. Other forelimb movements were not noticeably impaired, except for their increased thresholds. These latter changes diminished as a function of the lenght of postoperative period, so that six months.after surgery thresholds on the affected hemisphere approximated the values obtained on the intact hemisphere. A similar time-table of recovery were obtained in the present experiments. Five to six months after surgery the performance of the pyramidotomized animals was any longer inferior than in normal dogs, except for those tasks which required the most extensive involvement of distal muscles. Although sectioning the medullary pyramids in dogs affected their manipulative abilities, the results of such lesions were of much less consequence than in other species with better developed digital skill. In monkeys, pyramidal lesions abolish the ability to move fingers separately (1, 24, 25, 35). The operated animals were not longer able to use apposition of the thumb and index for picking up small items or to retrieve food from deep and narrow holes with the index finger. No recovery of these movements was observed in adult pyramidotomized monkeys for one year after surgery (1, 25, 35) and in animals operated as infants, the relatively independent finger movements did not appear in the first 3 years of life (24). These deficits have been attributed to interruption of direct corticomotoneuronal connections assumed to provide the main anatomical substrate for independent finger movements f(2, 22, 24, 31, 32). In subprimates, in which the ability to move digits separately is in general lacking, pyramidal lesions also result in an impairment of movements of distal muscles, proportional to the degree of digital skill in various species. Sectioning the pyramid in rats resulted in their inability to grasp small food pellets, observed for at least 3 mo after surgery (4). In the hamster pyramidal lesions also produced an impairment of the manipulative abilities of the paw, and even one year after the surgery these movements were less dextrous than on the normal side (20). A pronounced effect of pyramidotomy on movements of distal muscles was found in cats (13). The operated animals were unable to retrieve small pieces of food from narrow horizontal and vertical tubes for at least one year after surgery. In

other easier tasks the operated cats, instead of grasping food, as they did before surgery, used to rake it back over the floor and this behavior reminded the way of solving similar tasks by normal dogs. The diminished motor skill was correlated with an increase in thresholds for movements elicited from the hemisphere with the transected pyramid (14). The rudimentary character of movements of distal muscles in the dog together with a relatively little impairment of these movements after pyramidotomy as compared with cats, suggest that the anatomophysiological organization of the pyramidal tract is different in these two species. Comparative data concerning the organization of the pyramidal tract in the dog are, however, scanty and limited to such questions as the cortical origin of pyramidal fibres (27, 28), their number (23, 27), conduction velocity (21), site of termination in the spinal gray (3) which are, in general, similar to the cat (5, 6, 29, 30). To what extent other features of the organization of the pyramidal tract in the dog and the cat are similar remains to be determined. In the present study the contribution of the pyramidal system to forelimb manipulatory movements in the dog was evaluated in animals on which the pyramids had been transected at the age of 3 months. This age was deliberately chosen since in the dog the process of maturation of the pyramidal tract (3, 21) and of the motor cortex (10, ll) was found to last about 3 mo. In addition, our previous experiments (12) suggested that the manipulatory dexterity in the dog did not improve after they reached the age of 3 mo. It was, therefore, assumed that the effects of exclusion of pyramidal influences, after this system had achieved its full maturity should be similar irrespectively of the animals' age. In our next paper, dealing with the problem of age determinants in motor impairment after pyramidotomy in dogs (in preparation) it will be shown that pyramidal lesions performed in older dogs do not produce greater motor deficits than in animals operated on at the age of 3 mo. The authors are indebted to Dr. P. Jastreboff for his helpful advices concerning the statistical analysis of the material and wish to express their thanks to Mrs. J. Leoniak and Mrs. I. Malinowska for their technical assistance. This investigation was supported by Foreign Research Agreement No 05.001.0., annex 284A of the U.S. Department Health, Education and Welfare under PL 480. REFERENCES 1. BECK, C. H. and CHAMBERS, W. E. 1970. Speed, accuracy and strenght of forelimb movement after unilateral pyramidotomy in Rhesus monkeys. J. Comp. Physiol. Psychol. 70: 1-22.

2. BERNHARD, C. G. and BOHM, E. 1954. Cortical representation and functional significance of the corticomotoneuronal system. Arch. Neurol. Psychiatr. 72: 473-502. 3. BUXTON, D. F. and GOODMAN, D. C. 1967. Motor function and the corticospinal tracts in the dog and racoon. J. Comp. Neurol. 129: 341-360. 4. CASTRO, A. J. 1972. Motor performance in rats. The effects of pyramidal tract section. Brain Res. 44: 313-323. 5. CHAMBERS, W. W. and LIU, C. N. 1957. Cortico-spinal tract of the cat. An attempt to correlate the pattern of degeneration with the deficits in reflex activity following neocortical lesions. J. Comp. Neurol. 108: 23-56. 6. CREVEL, H. van and VERHAART, W. J. C. 1963. The "exect" origin of the pyramidal tract. A quantitative study in the cat. J. Anat. (Lond.) 97: 495-515. 7. DUNCAN, D. B. 1955. Multiple range and multiple F test. Biometrics 11: 1-42. 8. GAMBARJAN, L. S., GARIBJIAN, A. A. and OGANISJIAN, S. S. 1964. A method of sectioning pyramidal tracts in dogs (in Russian). Izv. Akad. Nauk Arm. SSR. 17: 23-27. 9. GORSKA, T. 1967. Instrumental conditioned reflexes after pyramidotomy in dogs. Acta Biol. Exp. 27: 103-122. 10. GORSKA, T. 1974. Functional organization of cortical motor areas in adult dogs and puppies. Acta Neurobiol. Exp. 34: 171-203. 11. GORSKA, T. and CZARKOWSKA, J. 1973. Motor effects of stimulation of the cerebral cortex in the dog. An ontogenetic study. In A. A. Gydikov, N. T. Tankov and D. S. Kozarov (ed.), Motor control. Plenum Press, New York, p. 147-166. 12. GORSKA, T. and CZARKOWSKA, J. 1978. Motor cortex development in the dog. Some cortical stimulation and behavioral data. Neuroscience 3: 129-131. 13. GORSKA, T. and SYBIRSKA, E. 1980. Effects of pyramidal lesions on forelimb movements in the cat. Acta Neurobiol. Exp. 40: 843-859. 14. GORSKA, T., SYBIRSKA, E., CZARKOWSKA, J. and ABRAHAM, W. 1975. Changes in the excitability of cortical motor areas after lesions of the lateral descending system in cats (in Polish). XI11 Congr. Polish Physiol. Soc., Gdansk 1975, p. 106 (Abstr.). 15. GORSKA, T., WOOLSEY, C. N. and WETZEL, A. 1980. Motor effects of cortical stimulation after chronic lesions of medullary pyramid in the dog. Acta Neurobiol. Exp. 40: 861-879. 16. GORSKA, T. and ZALEWSKA-WALKOWSKA, M. 1979. Effects of pyramidectomy on manipulatory movements in dogs. Acta Physiol. Pol. 30: 73-74. 17. IOFFE, M. E. 1973. Pyramidal influences in establishment of new motor coordinations in dogs. Physiol. Behav. 11: 145-153. 18. IOFFE, M. E. 1973. Supraspinal influences on spinal mechanisms activated prior to learned movement. Acta Neurobiol. Exp. 33: 729-741. 19. IOFFE, M. E. and SAMOILOV, M. I. 1971. Interaction of pyramidal and rubrospinal systems in the regulation of local movements (in Russian). In P. G. Kostyuk (ed.), Mekhanizmy nishodyaschego kontrolya aktivnosti spinnogo mozga. Nauka, Leningrad, p. 140-143. 20. KALIL, K. and SCHNEIDER, G. E. 1975. Motor performance following unilateral pyramidal tract lesions in the hamster. Brain Res. 100: 170-174.

KIRK, G. R, and BREAZILE, J. E. 1972. Maturation of the corticospinal tract in the dog. Exp. Neurol. 35: 394-397. LANDGREN, S. C. G., PHILLIPS, C. G. and PORTER, R. 1962. Cortical fields of origin of the monosynaptic pyramidal pathways to some alpha motoneurones of the baboon's hand and forearm. J. Physiol. 161: 112-125. LASSEK, A. M. and RASMUSSEN, G. L. 1940. A comparative fiber and numerical analysis of the pyramidal tract. J. Comp. Neurol. 72: 417-428. LAWRENCE, D. G. and HOPKINS, D. A. 1976. The development of motor control in the Rhesus monkey: Evidence concerning the role of corticomotoneuronal connections. Brain 99: 235-254. LAWRENCE, D. G. and KUYPERS, H. G. J. M. 1968. The functional organization of the motor system in the monkey. I. The effects of bilateral pyramidal lesions. Brain 91: 1-14. LINDQUIST, E. F. 1953. Design and analysis of experiments in psychology and education. Riverside Press, Cambridge, p. 393. MAFFRE, S. 1955. Etude physiologique et morfologique du faisceau pyramidal chez le chien d'apres les cons6quences de son exclusion "chronique". M. Lecomte, Marseille, 170p. MORIN, G., POURSINES, Y. and MAFFRE, S. 1951. Sur l'orgine de la voie pyramidale. Documents obtenus par la methode des degenerescences descendantes chez le chien. J. Physiol. (Paris) 43: 75-96. NYBERG-HANSEN, R. 1969. Further studies on the origin of corticospinal fibres in the cat. An experimental study with the Nauta method. Brain Res. 16: 39-54. NYBERG-HANSEN, R. and BRODAL, A. 1963. Sites of termination of corticospinal fibres in the cat. An experimental study with silver impregnation methods. J. Comp. Neurol. 120: 369-391. PHILLIPS, C. G, and PORTER, R. 1977. Corticospinal neurones. Their role in movement. Academic Press, London, 450 p. PRESTON, J. B. and WHITLOCK, D. G. 1961. Intracellular potentials recorded from motoneurones following precentral gyrus stimulation in primate. J. Neurophysiol. 24: 91-96. STARLINGER, J. 1895. Die Durchschneidung beider Pyramiden beim Hunde. Neurol. Zbl. 14: 390-394. SINGER, M. 1962. The brain of the dog in section. W. B. Saunders Co., Philadelphia, 124, XL p. WOOLSEY, C. N., GORSKA, T., WETZEL, A,, ERICKSON, T. c., EARLS, F. J. and ALLMAN, J. M. 1972. Complete unilateral section of the pyramidal tract at the medullary level in Macaca mulatta. Brain Res. 40: 119-123. Accepted 15 June 1980 Teresa GORSKA and Maria ZALEWSKA-WALKOWSKA, Nencki ~nstitute of Experimental Biology, Pasteura 3, 02-093 Warsaw, Poland.