an object moving slowly in different ways in front of him or her. In the middle of each session there was also a 1-minute period with no object present. The results.
Copyright 1984 by the American Psychological Association, lnc
Developmental Psychology 1984, Vol 20, No 3. 378-388
Developmental Changes in the Organization of Prereaching Movements Claes von Hofsten University of Uppsala, Uppsala, Sweden How does the kind and amount of prereaching activity change with age during the period preceding successful grasping of objects, that is, during the first 4 months oflife? This question was studied longitudinally in 23 infants. They were seen every third week from the first week of life to 16 weeks of age. Twelve of the infants were also seen at 19 weeks of age. At each session the subject was presented with an object moving slowly in different ways in front of him or her. In the middle of each session there was also a 1-minute period with no object present. The results show that the amount of prereaching goes down at 7 weeks and that this decrease is contingent on the presence of the object. The infant does not seem to lose interest in the object at this age, but attending to the object inhibits prereaching activity in some way. The form of prereaching changes at this age, too. Instead of opening the hand during the forward extension of the arm the hand will be fisted. After this age, the amount of prereaching activity will go up again and the hand will start opening during the forward extension, but only when the infant looks at the object.
A neonate, placed on its back in a horizontal or semiupright position will, if it is alert and has support for trunk and head, engage in rather extensive spontaneous movements of its arms and hands. These movements show a specific patterning, for example, the forward extension of the arm is often accompanied by an opening of the hand (von Hofsten, 1982; Trevarthen, 1974). This synergistic extension of arm and hand contrasts with the passive traction reflex where the arm is extended forward by pulling it at the wrist (Twitchell, 1965). In the latter case, all muscles will flex, including those of the fingers, that is. the hand will get fisted. Fisted hands are rare in spontaneous forward extensions of the arm. DiFranco,
Funds for this investigation were provided by grants to the author from the Swedish Council for Research in the Humanities and Social Sciences. I wish to thank the mothers and nurses at Ringblomman, Akademiska Sjukhuset, Uppsala, Sweden, for their helpful cooperation. I also wish to thank Annica Lindgren-Rydberg for her experimental assistance, patient scoring of videotapes, and computation of data. Finally, I am indebted to Herbert L Pick, Jr for his useful comments on the manuscript Requests for reprints should be sent to Claes von Hofsten, Department of Psychology, University of Uppsala, Box 227, S-75104 Uppsala, Sweden.
Muir, and Dodwell (1978) found that only 8% of neonates arm movements could be so described. That the arm movements of the neonate may come under visual control has been revealed through careful analysis of the aiming of arm movements in three-dimensional space (von Hofsten, 1982). It was found that these movements, performed while the infant fixated an object, were aimed closer to that object than movements performed while the infant looked elsewhere or closed her eyes. However, the aiming was not very precise. Furthermore, the infant never grasped the object and rarely touched it. The fact that some aspects of reaching, like the oriented movements of the arm, may be controlled by vision in the neonate but not others, like the catching, grasping, and manipulative movements of the hand, makes it justified to call this activity prereaching (Trevarthen, 1974). The questions asked in the present article concern the way in which prereaching movements change with age during the period preceding successful grasping of objects, that is, during the first 4 months oflife. Such changes may carry information about restructurings or reorganizations of the reaching activity and, thus, also about developmental changes in the
378
DEVELOPMENTAL CHANGES OF PREREACHING MOVEMENTS
maturing nervous system. Two types of suggestions of developmental discontinuity will be considered. Amount of Activity
379
continuous way. In a series of articles, Mounoud (Mounoud, 1976, 1981; Mounoud & Vinter, 1981) has argued that the coordinated, prestructured movement patterns of the neonate become dissociated into isolated activities during the second and third month of life. For instance, in the case of prereaching, the movements of the arm and the hand should no longer be coordinated. The decomposition of activity is said to be a necessary part of an adaptive restructuring process. The constituent elements or local programs are elaborated individually before they enter into new compositions of behavioral organization. Through this process, the activities of the infant and the properties of reality acquire meaning that will enable the infant to control her actions better. Observations by Trevarthen (1974) and others (Cunningham, 1979; White, Castle, & Held, 1964) suggest that the extension synergy of neonatal prereaching is broken up in the second and third month of life. Trevarthen (1974) found that during this period there is a loss of fluency in the extension movements of the fingers. "The grasp movement may be lost, while vigorous, usually jerking arm extensions are made as the object is intensely fixated" (p. 577). However, Trevarthen has another explanation of the phenomena than Mounoud. He believed (Trevarthen, Murray, & Hubley, 1981) that the observed blocking of hand opening depends on a proportionally much greater growth of the proximal limbcontrol system than of the distal-hand control system during the first months of life. In the present longitudinal study, which covers the whole prereaching period, amount as well as type of prereaching activity was measured.
A classical hypothesis states that neonatal motor activity should disappear before reappearing again in a more mature form (McGraw, 1943). The described sequence is said to be due to the takeover of behavior by cortical centers. When this happens, the centers previously regulating the behavior will first be brought under control by inhibition before excitatory connections are formed. From that standpoint, Humphrey (1969) argued that prereaching activity should disappear shortly after birth. Existing data (Cunningham, 1979; McDonnell, 1979; Trevarthen, 1974) do not support a hypothesis of complete disappearance, but the activity might well diminish. There is, to my knowledge, no study of the complete age range that has looked at amount of activity. However, even if the amount of prereaching diminishes during some period, this does not necessarily mean that Humphrey's assumptions are right. There are at least two other ways of explaining such an effect. Bower (1974) believed in such a decrease in activity but thought that it was due to the fact that neonatal prereaching is seldom successful. In other words, the activity will be temporarily extinguished from lack of reinforcement. Another way to explain a decrease in activity is in terms of the focus of activity. There is an extensive social development during the second and the third month of life with, for instance, the appearance of the smile. It is quite possible that the emerging social activity will, for a while, make the infant less focused on objects and prereaching. It should be noted that only Method Humphrey's (1969) account for an eventual decrease in amount of prereaching postulates Subjects any reorganization of these. The two alterNine male and 14 female infants completed the lonnative explanations are both based on as- gitudinal program. The first recording of each infant was sumptions of changes in the behavioral bias made at a clinic in Uppsala, Sweden, to which mothers can go from the delivery ward with their newborns on the of the infant. Kind of Activity Prereaching activity may remain throughout the period but change in a qualitative and dis-
third day after delivery They usually stay 5 days. At the clinic, which is of rooming-in type, the mothers rest and learn to nurse and breast-feed their infants. Staying at the clime is cost-free. The mothers represent a fair cross section of the Uppsala population. At the time of the study, between 8 and 11 mothers were simultaneously attending the clinic.
380
CLAES VON HOFSTEN
Only those infants with Apgar scores between 8 and 10, who were born at a concepuonal age between 38 and 42 weeks, and had a birth weight between 2750 and 4500 g were considered for the longitudinal program All the subjects were born between January 4 and February 9, 1980. Two of the infants were monozygotic twins (G.Ji and G.Ma.).
Stimulus The object presented to the infants was a spherical tuft made of a bright red, blue, and yellow yarn. It was hanging down from the end of a 70 cm long horizontal hollow rod attached via a felt coupling to the perpendicular shaft of an electric motor with variable speed and direction Thus, the object moved along a horizontal circular path of 140 cm diameter. During testing the motor was placed straight behind the infant. The attached rod moved above the infant's head, and the object hung down at the height of the infant's eyes, at a nearest distance of approximately 12 cm from them (see Figure 1)
Experimental Conditions In the experiment the object either moved back and forth in front of the infant between a position about 25 cm to the left and 25 cm to the right or remained stationary straight ahead The motion of the object was either smooth or irregular. The irregularity of the motion was created by having an experimenter starting and stopping the motor about once a second When in continuous motion the object had a velocity of either 3.2 cm or 6.4 cm per sec The irregularly moving object covered only about half the distance of the smoothly moving object in 1 s The chair A chair was constructed from sketches provided by Trevarthen (1979), it can be seen in Figure 1. The chair allowed free movement of the infant's arms while supporting the head and trunk. A 13-cm-wide band of linen was laced around the infant's chest, just under the armpits, to hold him or her securely The inclination of the chair was 50°. Recording arrangements To allow three-dimensional analysis of the hand movements, the situation was recorded by two SONY AVC-3250 CE video cameras placed as shown in Figure 1, with the axis of one camera parallel to the axis of the infant's body (50° inclination) and the axis of the other camera perpendicular to the first one
Procedure The experiment was divided into two blocks that were identical, with the exception of object velocity. In one of the blocks the lower object velocity was used and in the other the higher. Between the two blocks, there was a 1minute period during which the object was absent.1 Which block to present first was randomly determined In each block the object passed six times in front of the infant. The object moved alternatively from the left to the right and from the right to the left. Starting direction was randomly determined. The sequence of movements was always the same. The first time the object passed in front of the infant it moved smoothly and the second time
Figure 1 Experimental situation.
it moved irregularly At the third passage it moved smoothly until straight ahead, where it stopped for about 30 s before continuing its interrupted motion. In the middle of the stopping period, the object was shaken gently for about 10 s After this sequence of motions another identical sequence of smooth, irregular, and stopping motions followed This time the directions of the motions were reversed. Thus, in each block, the object made four smooth-motion passages in front of the infant (2 complete and 2 interrupted) and two irregular-motion passages As the irregularly moving object only covered half the distance of the smooth-moving one per time unit, the exposure time for the two types of motions were about the same The exposure time for each type of motion was 60 s in the slowvelocity block and 30 s in the fast-velocity block In each of the two blocks the object remained stationary straight ahead, for a total period of 60 s The longitudinal program At the first session the infants were between 4 and 8 days old (Mdn = 6 days). After the first session the infants were seen every third week until age 16 weeks Fourteen of the infants were also seen at 19 weeks The remaining 9 infants joined the study too late to be included in the final session. Summer vacation had already started then Ten infants missed one of the first six sessions, two at 1 week, one at 4 weeks, two at 7 and 10 weeks, and three at 16 weeks.
1 The original plan was to present a static object in front of the infant during the middle 1-minute period. However, as it was impossible to keep the object totally stationary, this condition was replaced by the object-absent condition but only after the investigation had started. Thus, five infants were, at the first recording, presented with a "stationary" object at the middle 1-minute period.
381
DEVELOPMENTAL CHANGES OF PREREACHING MOVEMENTS
Results For each infant and session, every arm movement extending forward more than 7.5 cm from the frontoparallel plane tangential to the body of the infant was scored from the videotape, except those originating from Moro reflexes or startles and those associated with yawning, sneezing, or sudden forward head movements. The task was rather simple as the 7.5 cm borderline was clearly indicated on the TV screen. There was one problem with the older infants. They sometimes tracked the moving object with their extended arms and hovered around the stationary object. In such cases, two movements were scored if the forward extension exceeded 4 s, three movements if the forward extension exceeded 8 s, and so forth. The intracoder reliability for this task was 0.94. Three infants scored no forward extensions at all for the first three sessions of the study. They were omitted from further data treatment for that reason. For the remaining 20 infants, 1,954 forward extensions were scored altogether. Each forward extension was further scored with respect to the looking behavior of the infant and the posture and movement of the hand. The following four categories of looking behavior were used. 1. Fixation (F). The infant directed his or her eyes toward the target area. If the object moved, the infant changed gaze in accordance with the motion of the object. 2. Nonfixation (NF). The eyes were opened but the infant was not looking at the object. 3. Eyes closed (CL) The eyes were closed. 4. Indeterminate gaze (IG) The eyes were opened, and the object was within the infant's field of view, but it was uncertain whether the infant was looking at the object or not. The behavior of the hand was scored in one of the following four categories: 1. Fisted (f). All the fingers were fully flexed before and during the reach. 2. Half-open (ho). The fingers were semiflexed before and during the reach. 3. Open before (o). The fingers were extended before the reach started. 4. Opening during (od). Two or more digits clearly extended during the forward extension of the arm.
Table 1 Mean Number of Forward Extensions per Subject When the Object Is Present and Absent Distributed Over Age and Types of Looking Behavior Object present Age (weeks) 1 4 7 10 13 16 19
n
F
IG
NF + CL
Total
18" 13" 19 18 18 20 17 12
25 2 1 3.8 1.2 7 1 9.7 18.3 36 6
09 09 1.1
7.7 88 41 3.5 2.3 14 1.8 2.0
11.2 11.8 9.1 5.6 11 3 12.0 21.2 38 9
0.8 1.9 10 1.2 0.3
Object absent — 1.9 2.1
2.2 1.7 1.0 1.7 2.3
Note F = fixation, IG = intermediate gaze, NF = nonfixation, CL = closed eyes, n = number of subjects. * At one week of age, only 13 of the 18 subjects received the object-absent condition Therefore, two sets of means have been calculated, one for the whole group and one for those who received the object-absent condition.
Reliability of scoring was evaluated in two ways. The forward extensions of the three youngest age groups were classified independently by two observers. Interobserver reliability was found to be 0.91 for looking behavior and 0.92 for hand behavior. Intraobserver reliability for the main coder (A.R.), measured for four randomly chosen infants of each of the four oldest age groups, was above 0.90 for all the scored properties. Number of Forward Extensions The distribution of forward extensions over age and looking behavior is shown in Table 1. In Table 1 movements while not fixating the object (NF) and movements while the eyes were closed (CL) are pooled. Eyes closed while reaching was very rare except at the first week. The number of such cases at that age was 76. At the remaining six sessions, the total number of closed-eye reaches was on the average of only 4 per session. Table 1 shows that the number of forward extensions increases considerably during the last three sessions. Before that, the number of movements stays rather stable. The decrease at 7 weeks is not statistically significant. However, the number of movements while fixated
382
CLAES VON HOFSTEN
gives a different picture for the first four sessions. These movements increase in number from the first to the second session (t = 2.15, p < .05) after which there is a considerable decrease up to the third session (t = 3.34, p < .01). From there on the number of movements while fixated increases steadily up to the last session. Finally, it can be observed from Table 1 that the number of movements in the object-absent condition stays pretty much the same over age. An analysis of variance (ANOVA) showed that there was no systematic effect of age in this condition, F{6, 110) < 1. In Table 2 the total number of reaches as well as the number of fixated ones are shown for each subject and session. The upper part (subjects with B as first initial) of Table 2 shows the results of the 8 boys and the lower part (subjects with G as first initial) the results of the 12 girls of the study. The individual differences are great. One subject (G.N.) made only one fixated reach before 19 weeks of age, and one subject made as many as 121. However, there are also consistencies between subjects. Fourteen of the 18 subjects who partic-
ipated at 7 weeks showed a decrease in the number of fixated movements. Eleven subjects had no fixated movements at all at 7 weeks compared to 3 subjects at 4 weeks. The increase in reaching after 7 weeks is rather dramatic for most infants. Generally, it occurs after a period of practically no reaching. At the session when the number of reaches performed by the infant starts to increase, the number of forward extensions goes up from, on the average, 4.5 to 21.7 movements per session. However, as can be seen from Table 2, this increase in reaching activity starts at different ages for different infants. For 8 infants it occurred at 10 weeks of age, for 6 infants at 13 weeks of age, for 5 infants at 16 weeks of age and for one infant at 19 weeks of age. The effect of the different experimental conditions on amount of prereaching activity for different ages is shown in Figure 2. In Figure 2 the smooth and irregular motions of the target have been pooled for each velocity conditions. The differences in frequency of reaching between these two types of motions were small and unsystematic. Figure 2 shows that the decrease in reaching ac-
Table 2 Total Number of Movements and Number of Fixated Movements for Each Subject and Age (in Parentheses) Age (weeks) Subject
BT. B.F. B.S. B.M. B.B. B.W. B.A. B.G. G.E. G.N. G.M. G.L. G.Mo. G.Ka. G.Ha. G.Li. G.Ni. GJu.
GJi. G.Ma. Total
7
10
13
16
19
26(5) 15(9) 14(5) 19(3) 16(9) 17(5) 25 (10) 4(0)
4(0) 17(4) 8(0) 8(4) 0(0) 17(0) 0(0) 14(4)
29(11) 25(15) 8(1) 2(2) 19(11) 28(17)
10(5) 9(9) 1(0) 20(12) 18(8) 43 (39) 0(0) 6(0)
13(5) 12(6) 40(39) 31 (29) 3(0) 58 (55) 13(11) 10(4)
— — 35 (33) 20(18) 78 (73) 87 (80) 19(13)
11 (2) 8(1) 37(2) 10(3) 24(6)
4(1) 1(0) 10(5) 2(2)
—
9(5) 20(6) 5(2) 2(0) 6(2) 14(2)
4(2) 6(1) 10(5) 4(4) 21(5) 8(2) 6(0)
21(0) 3(0) 12(2) 0(0) 0(0) 4(0) 6(2) 14(2) 3(0) 9(4)
0(0) 2(0) 0(0) 0(0) 0(0) 34 (24) 4(1) 34(27)
225 (45)
212 (73)
1
4(0) 10(2) 2(0) 8(2) 30(4) 17(5) 8(1) —
4
—
—
9(1)
— —
4(4) 13(7) 23(7)
1(0) 0(0) 0(0) 22(16) 8(8) 4(4) 12(9) 1(0) 25(16) 3(0) 18(3) 58 (54)
140 (22)
234(128)
259(193)
—
2(2) 18(14) 40(40) 12(5) 16(9) —
19(13) 17(9) 23 (23) 63 (47) 390(311)
Total 86 (26) 88 (45) 73 (45) 123 (85) 106 (50) 258 (194) 133 (102) 62 (22)
—
37(3) 73 (55) 85 (27) 56 (38) 72 (54) 58 (35) 129(101) 93 (50) 133 (100) 57 (22) 68 (37) 164(110)
494 (439)
1954(1211)
59 (54) 24 (16) 4(3) 76 (74) 14 (10) 77 (65) 1(0)
DEVELOPMENTAL CHANGES OF PREREACHING MOVEMENTS
tivity at 7 weeks is contingent upon the presence of the object, as there is no decrease in the frequency of movements during the objectabsent interval at that age. The 7-week session is the only one in the study where the reaching activity in the object-absent condition is higher than in the other conditions (/ = 2.11, p < .05). Furthermore, Figure 2 shows that the Stop condition elicited more reaches per minute than the motion conditions at all ages from 4 weeks on and that the slow motion elicited more reaches than the fast one, at least from 10 weeks on. A means ANOVA (Kirk, 1968) showed that the main effect of object motion
'
383
was statistically significant, F\2, 12) = 7.57, p < 0.01. The Stop condition should not be confused with a static-object condition. It is, rather, a transfer from translatory motion to jiggling motion. This change in event seems to be an efficient elicitor of reaches. As shown by the present data, forward-extended arm movements does not cease altogether when the infant does not look at the object or when the object is not in front of the infant. Rather, there seems to be some base-level activity that is independent of condition. To appreciate the effect of different types of object conditions on the reaching ac-
REACHES/MIN. STOP
SLOW
FAST 5
NO OBJECT (BASELINE)
10
13
16
19
AGE(WEEKS) Figure 2 Number of forward extensions per minute for the fast, slow, and stop conditions of the experiment as a function of age. BASE is the number of forward extensions per minute m the no-object condition.
384
CLAES VON HOFSTEN
Table 3 Estimated Percentage Time Spent Looking at the Object and Frequency of Reaching During That Period for the Motion Conditions and the Stop Condition at Each Age Level Age (weeks) Condition Motion % fixation Reaches/min Stop % fixation Reaches/min - Base level Reaches/min N
1
4
7
10
13
16
19
17 2.2
50
69 .9
.1
51 2.0
56 2.6
68 4.0
75 7.6
27 2.1
69 2.2
66 1.1
78 3.1
71 3.9
82 6.2
92 10.1
1.9 13
2.1 19
2.2 18
1.6 18
1.0 20
1.7 17
2.3 12
Note The base level of reaching at the object-absent condition is also shown for each age level.
tivity of the infant, reaching activity should be related to this base level. An estimation of the base-level activity is given by the frequency of reaching in the object-absent period. However, just to compare the overall frequency of reaching in an object condition with the base frequency in the object-absent condition would not do justice to the effect of the object on reaching frequency. This follows from the fact that the infant does not look at the object all the time when it is present. To get a better estimation of how the object affects reaching, we need to know the relative time spent looking at the object in each condition. The present study did not measure looking time directly, but we do know for each reach whether or not the infant looked at the object. If we assume that the base-level activity during an unattended period of an object condition is the same as the base-level activity during the object-absent condition, we can calculate, for each condition, the amount of time spent looking at the object. Dividing the number of nonfixated reaches in a condition with the base-level frequency will give us the amount of time not looking at the object in that condition. The remaining time would then be the estimated looking time. In Table 3 the calculated proportion time spent looking at the object and the frequency of reaching during this time is shown for the motion conditions and the Stop condition at each age level. At the youngest age level the figures in Table 3 are only based on data from the 13 infants who were presented with the object-absent
condition. Table 3 suggests that the infant spends more time looking at the object and reaches more frequently for it in the Stop condition than in the motion conditions. Second, looking seems to increase monotonically with age. Third and most important, Table 3 suggests that reaching activity while the infant watches the object decreases from being slightly above base level in the neonate to being well below base level activity in the 7-week old. At that age, looking at the object almost inhibits reaching completely. After 7 weeks of age, reaching while watching the object rises monotonically. Already at 10 weeks the frequency is above base level. Type of Movements The results of the analysis of hand posture and hand movement during reaching are shown in Figures 3,4, and 5. The three figures show the proportion of reaches where the hand was fisted (f), open before (o), and opening during (od) the forward extension of the arm for the fixated (F) as well as the nonfixated reaches (NF) at each age level. Figure 3 shows that the proportion of fisted reaches increases dramatically at 7 weeks of age. After this age, the proportion of fisted reaches decreases again. An ANOVA was performed on the data of Figure 3. The individual scores were thereby treated as replications. As 8 of the subjects were missing in the last session, the analysis was only based on the first six sessions. Missing data were either inter-
DEVELOPMENTAL CHANGES OF PREREACHING MOVEMENTS
polated (7 cases) or extrapolated (3 cases). Only the main effect of age was found to be significant, F(5, 218) = 6.94, p = 0.01. Figure 4 shows that the reaches where the hand is opened before the movement starts decreases in frequency up to 7 weeks of age and then increases. There are no systematic differences between fixated and nonfixated reaches in this respect. An ANOVA was performed on the data of Figure 4. As in the previous case, the individual scores were treated as replications, and the last session was deleted from the analysis. The main effect of age was found to be significant, F\5, 218) = 10.5, p = 0.01. Finally, Figure 5 shows that the number of reaches where the hand is opened during the forward extension decreases up to 7 weeks of age. After that there is an increase for the fix-
385
ated reaches but not for the nonfixated ones. It should also be noted that, at each age, there are more fixated open-during-reaches than nonfixated ones. An ANOVA was performed on the data of Figure 5. The procedure was as in the two previous cases. The main effect of fixation was found to be significant, F{1, 218) = 12.61, p = 0.01, as well as the interaction between fixation and age, ^ 5 , 218) = 3.77, p = 0.01. The main effect of age was significant, F\5, 218) = 2.95,/> = 0.05. A note on laterality At all age levels there were more right-hand than left-hand movements. The percentage of right-hand movements was 72, 63, 82, 67, 63, and 55 at 1, 4, 7,10,13,16, and 19 weeks of age, respectively. Seven infants used the right hand more than the left one throughout the study, but none used the left hand more throughout the study.
A % fisted
50.
1
10
13
16
AGE(weeks) Figure 3 Percentage of fixated movements (F) and nonfixated movements (NF) at which the hand is fisted for the different age levels of the study.
386
CLAES VON HOFSTEN
the forward extension of the arm. Such an explanation fits with McGuire and Turkewitz's (1979) application of Schneirla's approachwithdrawal theory to infant behavior. This theory states that if the intensity of the stimulus in some dimensions exceeds some critical value the infant will tend to withdraw from it instead of approach it. However, the approach-withdrawal theory would also predict that the infant would look away from a too-intense stimulus, but this does not seem to happen in the present situation. Rather, the object seems to be intensely fixated (Trevarthen, 1974). Another possibility is that, in the excited 2month-old infant, not only the agonist but also the antagonist muscles will be activated and consequently block the movement. Gatev (1972) found that arm movements in the 1to 2-month-old infant usually begin without antagonist inhibition. The two explanations presented above may account for some aspects of the changes in
At the most, the left hand dominated in half the sessions. This happened in three cases. Discussion The present investigation gives evidence of substantial changes in the pattern of prereaching around 2 months of age. These changes affect the amount as well as the form of reaching. The amount of reaching goes down and the hand starts to get fisted instead of opened during the forward extension of the arm as at earlier age levels. The decline in reaching does not seem to have anything to do with loss of interest in the task. According to the calculated looking times, the fixation of the object seems to increase during the same age period. The data, rather, indicate that the actual looking at the object inhibits, in some way, the execution of forward-extended arm movements. One way to explain this effect is in terms of a change in state of the infant. The excitement caused by looking at the target might inhibit ' > %
open before
50
1
4
7
10
13
16
19
AGE ( w e e k s )
Figure 4 Percentage of fixated movements (F) and nonfixated movements (NF) at which the hand is opened before the forward extension starts for the different age levels of the study.
DEVELOPMENTAL CHANGES OF PREREACHING MOVEMENTS
prereaching pattern occurring at 2 months of age, but neither of them can account for the fact that the form of movement is changed whether the object is fixated or not. The extension synergy, the forward extension of the arm, and the opening of the hand are broken up. At 2 months of age the hand was fully flexed during the extension of the arm. It is indeed difficult to explain this development in other terms than organizational changes in the prereaching pattern itself. At one level the things happening at 2 months of age indicate important changes in the organization of the nervous system. It has been well established (Kuypers, 1962, 1964) that at least two separable systems are responsible for the control of the upper limb; one "proximal" motor system organized mainly on brainstem level and responsible for the gross movements of the arm and hand and one "distal" motor system organized cortically and responsible for the fine coordination of the hand. The dissociation between arm and
opening
387
hand movements at 2 months of age and the inhibition of directed reaching may indicate that the cortically "distal" motor system has started functioning but is not yet synchronized with the proximal motor system. Consequently, coordination is adversely affected. It may also be as Trevarthen et al. (1981) have suggested that both systems are functioning earlier but that, during the first month, the proximal control system grows more than the distal and that this is the cause of incoordination. The discussion of "distal" and "proximal" motor systems and their role in development of reaching has certain affinities with McGraw's (1943) discussion of the neuro-muscular maturation of the human infant. In both cases there is a notion of increased cortical involvement with age. However, in McGraw's model there are also some explicit statements describing how this comes about. The rather well-organized but reflexive behavior of the newborn is said to disappear through the pro-
during
50
19 AGE(weeks)
Figure 5 Percentage of fixated movements (F) and nonnxated movements (NF) at which the hand is opening during the forward extension for the different age levels of the study.
388
CLAES VON HOFSTEN
cess of inhibition before reappearing again in a more mature and cortically organized form. The inhibition should cause the behavior not only to decrease in amount but also to become disorganized in form. This is in general agreement with the present results. However, some of McGraw's studies have also more direct relevance for the discussion of the development of arm and hand control. For instance, she studied the ability of an infant to hold itself suspended by the force of a hand grip. In the newborn this behavior is a flexion synergy by which the infant by 4 weeks of age can hold itself suspended on the average almost half a minute. However, at 7 to 10 weeks of age the suspension time has decreased to 6 s on the average indicating a breaking up of the flexion synergy that in time parallels the breaking up of extension synergy observed in the present study. At a more functional level of description, Mounoud's (Mounoud, 1976, 1981; Mounoud & Vinter, 1981) model of senso-motor development fits the obtained results quite well. The breakdown of the extension synergy of forward extended arm movements at 2 months of age can be seen as the dissociation of a prestructured movement pattern. It is paralleled by the breakdown of coordination between auditory perception and head movements in the orientation toward a sound source (Field, Muir, Pilon, Sinclair, & Dodwell, 1980). When, after 2 months of age the hand starts to open again during the forward extension of the arm, it only occurred when the infant fixated the object. The opening of the hand is not, any longer, just a part of an extension synergy and an orientation reaction but a meaningful adaptive behavior in itself. The hand now starts to open as a preparation for manipulating of the object. References Bower, T. G. R. Development in infancy (1974) San Francisco. Freeman Cunningham, C. C (1979) Aspects of early development in Down Syndrome infants Unpublished doctoral dissertation, University of Manchester, Manchester, England. DiFranco, D., Muir, D. W., & Dodwell, P. C. (1978) Reaching in very young infants. Perception, 7, 385-392. Field, J., Muir, D., Pilon, R., Sinclair, M., & Dodwell, P.
(1980). Infants' orientation to lateral sound from birth to three months Child Development, 51, 295-298. Gatev, V. (1972) Role of inhibition in the development of motor co-ordination in early childhood. Developmental Medicine and Child Neurology, 14, 336-341. Hofsten, C. von. (1982). Eye-hand coordination in the newborn. Developmental Psychology, 18, 450-461. Humphrey, T. (1969). Postnatal repetition of human prenatal activity sequences with some suggestions of their neuroanatomical basis. In R J Robinson (Ed.) Brain and early behaviour. Development in thefetus and infant (pp 43-84). London' Academic Press. Kuypers, H G. J M. (1962). Corticospinal connections Postnatal development in the Rhesus monkey Science. 138. 678-680 Kuypers, H G. J. M (1964) The descending pathways to the spinal cord, their anatomy and function. In J. C Eccles, & J C. Shade, (Eds.), Organization of the spinal cord Amsterdam: Elsevier. McDonnell. P (1979) Patterns of eye-hand coordination in the first year of life Canadian Journal ofPsychology, 33, 253-267. McGraw, M. B (1943). The neuro-muscular maturation of the human infant New York: Columbia University Press McGuire. I., & Turkewitz, G (1979). Approach-withdrawal theory and the study of infant development In M Bortner (Ed ), Cognitive growth and development Essays in memory of Herbert Birch (pp 57-84) New York' Brunner/Mazel Mounoud. P (1976). Revolutionary periods in early development. Archives de Psychologic XLIV, 171, 103— 114 Mounoud. P (1981, October). L'evolulion des conduttes de prehension comme illustration dun model du developpement Paper presented at the seventeeth Joumees d'Etudes de l'Association de Psychology Scientifique de Langue Francaise (APSLF) Grenoble, France Mounoud, P., & Vinter, A (1981). Tire-a-part: Representation and sensonmotor development. In G Butterworth (Ed ), Infancy and epistemology An evaluation of Piaget's theory (pp 200-235) Brighton, Susex: Harvester Press Trevarthen, C (1974). The psychobiology of speech development In E H Lenneberg (Ed ), Language and brain Developmental aspects Neurosciences Research Program Bulletin (pp 570-585), 12 Trevarthen, C (1979). Sketches of chair (Available from C Trevarthen. Department of Psychology, University of Edinburgh. 7 George Square, Edinburgh EH8 9YL, Scotland.) Trevarthen, C , Murray, L., & Hubley, P. (1981) Psychology of infants lnJ.A Davies&J Dobbing(Eds ), Scientific foundations ofpaediatrics (2nd ed.; pp 211-274). London William Hememann Medical Books. Twitchell, T. E. (1965). The automatic grasping responses of infants. Neuropsychologia, 3, 247-259 White, B. L., Castle. P , & Held, R. (1964). Observations on the development of visually directed reaching Child Development, 35, 349-364.
Received August 30, 1982 Revision received November 30, 1982 •
