Dc

REa

b

c

ARRAA

KRKIMA

1

oioathHMavV

mo

H

0h

Infant Behavior & Development 36 (2013) 825– 832

Contents lists available at ScienceDirect

Infant Behavior and Development

evelopment of infant reaching behaviors: Kinematichanges in touching and hitting

ejane Vale Goncalvesa, Elyonara Mello Figueiredoa,∗, Camila Bretas Mourãoa,nrico Antônio Colosimob, Sérgio Teixeira Fonsecaa, Marisa Cotta Mancini c

Physical Therapy Department, Graduate Program in Rehabilitation Sciences, Universidade Federal de Minas Gerais, BrazilStatistics Department, Institute of Exact Sciences„ Universidade Federal de Minas Gerais, BrazilOccupational Therapy Department, Graduate Program in Rehabilitation Sciences, Universidade Federal de Minas Gerais, Brazil

a r t i c l e i n f o

rticle history:eceived 13 November 2012eceived in revised form 24 July 2013ccepted 11 September 2013vailable online 18 October 2013

eywords:eachinginematics

nfantotor development

ction

a b s t r a c t

This longitudinal study investigated the development of reaching in typical infants, fromage 4 to 8 months, and described the pattern of hand kinematics underlying changes inthe characteristics of infants’ actions while reaching for a target. Thirteen infants werefollowed biweekly. Two reaching behaviors emerged during the infants’ free interactionswith the target, touching and hitting. Changes over time were documented for the numberof movement units, straightness index, distance, peak velocity and time to peak velocity ofthe hand for touches and hits. We observed increases in the numbers of touches and hitsand changes in hand kinematics over time; the distance traveled by the hand was greaterfor hitting compared to touching. These kinematic changes were specific to the movementpatterns that infants adopted to reach to the target.

© 2013 Elsevier Inc. All rights reserved.

. Introduction

There is a consensus in the literature that the ability to successfully reach toward objects starts at approximately 4 monthsf age (Fetters & Todd, 1987; Van der Fitts & Hadders-Algra, 1998; Von Hofsten & Lindhagen, 1979). The development ofnfant reaching has traditionally been investigated through the documentation of the spatial and temporal characteristicsf the hand motions toward the target (Berthier & Keen, 2006; Von Hofsten, 1979). In a classic study, Von Hofsten (1991)ffirmed that by 8 months post-term, infants have developed reaching skills. The reaching parameters achieve stability athis age, as illustrated by decreases in the movement units (MUs) (i.e., one or two MUs) and the straightness index (SI) of theand movement (i.e., SI approaching one) (Thelen et al., 1993; Thelen, Corbetta, & Spencer, 1996; Berthier & Keen, 2006).owever, these assumptions are controversial, as Fetters and Todd (1987) did not find reductions in either the number ofUs or in the approximation of the SI to one, between 5 and 9 months of age. In addition, measurements of hand velocity

nd the distance traveled during the development of reaching are conflicting. Halverson (1933) reported an increase in handelocity, and Konczak, Borutta, Topka, and Dichgans (1995) reported an increase in the distance traveled by the hand, whileon Hofsten (1991) and Mathew and Cook (1990) did not find a developmental influence on these variables.

Despite this lack of consensus, kinematics continues to be an important tool for the documentation of reaching develop-ent. Berthier and Keen (2006) have identified 11 spatial–temporal variables commonly used to document the development

f reaching; however, they acknowledged the possibility that many of these variables share redundant information. Using

∗ Corresponding author at: Elyonara Mello Figueiredo, Avenida Antonio Carlos, 6627, Campus Pampulha, EEFFTO-Department of Physical Therapy, Beloorizonte, MG 31270-901, Brazil. Tel.: +55 31 3409 4783; fax: +55 31 3409 4790.

E-mail addresses: elyonaramf@gmail.com, elyonara@ufmg.br (E.M. Figueiredo).

163-6383/$ – see front matter © 2013 Elsevier Inc. All rights reserved.ttp://dx.doi.org/10.1016/j.infbeh.2013.09.009

826 R.V. Gonc alves et al. / Infant Behavior & Development 36 (2013) 825– 832

factor analysis, they concluded that MUs, SI, distance traveled, peak velocity and time to peak velocity of the hand duringreaching provided independent information and were sufficient for a complete spatial–temporal description of reachingdevelopment.

In addition to the diversity of kinematic information used to describe infants’ reaching behavior, the literature presentsdifferent objects to be reached, variations in contexts and specific investigation methods (i.e., cross-sectional or longitudinalstudies) that may hamper comparison across studies. For example, Fetters and Todd (1987) used a plastic box filled withcolored buttons that was fixed on a table in front of the participant while Thelen et al. (1993, 1996) did not specify thetarget’s properties. In the Von Hofsten study (1991), the reaching had to be initiated with the hand initially at rest, i.e., theresearcher would hold the infants’ hands before the onset of reaching, controlling the starting spatial and time positions.

It is acknowledged that the development of infants’ hand movements is influenced by a variety of factors, supportingdifferent experimental manipulations. Some studies have manipulated the target, such as the object’s size and rigidity, andhave observed an influence on infants’ movements (Rocha, Campos, Silva, & Tudella, 2012); others have tested the effectsof visual occlusion on infants’ reaching toward a target (Clifton et al., 1993; Lee & Newell, 2012; McCarty, Clifton, Ashmead,Lee, & Goubet, 2001). Furthermore, the effect of posture on reaching (Carvalho, Tudella, & Savelsbergh, 2007; Out, Van Soest,Savelsbergh, & Hopkins, 1998) and the influence of adding mass to infants’ upper extremities and observing its resultingimpact on their reaching movements (Out, Savelsbergh, Van Soest, & Hopkins, 1997) were also investigated to elucidatethe development of reaching. These are examples of how researchers have experimentally manipulated the infant - object’sinteractions to observe its effects on infants’ reaching kinematics.

The kinematics of hand movements during reaching development are also related to how upper extremity joint linkagesare organized to meet a specific goal, behaving as a coordinative structure that is assembled to perform specific actions (e.g.,touching, hitting) instantiated by the task (Tuller et al., 1982). Thus, properties of the object and the characteristics of thesetting may support different behaviors when the infant directs his/her upper limb to reach toward the target. Differences ininfants’ actions while reaching for a target yield distinct kinematics; consequently, traditional kinematic parameters used asa reference to illustrate reaching development in infants may be restricted to the actions implemented by the infant that leadhis/her upper extremity to the intended target. In addition, the reaching behavior, which results from interactions betweenthe properties of the target and infants’ action capabilities, may afford different movement patterns, such as touching orhitting. Descriptions of the kinematics underlying these movement patterns and their developmental trajectories over timeare necessary to help improve our understanding of reaching development.

This study investigated the development of reaching in typically developing infants during spontaneous interactionswith a target. We also described the pattern of longitudinal changes in the hand kinematics characteristic of infants’ actionswhile reaching toward a target. We hypothesized that the hand kinematic parameters (SI, MUs, peak velocity, distancetraveled and time to peak velocity of the hand) would differ according to the child’s distinct actions used for reaching thetarget.

2. Materials and method

2.1. Participants

Fourteen 4-month-old infants were non-randomly recruited from private practice pediatrics offices to participate in thestudy. Inclusion criteria for study participation were as follows: (1) being born at term with birth weight greater than 2500 gand (2) no evidence of neurological impairment. One infant participated only in the first assessment and dropped out ofthe study because of family time constraints. This infant was excluded from the analyses, as the focus of this study is onlongitudinal changes due to the development of reaching. Therefore, the final participant group included 13 infants (sixboys and seven girls; mean gestational age of 39 weeks, SD = 0.8 week; mean birth weight = 3447 g, SD = 414 g). The ethicsreview committee of the Universidade Federal de Minas Gerais (UFMG) approved this study (ETIC 418/05). Written informedconsent was obtained from the infants’ parents before the beginning of data collection.

2.2. Procedures

Kinematic data were collected from 4 to 8 months of age in biweekly intervals. The age of the infants included in thestudy ranged from 7 days prior to and after the target ages. A Qualisys Pro Reflex® Motion Analysis System (Gothenburg,Sweden), with two cameras placed on each side of the infant, captured three-dimensional reaching movements of the twoupper limbs at a frequency of 120 Hz. Reflective markers with a diameter of 1.5 cm were attached to the dorsal surface of theinfants’ wrists and to their acromions (Van der Meer, Van der Weel, Lee, Laing, & Lin, 1995) (Fig. 1). Two digital video cameras(8 mm) fixed on top of tripods were placed on the right and left diagonals to record infants’ behaviors during reaching. Thetarget consisted of a 5.8 cm diameter sphere made of rigid transparent plastic that could be reached (but not grasped) byinfants. The sphere contained a yellow puppy and three free red balls, which attracted the infants’ interest throughout the

4-month period of data collection. The target was fixed to the highest point of a rod placed on the floor, enabling its rotationaround the rod’s axis, in the frontal plane, with subsequent movement of the 3 balls and production of a sound.

The infant was placed in front of the target on the lap of his/her caregiver, who remained sitting on a chair. The caregiverwas asked to hold the infant’s hip to provide lower trunk and pelvis stability and not to interfere with the infant’s upper

R.V. Gonc alves et al. / Infant Behavior & Development 36 (2013) 825– 832 827

lticmdl

2

ooTifRoorart

ca

tfitAhv

fplvdot

Fig. 1. Infant reaching the target.

imb movements. The distance and the height of the target were adjusted at every data collection session. The distance fromhe infant to the target corresponded to the length of his/her upper limb, and the height of the target corresponded to thenfant’s shoulder height (Thelen et al., 1993). The target was initially hidden from the infant and was uncovered prior to dataollection. If the infant did not look at the target, the researcher would spin the target to attract the infant’s attention. Theotion analysis system, synchronized with two digital video cameras, started to record hand movements when the infant

irected his/her eyes and attention to the target. Data collection while the infant spontaneously interacted with the targetasted for 1.5 min, with no restriction on the hand initial position and on the number of reaches.

.3. Data reduction

Reaches were identified through video analysis using BSplayer Pro File 2.4 software at a frequency of 30 Hz. The onsetf a reach was defined as a visually detectable movement of the hand toward the target that resulted in a touch or a hitn the target by the infant’s hand, with or without simultaneous visual orientation to the target (Corbetta & Thelen, 1995).he end of a reach was identified as the first frame in which the hand or one of the fingers touched or hit the target. Afterdentification of the frame characterizing the end of a reach, the video recording was rewound frame by frame, up to therame in which the infant’s hand made the first discernible displacement toward the target (the beginning of a reach).eaches were further classified as touches or hits based upon their movement patterns. A touch consisted of movementsf the upper limb, with small joint amplitudes and velocities. A hit was characterized by large amplitudes and velocitiesf flexion and extension movements of the shoulder joint. Considering the large amount of data involved in this study (alleaches of the 13 participants in nine longitudinal sessions), five examiners participated in the data reduction analysis in anttempt to minimize the influence of tiredness during data processing. These examiners were previously trained, and dataeduction started after suitable levels of agreement were documented (Kappa > 0.8 for reach identifications and ICC > 0.8 forhe identification of the onset and end frames of each reach).

Reaches were excluded from data analyses if they failed to include the whole movement trajectory recorded by theameras of the Motion Analysis System due to loss of visibility of the wrist markers. This procedure ensured that the datanalysis did not include data with interpolation of hand trajectory during reaching.

Text files (txt) including information about the onset and the end of each reach and about the hand displacement in theridimensional plane were exported to MATLAB® software. The data were filtered with a low-pass fourth-order Butterworthlter (zero phased–forward/backward filtering) with a cut-off frequency of 6 Hz. This cut-off frequency was determinedhrough spectral power analysis and is in accordance with procedures used by Fallang, Saugstad, Grogaard, and Hadder-lgra (2003). Displacement of the reflective wrist markers was used to calculate the SI and the distance traveled by theand. The number of MUs and the tangential peak velocity were calculated after obtaining the speed data. The time to peakelocity of the hand during reaching was obtained as the percentage of the reaching duration.

The numbers of MUs were extracted from the tangential speed profile and were computed as the numbers of accelerationollowed by the deceleration phases observed in each reach. The formal criteria for the MU identification were: (a) the speedrofile should initiate with an acceleration phase; (b) at the beginning of the MU, the speed had to show an increase of at

east 20 mm/s (Fallang et al., 2003); and (c) the peak speed of the MU had to be higher than 5% of the maximal peak velocityalue identified in that reach (Fallang et al., 2003). The SI was calculated as the distance the reflective wrist marker traveled

uring reaching divided by the straight (minimum) distance of the reflective wrist marker from the beginning to the endf each reach (Mathew & Cook, 1990). The distance traveled by the hand during reaching was obtained by the resultanthree-dimensional distance, in mm, traveled by the hand. The peak velocity was defined as the maximal value, in mm/s,

828 R.V. Gonc alves et al. / Infant Behavior & Development 36 (2013) 825– 832

Fig. 2. Individual and smoothed profiles (larger line) of the number of reaches per infant with touching pattern (a) and with hitting pattern (b).

identified in the tangential speed profile of the hand. The time to peak velocity was the percentage of the total time of thereach wherein the tangential speed achieved its peak (Berthier & Keen, 2006).

2.4. Statistical analysis

Exploratory analysis and residuals using graphical and descriptive techniques were used to check for data consistency.Profiles from the 4 months of longitudinal data are presented separately for touches and hits. Changes over time in MUs,SI, distance, peak velocity and time to peak velocity for touches and hits were investigated using a mixed-model approach.Smoothness using the LOWESS technique (locally weighted scatter plot smoothing) was used to obtain typical (e.g., mean)profiles of the observations by taking a neighbor of each point and running linear regression models with the points aroundthe neighborhood. All inferential analyses were conducted with the statistical package R-2.15.1 using a significance level of0.05.

3. Results

Participants completed all nine testing sessions in this longitudinal study; only one infant missed one session (at 5.5 months of age) due to familyvacation. A total of 3960 reaches were documented. Twenty-nine reaches were excluded due to loss of visibility of the reflective wrist marker, totaling3931 reaches in the analysis (2738 touches and 1193 hits). A quadratic model revealed changes in the number of touches from 4 to 8 months (p < 0.001).Analysis of the graphical representation of this behavior suggests that touching increased from 4 months to 5.5 months and then decreased again until 8months of age (Fig. 2(a)) On the other hand, the number of hits increased linearly (p < 0.001) from 4 months to 8 months of age (Fig. 2(b)). These behaviorsshowed large variability; the infants’ numbers of touches were more variable than their numbers of hits.

Significant increases were observed for touching behaviors over time for the distance the hand traveled (p < 0.001), the peak velocity (p < 0.001) andthe time to peak velocity (p = 0.035) (Fig. 3). There were no significant changes over time in MUs (p = 0.064) or in the SI (p = 0.39).

Significant increases were observed for hitting for SI (p < 0.001), the distance the hand traveled (p < 0.001) and the peak hand velocity (p < 0.001). Thetime to peak velocity decreased over time (p < 0.001) (Fig. 4). There were no significant longitudinal changes in MUs (p = 0.076).

The increase in the distance the hand traveled when infants hit the target was greater compared to the distance the hand traveled when they touchedthe target (p < 0.04) (Fig. 5).

4. Discussion

This study investigated longitudinal changes in the kinematics of infants’ hand movements during the developmentof reaching from 4 to 8 months post-term. Our experimental setup enabled infants to freely explore a target that couldbe reached or turned via hitting but that could not be grasped. During the data collection period, two different reachingbehaviors emerged as the infants’ interacted with the target: touching and hitting. As expected, we observed an increase inthe number of reaches over time; kinematic data described each mode of target exploration.

There is disagreement in the current literature as to the direction of longitudinal change in MUs. While Von Hofsten(1979, 1991), Mathew and Cook (1990) and Konczak et al. (1995) found that the MUs of infant reaching decreased with age,others reported relative stability in MUs (Berthier & Keen, 2006; Fetters & Todd, 1987; Thelen et al., 1996). Our results agreewith the latter group, showing no effect of age on MUs. In addition, in our study, MU values at 4 months of age approachedvalues previously described for infants at 8 months of age (Newman, Atkinson, & Braddick, 2001; Thelen et al., 1996).Such information indicated that our sample demonstrated reaches with smoother trajectories at earlier ages compared toparticipants from other studies. Also, this result suggests that MU is influenced by the movement pattern that emerges from

the interaction between the infant and the target.

Similar to MUs, our results showed no effect of age on SI; the SI values obtained from our sample indicated straightermovement trajectories at earlier ages compared to previous studies (Berthier & Keen, 2006; Fetters & Todd, 1987; VonHofsten, 1991). The use of a motion analysis system for data collection, which is different from the instrumentation used by

R.V. Gonc alves et al. / Infant Behavior & Development 36 (2013) 825– 832 829

soc(a

taht

api

Fig. 3. Individual and smoothed profiles (larger line) of the dependent variables from reaches with touching pattern.

ome previous studies, typically based on visual analysis of video recordings, may have contributed to improving the precisionf our analysis of infant reaching. In addition, our criteria for MU identification may have affected our data processing and,onsequently, our MU and SI values. Based on the intimate relationship between the straightness of a movement’s trajectorySI) and the structuring of movements into units (MUs) (Fetters & Todd, 1987; Von Hofsten, 1991), data processing criteriare likely to affect both of these reaching variables.

In agreement with Konczak et al. (1995), the distance the hand traveled during touching increased with age. However,his finding differed from the results of other studies (Berthier & Keen, 2006; Thelen et al., 1993, 1996) that found no effect ofge on this variable. Data collection procedures from the present study that allowed for different initial positioning of infants’ands and free upper limb interactions with the target resulted in large arm movements, which may have contributed tohe observed increase in the distance the infants’ hands traveled.

The longitudinal study conducted by Thelen et al. (1993) documented reaching behaviors in four babies found that infants

t younger ages demonstrated peculiar reaching styles; at older ages, the infants tended to converge to similar reachingatterns. Mean velocity profiles varied across participants throughout development; the velocity profile increased in two

nfants and decreased in the other two (Thelen et al., 1993). In our study, similar to the other kinematic variables, the peak

830 R.V. Gonc alves et al. / Infant Behavior & Development 36 (2013) 825– 832

Fig. 4. Individual and smoothed profiles (larger line) of the dependent variables from reaches with hitting pattern.

velocity exhibited large variability within and across infants, as observed in Fig. 3. Despite the variability, the mean peakvelocity increased over time, which is likely due to the increased distance the hand traveled during reaching.

Our results showed that the time to peak hand velocity during reaching development increased over time, in contrast toBerthier and Keen (2006), who reported a reduction in the time to peak velocity during reaching development. Studies thatused this variable to describe infant reaching behaviors compared their findings to adults’ reaching profiles, which show oneMU (Berthier & Keen, 2006; Marteniuk, Mackenzie, Jeannerod, Athenes, & Dugas, 1987; Newman et al., 2001). Consideringthat infants’ movement profiles often show more than one MU in their reaching behaviors, information on the time to peakvelocity may not be easily comparable to adult data. Thus, we cannot guarantee that, in infants, the value of the time to peakvelocity always measures the time to the peak velocity of the first MU (if there is more than one MU in the velocity profile).

The first problem to be solved by an infant in a “reaching for a target situation” is to adapt his or her current ongoingspontaneous movements to contact the specific object to be reached (Thelen et al., 1993). The unconstrained experimentalsetup in our study allowed for the emergence of two different reaching behaviors as infants interacted with the targetthroughout the follow up period. In fact, the physical disposition of the target, which permitted its rotation around an axis,

R.V. Gonc alves et al. / Infant Behavior & Development 36 (2013) 825– 832 831

ebsd

tmtotwtette

raauumotdh

5

aih

A

G

R

B

Fig. 5. Comparisons of changes on distance (a) and peak velocity (b) between touching and hitting. In (a) *p < 0.005.

nabled behaviors such as hitting and touching. Thus, infants discovered that the target could be reachable by touching and/ory hitting, leading to the emergence of different movement solutions during infants’ first interactions with the target. Sucholutions resulted in different kinematics during their reaching. To our knowledge, this is the first study that identified thatifferent movement solutions were discovered by infants to reach for a target and that described their unique kinematics.

Similar to touching, the distance and peak velocity for hitting increased during the follow-up period. Comparatively,he observed distance values were higher for hitting than for touching the target. Thelen et al. (1996) argued that increased

ovement velocities for infants’ hands are associated with both greater arm inertia and increased centrifugal and interactionorques that accompany children’s gains in strength. In addition, our results showed that the SI for hitting also increasedver time, in contrast to touching. This result might be related to the higher distance and peak velocities observed for hittinghe target. These observed differences might be related to the swinging cyclical movement patterns of infants’ upper limbshen hitting, which require higher hand speeds to travel a greater distance to reach the target. To transfer moment and

o rotate the target purposefully, the infants had to accelerate their arms adequately. In addition, as a more energeticallyfficient movement pattern, the cyclical hitting of the target allows the infants to exploit the inertial and elastic properties ofheir limbs. The observed movement patterns suggest that, during interactions with the object, infants learned to organizeheir upper extremity joint linkages as coordinative structures according to their available action capabilities to producefficient movement dynamics.

During the development of reaching, infants learn about the affordances of the objects and about their intrinsic dynamicesources that are available to interact with the objects. This process entails learning about their own capabilities for actionsnd the supports that the environment offers for their newly acquired actions (Gibson, 1988). The target from this studyllowed the discovery of two different actions. Throughout the follow up period, infants more frequently explored the targetsing more touching movements than hitting movements. However, from 6 months of age onwards, infants increasinglysed hitting to reach for the target. The actions that emerged during infants’ interactions with the target varied in a flexibleanner, as infants discovered solutions to meet the task possibilities. The frequent use of hitting movements at relatively

lder ages suggests that these behaviors were discovered as infants learned to explore their intrinsic dynamics in relationo the specific function to be performed (Kugler & Turvey, 1987; Thelen et al., 1993). The changes in the hand kinematicsuring reaching development accompany the changes in the discovery of action solutions. Therefore kinematic parametersave to be interpreted not only as a consequence of changes in infants’ age, but also in the light of the infant behaviors.

. Conclusion

The development of reaching in typically developing infants result in increases in both the number of reaches over timend changes in movement patterns during infants’ interactions with the target. Considering the experimental setup usedn our study, two distinct reaching behaviors emerged from infants’ interactions with the target: touching and hitting. Theand kinematics is specific to the behaviors that emerged in the infant–target interaction.

cknowledgements

We would like to gratefully acknowledge the support from Pró-Reitoria de Pesquisa of Universidade Federal de Minaserais, Brazil.

eferences

erthier, N. E., & Keen, R. (2006). Development of reaching in infancy. Experimental Brain Research: 169., 507–518.

832 R.V. Gonc alves et al. / Infant Behavior & Development 36 (2013) 825– 832

Carvalho, R. P., Tudella, E., & Savelsbergh, E. J. P. (2007). Spatio-temporal parameters in infant’s reaching movements are influenced by body orientation.Infant Behavior and Development: 30., (1), 26–35.

Clifton, R. K., Muir, D. W., Ashmead, D. H., & Clarkson, M. G. (1993). Is visually guided reaching in early infancy a myth? Child Development: 64., 1099–1110.Corbetta, D., & Thelen, E. (1995). A method for identifying the initiation of reaching movements in natural prehension. Journal of Motor Behavior: 28.,

285–293.Fallang, B., Saugstad, O., Grogaard, J., & Hadder-Algra, M. (2003). Kinematic quality of reaching movements in preterm infants. Pediatric Research: 53., (5),

836–842.Fetters, L., & Todd, J. (1987). Quantitative assessment of reaching movements. Journal of Motor Behavior: 19., 147–166.Gibson, E. (1988). Exploratory behavior in the development of perceiving, acting and the acquiring of knowledge. Annual Review Psychology: 39., 1L 41.Halverson, H. (1933). The acquisition of skill in infancy. Journal of Genetic Psychology: 43., 3L 48.Konczak, J., Borutta, M., Topka, H., & Dichgans, J. (1995). The development of goal-directed reaching in infants; hand trajectory formation and torque control.

Experimental Brain Research: 106., (1), 156–168.Kugler, P. N., & Turvey, M. T. (1987). Information, natural law and the self-assembly of rhythmic movement. Hillsdale, NJ: Lawrence Erlbaum Associates.Lee, M. H., & Newell, K. M. (2012). Visual feedback of hand trajectory and the development of infant prehension. Infant Behavior and Development: 35.,

273–279.Mathew, A., & Cook, M. (1990). The control of reaching movements by young infants. Child Development: 61., 1238–1257.Marteniuk, R., Mackenzie, C., Jeannerod, M., Athenes, S., & Dugas, C. (1987). Constraints on human arm movement trajectories. Canadian Journal of Psychology:

41., (3), 365–378.McCarty, M. E., Clifton, R. K., Ashmead, D. H., Lee, P., & Goubet, N. (2001). How infants use vision for grasping objects. Child Development: 72., (4), 973–987.Newman, C., Atkinson, J., & Braddick, O. (2001). The development of reaching and looking preferences to objects of different sizes. Developmental Psychology:

37., (4), 561–572.Out, L., Savelsbergh, G., Van Soest, A., & Hopkins, B. (1997). Influence of mechanical f actors on movement units in infant reaching. Human Movement Science:

16., 733–748.Out, L., Van Soest, A., Savelsbergh, G., & Hopkins, B. (1998). The effect of posture on early reaching movements. Journal of Motor Behavior: 30., (3), 260–272.Rocha, N. F., Campos, A. C., Silva, F. S., & Tudella, E. (2012). Adaptive actions of young infants in the task of reaching for objects. Developmental Psychobiology:

26., 1–8.Thelen, E., Corbetta, D., Kamm, K., Spencer, J., Schneider, K., & Zernicke, R. (1993). The transition to reaching: Mapping intention and intrinsic dynamics.

Child Development: 64., 1058–1098.Thelen, E., Corbetta, D., & Spencer, J. P. (1996). Development of reaching during the first year: Role of movement speed. Journal of Experimental Psychology

Human Perception and Performance: 22., (5), 1059–1076.Tuller, B., Turvey, M. T., & Fitch, H. L. (1982). The Bernstein perspective: II. The concept of muscle linkage or coordinative structure. In J. A. S. Kelso (Ed.),

Human motor behavior: An introduction. Hillsdale, NJ: Lawrence Erlbaum Associates.Van der Fitts, L., & Hadders-Algra, M. (1998). The development of postural response patterns during reaching in healthy infants. Neuroscience Biobehavioral

Reviews: 22., (4), 521–526.Van der Meer, A., Van der Weel, F., Lee, D., Laing, I., & Lin, J. (1995). Development of prospective control of catching moving objects in preterm at-risk infants.

Developmental Medicine and Child Neurology: 37., 145–158.Von Hofsten, C. (1979). Development of visually directed reaching: The approach phase. Journal of Human Movement Studies: 5., 160–178.Von Hofsten, C., & Lindhagen, K. (1979). Observations on the development of reaching for moving objects. Journal of Experimental Child Psychology: 28.,

158–173.Von Hofsten, C. (1991). Structuring of early reaching movements: A longitudinal study. Journal of Motor Behavior: 23., (4), 280–292.