GMFM in Infancy: Age-Specific Limitations and Adaptations

Hielkema, Tjitske MD; Hamer, Elisa G. BSc; Ebbers-Dekkers, Ilse MSc; Dirks, Tineke PT; Maathuis, Carel G.B. MD, PhD; Reinders-Messelink, Heleen A. PhD; Geertzen, Jan H.B. MD, PhD; Hadders-Algra, Mijna MD, PhD

Pediatric Physical Therapy:
doi: 10.1097/PEP.0b013e318288d370
Research Article

Purpose: To evaluate longitudinal applicability of the Gross Motor Function Measure (GMFM) in infants younger than 2 years.

Methods: Twelve infants at very high risk for cerebral palsy were enrolled between 1 and 9 months corrected age. The children were assessed 4 times during 1 year with the GMFM-66, GMFM-88, and other neuromotor tests.

Results: Longitudinal use of the GMFM in infancy was hampered by age and function-specific limitations. The GMFM-66 differentiated less at lower-ability levels than at higher-ability levels. The GMFM-88 demonstrated flattening of the developmental curve when infants had developed more motor abilities. We formulated adaptations for the longitudinal use of GMFM in infancy.

Conclusions: To facilitate use of the GMFM in infancy, an adapted version may be an option. Further research is required to assess reliability and validity, and in particular, the sensitivity to change of the suggested adaptations.

In Brief

The authors examined the limitations of the GMFM at younger ages, proposed adaptations to increase sensitivity to change at younger ages, and recommend further study of the validity of these modifications.

Author Information

Departments of Paediatrics—Developmental Neurology (Drs Hielkema and Hadders-Algra and Mss Hamer, Ebbers-Dekkers, and Dirks) and Rehabilitation Medicine, Center for Rehabilitation (Drs Hielkema, Maathuis, Reinders-Messelink, and Geertzen), University Medical Center Groningen, University of Groningen, Groningen, the Netherlands; Rehabilitation Center “Revalidatie Friesland” (Dr Reinders-Messelink), Beetsterzwaag, the Netherlands.

Correspondence: Mijna Hadders-Algra, MD, PhD, Departments of Paediatrics—Developmental Neurology, University Medical Center Groningen, Hanzeplein 1, 9713 GZ Groningen, the Netherlands (

Grant Support: The study was part of the national LEARN2MOVE research program and was supported financially by ZonMW, Johanna Kinderfonds, Stichting Rotterdams Kinderrevalidatie Fonds, Adriaanstichting, Revalidatiefonds, Phelps Stichting, Revalidatie Nederland, and the Nederlandse Vereniging van Revalidatieartsen.

At the time of the study, Elisa G. Hamer was involved as an employee of the University Medical Center Groningen as a PhD student in combination with her medical study (master degree) at the University of Groningen.

The authors declare no conflicts of interest.

Article Outline
Back to Top | Article Outline


The Gross Motor Function Measure (GMFM) is a reliable and valid instrument to measure gross motor functioning in children with cerebral palsy (CP).14 It is widely used for both clinical and research purposes. The GMFM was developed in the 1980s, as a measurement tool to detect changes in gross motor development in children with CP. The originally published GMFM had 88 items, subdivided into 5 dimensions: A, lying and rolling; B, sitting; C, crawling and kneeling; D, standing; and E, walking, running, and jumping. A child developing typically will perform all items around the age of 5 years. The more recent version of the GMFM-88, the GMFM-66,1 adapted the original instrument to be an interval measure. It takes “not observed” behavior into account by estimating scores for items not shown on the basis of shown items. The resulting total score has limited the contribution of prior dimensions A and B and, to a lesser extent, of the prior dimension C. The GMFM-66 can be used to calculate only total scores and not dimension scores.1 It is conceivable that the limited contribution of dimensions A and B makes the GMFM-66 less suitable for young infants, as they will score only or primarily in the first dimensions. To improve use of the GMFM in busy clinical practices, subsets of items based on a child's ability are used in the GMFM-66 item sets and the GMFM basal and ceiling approach. Both versions have a good reliability and validity in children aged 2 to 6 years.5,6

The achievement level of children with CP depends on the severity of impairment. Severity of gross motor impairment is currently expressed in terms of the Gross Motor Function Classification System (GMFCS).7 For children with CP at GMFCS levels I and II, gross motor development as measured by the GMFM-66 reaches a plateau around 6 to 7 years.8 Children at GMFCS levels III, IV, and V will reach their plateau at an earlier age.9 This means that the upper age limit of the GMFM depends on the child's GMFCS level. Lower age limits for the application of the GMFM have not been indicated.1 However, it could be argued that a lower age limit exists, as CP is often not reliably diagnosed before the age of 18 months.10 Limited information is available on the use of the GMFM in infants younger than 18 months who are at risk for or suspected of developing CP.

The original validation sample of the GMFM of 111 children with CP, aged 5 months to 15.4 years, included assessments of 21 children aged between 5 months and 2 years. It was reported that younger children showed more change in GMFM scores with increasing age. The youngest children showed most changes in dimensions A, B, and C. Interrater reliability and agreement between the professionals' and the parents' prior judgments were lowest on these 3 dimensions.1,11 Three other groups of authors addressed characteristics of GMFM scores in infancy.1214 They did not agree on the applicability of the GMFM to monitor gross motor development at an early age. Jelsma et al12 described the development of GMFM scores over a period of 8 months in 12 infants who showed clear neurological dysfunction at the age of 4 months. They reported that GMFM scores do not increase in a linear fashion and may vary to a great extent.12 Difficulties in assessing GMFM in very young children may have contributed to the variation. Kolobe et al13 described changes in GMFM scores over a 6-month period in 42 infants with CP or motor delay, indicating that application of the GMFM for children younger than 2 years is hampered by age and developmental characteristics of infancy. Wei et al14 studied the use of the GMFM-88 in children aged 3 to 36 months. On the basis of Rasch analysis, they excluded some items of dimensions B, C, and E and constructed a version with 73 items. This GMFM-73 was no more reliable than the GMFM-66. They therefore concluded that the prior GMFM-66 is a reliable and valid instrument for children younger than 3 years.14

The aim of this study was to describe developmental changes seen in infants at high risk for CP by using the GMFM-88 and GMFM-66 over a 1-year period. We planned to describe problems encountered in (a) the application of items, (b) scoring of items, and (c) developmental changes over time and to compare the infants' developmental trajectories on the GMFM-88 and GMFM-66, that is, the changes over the 1-year period, with the developmental trajectories measured with 3 other scales of infant motor development (the Alberta Infant Motor Scale [AIMS],15 the Bayley Scales of Infant Development–II Psychomotor Development Index [BSID-II PDI, Dutch version],16 and the Infant Motor Profile [IMP]).17 Our aim was also to present suggestions for adaptation of the GMFM for use in infancy.

Back to Top | Article Outline



Twelve infants (7 boys and 5 girls) from the LEARN2MOVE 0-2 years study18 participated. The medical ethics committee of the University of Groningen granted approval for the study. The trial is registered with the Dutch Trail Registry as NTR1428. Infants were included after their parents gave informed consent. Inclusion criteria were a corrected age between 1 and 9 months and a very high risk for CP. The latter was based on the presence of 1 of the following criteria: (a) cystic periventricular leukomalacia, diagnosed on serial ultrasound assessments of the brain19 (n = 2), (b) unilateral or bilateral parenchymal lesion of the brain20 (n = 4), (c) term/near-term asphyxia resulting in Sarnat21 score 2 or 3 (n = 4) with brain lesions on magnetic resonance imaging, and/or (d) neurological dysfunction during infancy, suggesting the development of CP (n = 2). Exclusion criteria were an additional severe congenital disorder, such as serious congenital heart disorder, or caregivers with insufficient understanding of the Dutch language. The median gestational age was 39 weeks (range, 27–41 weeks); median birth weight was 3140 g (range, 720–5400 g).

Back to Top | Article Outline
Study Design

The GMFM-88 and GMFM-66 were used as part of an extensive assessment battery that occurred over a 2- to 3-day period.18 The GMFM was examined at baseline (T0) and after 3 months (T1), 6 months (T2), and 12 months (T3). Most examinations were carried out at the children's homes with one or both parents present. Assessments were performed by 2 trained GMFM examiners, with the assistance of an assessor with a master's degree in human movement sciences. All assessments were video recorded. Total GMFM-88 scores and dimensions scores were determined. In addition, GMFM-66 scores were determined with the Gross Motor Ability Estimator.

Other neuromotor tests used were the AIMS,15 the BSID-II PDI, Dutch version,16 and the IMP.17 The AIMS is a reliable and valid15 instrument to assess gross motor development in infants from birth through 18 months or when a child is independently walking. The BSID-II PDI is a reliable and valid tool frequently used to measure fine and gross motor skills.16 Both the AIMS and the BSID-II PDI are discriminative, norm-referenced measures standardized on a population-based sample.15,16 Clinically, the AIMS and BSID-II PDI are not only used to detect children at risk for developmental disorders but they are also often applied as tools to describe developmental change.22,23 The IMP is a recently developed video-based assessment that provides information on a child's motor repertoire and his or her ability to adapt motor behavior to the specifics of the situation. It is suitable for the evaluation of motor development in infants developing typically or atypically. The IMP consists of 5 subscales: variation, variability, performance, symmetry, and fluency. The initial studies of the psychometric properties of the IMP indicate satisfactory to good reliability and validity.17,24 In this study, we used the IMP's performance scale to measure motor abilities in a way comparable to that of the other neuromotor tests.

Neurological condition was assessed with the Touwen Infant Neurological Examination.25 The Touwen Infant Neurological Examination assessment at T3 was used to describe the infant's neurological outcome at the end of the study period. The Touwen Infant Neurological Examination is a reliable neurological examination that includes traditional neurological signs and also quality of motor behavior.25 Children were classified as neurologically normal, minor neurological dysfunction, or neurologically abnormal. The latter implies the presence of a clear neurological syndrome, such as a hemisyndrome. The GMFCS levels were determined to give an impression of the level of functioning, keeping in mind that GMFCS levels are less precise at young ages and should be redetermined after the age of 2 years.26

Back to Top | Article Outline
Data Analyses

We first plotted developmental trajectories of the GMFM-88 and GMFM-66 scores and the other neuromotor tests (AIMS, BSID-II PDI, IMP performance scale) and assessed changes over time with non–parametric-related sample tests (Friedman and Wilcoxon sign rank tests). An α level of less than .05 was considered statistically significant. For the AIMS and the BSID-II PDI, raw scores were used because most infants scored below the minimum score for their age. The use of raw scores has been described for the BSID-II PDI, with the argument that using raw scores may have advantages in comparison with using developmental indexes for certain infants, for example, infants born preterm.27 Similar arguments may be applicable for the AIMS.

Next, we summarized the problems encountered in the application of GMFM-88 and GMFM-66 and formulated suggestions for adaptations to improve application of the GMFM in infancy. Finally, we applied the suggestions in our study sample and calculated adapted GMFM scores. We visualized and compared the developmental trajectories of the original and adapted GMFM scores and compared them with those of the other neuromotor tests.

Back to Top | Article Outline


Neuromotor Development

At the final examination, 6 children were classified as neurologically abnormal, the other 6 showed minor neurological dysfunction. Details of neuromotor condition measured with the various tests are provided in Table 1.

Back to Top | Article Outline
Developmental Trajectories of GMFM Scores and Other Neuromotor Tests

Both GMFM-88 and GMFM-66 scores and the scores on the other neuromotor tests (AIMS, BSID-II PDI, IMP performance scale) changed significantly over time (Friedman P = .0001; Figure 1A). During the first 6 months of observation, the infants showed a substantial increase in scores of all neuromotor tests (Wilcoxon P: AIMS, .002; BSID-II PDI, .003; IMP performance, .002; GMFM-88 total score, .002; GMFM-66, .002). In the second half year of observation, the scores of the AIMS (P = .007), BSID-II PDI (P = .007), IMP (P = .010), and GMFM-66 (P = .003) continued to show statistically significant increases, whereas the increase in GMFM-88 total score failed to reach statistical significance.

Inspection of the individual GMFM-88 data revealed that most infants clearly changed over time. Some children, however, showed only minor changes in GMFM-88 scores, that is, only minor improvement or deterioration in GMFM-88 scores. This phenomenon was observed especially between T2 and T3 (Figure 2A), where the curve of the GMFM-88 in comparison with the other neuromotor tests flattened (Figure 1A). The relative flattening of the GMFM-88 curve was brought about especially by dimension A, where scores dropped with increasing age (Figure 1B).

The GMFM-66 scores improved between T2 and T3 in 11 of 12 infants (Figure 2B). However, we also noted that application of the GMFM-66 resulted in identical scores in 8 of 24 assessments at T0 and T1 (score 22.66; Figure 2B), whereas the GMFM-88 (and the other measures of motor development) suggested more heterogeneity. The GMFM-88 scores at T0 and T1 were identical only on 2 occasions (Figure 2B).

The findings suggest that the GMFM-88 is the better tool to differentiate gross motor function in infants with relatively low motor abilities, whereas the GMFM-66 differentiates better when infants have developed more motor skills.

The advantages and limitations of the GMFM-88 and GMFM-66 in infancy may hamper, however, their longitudinal use in infancy. We therefore embarked on the development of adaptations of the GMFM, which may result in scores that reflect developmental change and within-group variation throughout infancy.

In the next paragraphs, we first describe the problems that we encountered when applying the GMFM-88. We continue with suggestions for solutions to improve the use of the GMFM in infancy to monitor development at young age. Finally, the resulting adaptation of the GMFM was applied to describe developmental change in the 12 infants included in the study.

Back to Top | Article Outline
GMFM: Practical Problems and Their Solutions

The practical problems and suggestions for solutions are summarized in Table 2. The nature of the practical problems during GMFM assessments was related to the infants' abilities, in particular, their ability to crawl. In the 48 measurements that were performed, we assessed 31 times (of the 48 measurements) an infant who did not show progression in the prone position; 6 times an infant who crept (abdominal crawling), and 11 times an infant who crawled (move on all fours, without abdominal support). During 5 assessments, the infant walked independently.

Back to Top | Article Outline
Infants Who Are Not Yet Able to Crawl on All Fours

Few problems were encountered for the infants who were not yet able to crawl. Infants who were not crawling demonstrated items in dimensions A and B, with an occasional addition of the “creeping” item of dimension C (C38). The difficulties met were related to the specifics of infant behavior. The items and the difficulties encountered are listed in Table 2. This table also includes suggestions for adaptation. The ability to creep is not included in the GMFM-66. In our opinion, creeping is an important motor ability, as it is the earliest form of locomotion,3032 and being able to locomote and thereby to explore the environment is associated with improved cognition.33 Therefore, we suggest including the “creeping” item in the evaluation of infant performance. We suggest giving credit for the creeping item if infants are able to crawl, because the presence of the ability to crawl usually puts an end to creeping behavior.

Back to Top | Article Outline
Infants Younger Than 2 Years Who Prefer the Prone Position (With or Without Progression in Prone Position)

When infants are able to roll, some will immediately turn to the prone position when placed in the supine position. This means that they are awarded no points for the supine position, although for most children, it is obvious that they are able to perform supine position items. Our suggestion for adaptation of GMFM scoring in infants who immediately turn into the prone position when placed in the supine position is the following: the score of dimension A would be equal to the score of all prone position items divided by the maximum prone position score of dimension A (including rolling into prone position; maximum 30 points).

Back to Top | Article Outline
Infants Younger Than 2 Years Who Are Able to Crawl on All Fours

By the time infants develop more motor abilities, they usually will not show the previously acquired easier ones. For example, when an infant is able to crawl, he/she usually does not show prone position behavior, as the infant has a strong urge to crawl. The easier items cannot be elicited either, as verbal instruction does not work very well in infancy. The inability to follow verbal instructions makes the situation in infants different from that in older children. As the GMFM is based on behavior that is observed, this may result in relatively low scores and no or rather small improvements in the total GMFM score despite favorable progression of motor development. Especially dimension A is difficult to elicit when an infant is able to crawl. This may induce a sudden drop in the score of dimension A with increasing age (Figure 1B), and interfere with an increase in total GMFM-88 score matching advanced development (Figure 2A).

Our suggestions for adaptation of GMFM scoring in infants who are able to crawl on all fours include the following (for details, see Table 2). First, if the infant does not show the items of dimension A, award an 80% score for dimension A. This is an arbitrary choice based on the dual notion of advanced development and the observation that most children do not reach a maximum GMFM score. Second, remove items A3, B19, B20, B34, C49, C50, D60, D61, E70, E73, and E74. Removal of the items results in lower maximum scores for the GMFM dimensions. Adapted maximum scores for the various dimensions are A, 48 (originally 51); B, 51 (originally 60); C, 36 (originally 42); D, 33 (originally 39); and E, 63 (originally 72). An alternative approach would be to use clinical judgment to infer what the infant would be able to achieve. However, we refrained from this solution to keep as closely as possible to one of the basic principles of GMFM assessment: scores are based on observed behavior.

Back to Top | Article Outline
Developmental Trajectories of the Adapted GMFM Scores

Application of the adapted GMFM scores resulted in GMFM trajectories that like the GMFM-66 trajectories did not flatten (Figure 2C). The adapted GMFM total score improved significantly over time (Friedman P = .0001), not only during the first 6 months of observation (Wilcoxon P = .002) but also during the second 6 months (Wilcoxon P = .003). Inspection of the individual data showed improvement over time in virtually all instances and substantial heterogeneity in scores (only 2 instances of identical scores at T0 and T1). Application of the adapted GMFM abolished the decrease of dimension A of the GMFM-88 with increasing age and changed it into an increase (Figure 1D). Visual inspection of the developmental curves of the various versions of the GMFM (eye balling) of the children who were diagnosed with CP (n = 6) and those not diagnosed with CP (n = 6) suggested that the curves of the 2 groups were largely similar, however, with lower scores in children with probable CP than in children without probable CP (Figure 3).

Back to Top | Article Outline


This study indicated that the application of the GMFM in infancy is associated with practical problems. This held true for the GMFM-88 and GMFM-66, but for each version in a different way, which hampers the longitudinal use of the GMFM during infancy. We therefore made some suggestions for an adapted GMFM for infants younger than 2 years.

The finding that GMFM-88 scores improved less when infants grew older could be largely attributed to behaviors specific to young children: infants, in general, do not show previously acquired motor abilities when they are able to perform more difficult ones. This means that infants who show progress in motor behavior “fail” to show “easier” behavior with negative consequences for the GMFM-88 score. This was also described by Kolobe et al13. They suggested awarding points for some items when other items are already passed,13 similar to some of our adaptations. Jelsma et al,12 who applied the GMFM-88 in infancy, reported nonlinear development. We found similar non-linearity that disappeared with the use of the adapted GMFM. Wei et al,14 however, did not describe practical problems when using the GMFM to assess children from 0 to 3 years and concluded that the GMFM-66 is a reliable and valid instrument at young ages. A possible explanation for the different appreciation of the applicability of the GMFM in the Wei et al study could be related to the group studied. While in the current study and the studies of Kolobe et al13 and Jelsma et al,12 infants at a very high risk for CP with and without an established diagnosis were assessed. On the contrary, Wei et al assessed children who all had the diagnosis CP at the ages of 3 to 36 months. In general, it is difficult to diagnose CP during the first postnatal year. In children with the more severe forms of CP, however, the probable diagnosis of CP can be made earlier.10 This might imply that the youngest infants studied by Wei et al had a severe type of CP and relatively low motor abilities. Such children usually are not able to crawl, which makes the GMFM easier to apply. In addition, a substantial part of the Wei et al sample was older than 2 years, which increases the possibility for instruction of the child. Moreover, in contrast to our study, the focus of the Wei et al study was not on developmental change, as only 40% of the children had more than 1 assessment.

We choose the GMFM-88 as a starting point for our suggestions for an adapted GMFM version to prevent loss of important information, especially in the lower-ability dimensions and to be able to profit from the fact that the GMFM-88 also furnishes dimension scores in addition to the total score. The presence of dimension scores is an advantage because, in infancy, dimension scores reflect functional level in more detail than the total score. The GMFM-66 contains few items from the lower-ability dimensions, thereby possibly limiting its sensitivity to change in the youngest infants. The GMFM-66 and adapted GMFM version reached about the same point after 12 months (Figure 1C). However, in the beginning, the GMFM-66 did not differentiate as well as the adapted version, because the GMFM-66 uses only a few lower-ability items. The adapted GMFM version is designed for clinical use, that is, the documentation of individual trajectories reflecting developmental change from an early age onwards by means of measurable and interpretable items. To this end, we combined the clinical usefulness of the GMFM-88, maintaining the dimension scores, with the longitudinal advantages of the GMFM-66. Inclusion of the lower-ability dimension items has the advantage of applicability in young infants and severely affected infants.

The strengths of our study are the exploratory character to discover and solve practical problems and the longitudinal design with 4 data points at early ages, which made it possible to analyze 48 assessments of 12 children. Infants were at risk of CP, according to well-defined strict inclusion criteria.18 At the data collection points, GMFM scores were compared with those of 3 other neuromotor tests, which allowed for a comprehensive comparison.

In contrast, the small sample size is a major limitation. We based our suggestions for adaptations on critically judging whether items would work in infancy taking into account the age-specific characteristics of the developing infant. It is good to note that some of our decisions have a clinically based character. We tested the adaptations of the GMFM in our small study population. This means that our study certainly precludes firm conclusions. More research in larger groups of infants is needed to study the applicability, reliability, and validity of the adapted GMFM version for infants, and in particular its sensitivity to change. Another limitation is the varying age at which the infants were included and examined. Infants were included between 1 and 9 months corrected age and were examined 3, 6, and 12 months after inclusion. We do not consider this a major limitation, because developmental change is more important in children with or at high risk of CP than age-related performance. In this study, half of the children were diagnosed with CP at 18 months. The age of 18 months is relatively early to diagnose CP. It is conceivable that some of the children grow out of their diagnosis, while others will grow into it.10 Interestingly, visual comparison of the developmental trajectories of GMFM scores of the infants with CP at 18 months with those of children without CP suggested that the GMFM curves in both groups were similar. In addition, it may be regarded as a limitation that we used the raw scores of the AIMS and BSID-II PDI. We used raw scores as they showed changes over time while percentile scores (AIMS) or developmental scores (PDI) were insensitive to change as the children consistently scored below the minimal required level to calculate these scores (see also Janssen et al27). In our comparison between GMFM scores and development measured with other neurodevelopmental tools, we were hampered by the fact that no reference standard exists for testing change in and evaluation of motor function in infancy.34

Finally, it is good to realize that at the end of the study, the infants' ages were 14 to 20 months. All had neurological dysfunction. But it was not clear whether all infants would develop CP, because especially the milder types of CP are difficult to diagnose at early ages.10,11

Back to Top | Article Outline


Application of the GMFM before the age of 2 years is hampered by age-specific limitations. We therefore suggest age and function-specific adaptations to facilitate the use of the valuable GMFM in infancy. Our preliminary data suggest that the sensitivity to change of the adapted GMFM version is promising. Further research is, however, required to assess the reliability and validity of the adapted version of the GMFM in infancy.

Back to Top | Article Outline


1. Russell DJ, Rosenbaum PL, Gowland C, et al. Manual for the Gross Motor Function Measure. Hamilton, ON, Canada,: McMaster University; 1993.
2. Bjornson KF, Graubert C, McLaughlin JF, Kerfeld CE, Clark EM. Test-retest reliability of the Gross Motor Function Measure in children with cerebral palsy. Phys Occup Ther Pediatr. 1998;18:51–61.
3. Bjornson KF, Graubert C, McLaughlin JF, Astley SJ. Inter-rater reliability of the Gross Motor Function Measure. Dev Med Child Neurol. 1994;36(suppl 70):27–28.
4. Palisano RJ, Hanna SE, Rosenbaum PL, et al. Validation of a model of gross motor function for children with cerebral palsy. Phys Ther. 2000;80:974–985
5. Russell DJ, Avery LM, Walter SD, et al. Development and validation of item sets to improve efficiency of administration of the 66-item Gross Motor Function Measure in children with cerebral palsy. Dev Med Child Neurol. 2010;52:e48–e54.
6. Brunton LK, Bartlett DJ. Validity and reliability of two abbreviated versions of the Gross Motor Function Measure. Phys Ther. 2011;4:577–588.
7. Palisano R, Rosenbaum P, Walter S, Russell D, Wood E, Galuppi B. Development and reliability of a system to classify gross motor function in children with cerebral palsy. Dev Med Child Neurol. 1997;39:214–223.
8. Beckung E, Carlsson G, Carlsdotter S, Uvebrant P. The natural history of gross motor development in children with cerebral palsy aged 1 to 15 years. Dev Med Child Neurol. 2007;49:751–756.
9. Rosenbaum PL, Walter SD, Hanna SE, et al. Prognosis for gross motor function in cerebral palsy: creation of motor development curves. JAMA. 2002;288:1357–1363.
10. Paneth E. Establishing the diagnosis of cerebral palsy. Clin Obstet Gynecol. 2008;51:742–748.
11. Ketelaar M, Van Petegem-van Beek E, Veenhof C, Visser J, Vermeer A. Gross Motor Function Measure: Nederlandse handleiding (Dutch Translation). Utrecht, the Netherlands: Universiteit Utrecht; 1999.
12. Jelsma J, Iliff P, Kelly L. Patterns of development exhibited by infants with cerebral palsy. Pediatr Phys Ther. 1999;11:2–11.
13. Kolobe THA, Palisano RJ, Stratford PW. Comparison of two outcome measures for infants with cerebral palsy and infants with motor delays. Phys Ther. 1998;78:1062–1072.
14. Wei S, Su-Juan W, Yuan-Gui L, Hong Y, Xiu-Juan X, Xiao-Mei S. Reliability and validity of the GMFM-66 in 0- to 3-year-old children with cerebral palsy. Am J Phys Med Rehabil. 2006;85:141–147.
15. Piper MC, Darrah J. Motor Assessment of the Developing Infant. Philadelphia, PA: Saunders; 1994.
16. Van der Meulen BF, Ruiter SAJ, Spelberg HCL, Smrkovsky M. Bayley Scales of Infant Development–II. Dutch Version. Lisse, the Netherlands: Swets Test Publishers; 2002.
17. Heineman KR, Bos AF, Hadders-Algra M. The Infant Motor Profile—a standardized and qualitative method to assess motor behaviour in infancy. Dev Med Child Neurol. 2008;50:275–282.
18. Hielkema T, Hamer EG, Reinders-Messelink HA, et al. LEARN 2 MOVE 0-2 years: effects of a new intervention program in infants at very high risk for cerebral palsy; a randomized controlled trial. BMC Pediatr. 2010;10:76.
19. De Vries LS, Eken P, Dubowitz LM. The spectrum of leukomalacia using cranial ultrasound. Behav Brain Res. 1992;49:1–6.
20. De Vries LS, Roelants-van Rijn AM, Rademaker KJ, Van Haastert IC, Beek FJ, Groenendaal F. Unilateral parenchymal haemorrhagic infarction in the preterm infant. Eur J Paediatr Neurol. 2001;5:139–149.
21. Sarnat HB, Sarnat MS. Neonatal encephalopathy following fetal distress. Arch Neurol. 1976;33:696–705.
22. Van Haastert I, De Vries LS, Helders PJ, Jongmans MJ. Early gross motor development of preterm infants according to the Alberta Infant Motor Scale. J Pediatr. 2006;149:617–622.
23. Mayes LC, Cicchetti D, Acharyya S, Zhang H. Developmental trajectories of cocaine-and-other-drug-exposed and non cocaine-exposed children. J Dev Behav Pediatr. 2003;24:323–335.
24. Heineman KR, La Bastide-Van Gemert S, Fidler VV, Middelburg KJ, Bos AF, Hadders-Algra M. Construct validity of the Infant Motor Profile: relation with prenatal, perinatal, and neonatal risk factors. Dev Med Child Neurol. 2010;52:e209–e215.
25. Hadders-Algra M, Heineman KR, Bos AF, Middelburg KJ. The assessment of minor neurological dysfunction in infancy using the Touwen Infant Neurological Examination: strengths and limitations. Dev Med Child Neurol. 2010;52:87–92.
26. Gorter JW, Ketelaar M, Rosenbaum P, Helders PJ, Palisano R. Use of the GMFCS in infants with CP: the need for reclassification at age 2 years or older. Dev Med Child Neurol. 2009;51:46–52.
27. Janssen AJWM, Akkermans RP, Stiener K, et al. Unstable longitudinal motor performance in preterm infants from 6 to 24 months on the Bayley Scales of Infant Development—Second Edition. Res Dev Disabil. 2011;32:1902–1909.
28. Van Haastert IC, Groenendaal F, Waarsenburg MK, et al. Active head lifting from supine in early infancy: an indicator for non-optimal cognitive outcome in late infancy. Dev Med Child Neurol. 2012;54:538–543.
29. Hadders-Algra M. Active head lifting from supine in infancy: a significant stereotypy? Dev Med Child Neurol. 2012;54:489–490.
30. McGraw MB. The Neuromuscular Maturation of the Human Infant. Reprinted: Classics in Developmental Medicine, No. 4. London, England: Mac Keith Press; 1989.
31. Adolph KE, Vereijken B, Denny MA. Learning to crawl. Child Dev. 1998;69:1299–1312.
32. Touwen B. Neurological Development in Infancy. London; England: Heinemann Medical Books; 1976.
33. Clearfield MW. The role of crawling and walking experience in infant spatial memory. J Exp Child Psychol. 2004;89:214–241.
34. Rosenbaum P, Russell D, Cadman D, Gowland C, Jarvis S, Hardy S. Issues in measuring change in motor function in children with cerebral palsy. A special communication. Phys Ther. 1990;70:125–131.
35. Van Haastert IC, Groenendaal F, Waarsenburg MK, et al. Active head lifting from supine in early infancy: an indicator for non-optimal cognitive outcome in late infancy. Dev Med Child Neurol. 2012;54:538–543.
36. Hadders-Algra M. Active head lifting from supine in infancy: a significant stereotypy? Dev Med Child Neurol. 2012;54:489–490.

cerebral palsy; child development; disability evaluation; infants; motor skills; movement; psychometrics

© 2013 Lippincott Williams & Wilkins, Inc.