Abstract
Background and Purpose Perception-action theory suggests a cyclical relationship between movement and perceptual information. In this case series, changes in postural complexity were used to quantify an infant's action and perception during the development of early motor behaviors.
Case Description Three infants born preterm with periventricular white matter injury were included.
Outcomes Longitudinal changes in postural complexity (approximate entropy of the center of pressure), head control, reaching, and global development, measured with the Test of Infant Motor Performance and the Bayley Scales of Infant and Toddler Development, were assessed every 0.5 to 3 months during the first year of life. All 3 infants demonstrated altered postural complexity and developmental delays. However, the timing of the altered postural complexity and the type of delays varied among the infants. For infant 1, reduced postural complexity or limited action while learning to control her head in the midline position may have contributed to her motor delay. However, her ability to adapt her postural complexity eventually may have supported her ability to learn from her environment, as reflected in her relative cognitive strength. For infant 2, limited early postural complexity may have negatively affected his learning through action, resulting in cognitive delay. For infant 3, an increase in postural complexity above typical levels was associated with declining neurological status.
Discussion Postural complexity is proposed as a measure of perception and action in the postural control system during the development of early behaviors. An optimal, intermediate level of postural complexity supports the use of a variety of postural control strategies and enhances the perception-action cycle. Either excessive or reduced postural complexity may contribute to developmental delays in infants born preterm with white matter injury.
Infant development is a process that involves the interaction of multiple systems within an environment.1 Perception-action theory proposes that movement (action) provides perceptual information that, in turn, influences movement and interaction with the environment. This perception-action cycle enhances motivation to move and promotes cognitive development.2 Because infants with brain injury may have difficulty generating movement and perceiving information, their development may be altered. However, the development of infants with brain injury varies greatly.3–5 An understanding of how brain injury influences perception and action may enhance the ability to detect delays and intervene early in development.
Postural control involves controlling the body's position to orient body segments to a specific environment.6 The integration of perceptual information and movement adaptation influences postural control, enables an individual to remain in a stable position to interact with the environment, and serves as a foundation for most motor behaviors.7,8 Thus, a lack of postural control may result from or lead to an atypical perception-action cycle and may contribute to developmental delays.9 In this case series, we describe how changes in postural control during development may have related to action and perception in 3 infants born preterm with brain injury.
Complexity in Development and Rehabilitation
Fluctuating periods of stability and variability are well-accepted features of typical development.10 Variability is a necessary part of development leading to adaptation.2,11 In recent years, the temporal structure of variability, known as complexity, has gained increased interest in development and rehabilitation literature. Optimal complexity, which has been described as a state midway between excessive order (high predictability) and excessive disorder (no predictability), characterizes healthy human body function.12 Optimal complexity signifies effective cooperation among the participating subsystems, enhancing the system's ability to integrate perceptual information and adapt actions to changing task demands.13
Postural complexity decreases as infants learn to sit, control their head, and reach while in the supine position.14,15 Previous research proposed that this change reflects the infant's ability to control the body and optimally use the available degrees of freedom.9,15 In other words, postural complexity represents the variety of postural control strategies that are used to attain a goal.9
Limited early postural complexity has been associated with atypical or delayed development.16,17 Infants diagnosed with or at risk for cerebral palsy demonstrated more repetitive (less complex) postural sway strategies in the early stages of sitting.16 Infants born preterm used more repetitive postural sway patterns at an adjusted age (AA; ie, chronological age minus weeks preterm) of 1 to 3 weeks and demonstrated less adaptive postural control strategies while learning to control their head in the midline position and during a reaching task.17,18
Evidence on the role of postural complexity in optimal health, early development, and developmental disabilities suggests that alterations in postural complexity may be predictors of disability.16,19 A better understanding of the role of postural complexity in the prediction of disability may support the identification of clinically useful tools for predicting motor disability.20,21
In this case series, we examined the relationship between altered postural complexity and development in infants who were at high risk for motor delay and considered how the findings relate to the perception-action theory. Changes in postural complexity were used to quantify an infant's action and perception during the development of early motor behaviors. For example, optimal postural complexity would represent the infant's ability to use a variety of postural control strategies to meet the task demands, providing varied perceptual information and enhancing further action (Fig. 1A); reduced postural complexity would signify the use of a limited number of postural control strategies to accomplish a motor behavior, resulting in limited perceptual information and action (Fig. 1B); and excessive postural complexity would signify the use of a large number of postural control strategies and the inability to select the most efficient strategies (Fig. 1C).9 Postural complexity was measured in a “no-toy condition” and in a “toy condition.” The use of a visual and reaching target in the toy condition changed the task demands, requiring the infant to adapt postural complexity. The findings observed in both conditions are presented to enable the reader to consider the implications of postural complexity under different task demands.
Theoretical model of the relationship between the perception-action cycle and postural complexity. Each colored line represents an option for perception and action. Postural complexity determines how many options are available and whether the infant can purposefully select a strategy. (A) Optimal variability supports strategy selection and the perception-action cycle. Each of the 4 postural control strategies provides different perceptual information. (B) Limited postural complexity results in few postural control strategies being available and restricts the information gained from action and perception. In this case, only 2 possible sets of action and perception from which to select or learn are available. (C) Excessive postural complexity results in many postural control strategies being available. However, the perception-action cycle may not provide the information needed, or the postural control system may not be able to select the most efficient strategy.
Case Descriptions
Infant 1
Infant 1 was an African American female surviving twin born at 24 weeks of gestation with a birth weight of 480 g and APGAR scores of 1 at 1 minute and 5 at 5 minutes. She was in the neonatal intensive care unit (NICU) for 149 days with complications, including discordant twin syndrome, respiratory distress syndrome, osteopenia, hypothyroidism, layringomalacia, stage 1 retinopathy of prematurity, and atrial septal defect. She was diagnosed with bilateral grade 4 intraventricular hemorrhage (IVH) and cystic periventricular leukomalacia. She developed hydrocephalus, and ventriculoperitoneal shunts were placed bilaterally. She was on a ventilator for respiratory support for 74 days.
Infant 1 was discharged with a nasal gastric feeding tube and oxygen to reside with both parents and a school-age sibling. Her mother reported having a college degree and a household income of 1.5 to 2 times the federal poverty line for a family of 4. Both parents were employed outside the home. Infant 1 was referred to a local early intervention program at NICU discharge and began weekly physical therapy intervention at an AA of about 4 months.
Infant 2
Infant 2 was an African American male born at 29 weeks of gestation with a birth weight of 1,280 g and APGAR scores of 2 at 1 minute and 4 at 5 minutes. He was in the NICU for 65 days with complications, including respiratory distress syndrome, right grade 4 IVH, left grade 2 IVH, and periventricular leukomalacia. He was on a ventilator for 5 days.
Infant 2 was discharged to home with both of his parents and his teenage sister. His mother reported having less than a high school education and a household income less than the federal poverty line for a family of 4. Neither parent had employment. Infant 2 was referred to a local early intervention program and began receiving physical therapy 1 or 2 times per month at an AA of about 2 months.
Infant 3
Infant 3 was an African American male born at 24 weeks of gestation with a birth weight of 840 g and APGAR scores of 1 at 1 minute and 5 at 5 minutes. He was in the NICU for 154 days with complications, including patent ductus arteriosus (which was closed with indomethacin), bronchopulmonary dysplasia, and a unilateral inguinal hernia. He had right grade 3 IVH, left grade 4 IVH, and periventricular leukomalacia.
Infant 3 was discharged home to reside with his mother, grandmother, and 2 siblings. His mother reported having a college degree and a household income of 1.5 to 2 times the federal poverty line for a family of 4. She worked part time outside the home; his grandmother worked full time. Infant 3 was referred to a local early intervention program and began receiving weekly occupational therapy at an AA of about 3 months. At an AA of 6 months, infant 3 developed hydrocephalus and received a ventriculoperitoneal shunt and medical management for the onset of seizures.
Clinical Impression 1
The 3 infants included in this case series were all at high risk of developmental disabilities because of their preterm birth status, IVH, and periventricular leukomalacia. All 3 infants had 2 involved caregivers and received early intervention services. These infants were part of a larger study of postural complexity and early motor behaviors.18 Early diagnosis of brain injury and similar developmental risk factors made the treatment of these 3 infants ideal for revealing the relationships among changes in postural complexity, early development, and developmental outcomes.
Examination Protocol
A combination of norm-referenced standardized assessments and sensitive measures of behaviors was completed to quantify global development, specific behaviors, and postural control. Assessment visits were completed every 2 weeks from 5 weeks before the infant's due date (35 weeks of gestation) or when the infant was medically stable through an AA of 3 months and monthly from an AA of 3 months to an AA of 6 months. Developmental follow-up visits were completed at an AA of 9 months and an AA of 12 months. Assessment visits were completed by 1 of 2 researchers who were trained in the reliable use of all assessment tools. During assessments, caregivers were encouraged to meet the infant's needs so that the infant was in a positive or neutral behavioral state. Because of the medical fragility of the infants, no infant was excluded because of missing or incomplete data. All assessments were video recorded for behavioral coding and scoring. Medical records were used to confirm developmental outcomes whenever possible.
Developmental Assessments
The Test of Infant Motor Performance (TIMP) was completed at each visit through an AA of 4 months. The TIMP is a reliable and valid norm-referenced assessment of postural and selective motor control for infants from 34 weeks of gestation to an AA of 4 months.22–24 The TIMP z score was used to compare infants' development with TIMP normative data and to evaluate changes over time.24 The Bayley Scales of Infant and Toddler Development, 3rd edition, were completed at 4, 5, 6, 9, and 12 months of age.25 The Bayley scales are a reliable and valid norm-referenced standardized assessment of gross and fine motor, cognitive, receptive, and expressive language abilities for infants and toddlers.25,26 Scale scores are presented as a comparison with the Bayley scales normative sample.
Reaching, Looking, and Head Control Behaviors
Behaviors and postural control were assessed while the infant was in the supine position under 2 conditions: without a visual stimulus (no-toy condition) and with a visual stimulus (toy condition). During the toy condition, five 60-second trials were completed with a toy presented at 75% of arm's length directly over the torso. Reaching was quantified as the percentage of time either hand was in contact with the toy during the toy condition. The sum of the percentages of time each hand was in contact with the toy was calculated. Looking was coded when the infant's eyes were directed at the toy and is presented as a percentage of the toy condition. Controlling the head in the midline position was coded any time the infant's head was within 30 degrees of the midline position. The percentage of the assessment during which the infant's head was in the midline position was calculated for each condition.18 Behavioral coding was completed by use of video recordings of the assessments and the MacShapa v1.1.2a (Department of Mechanical and Industrial Engineering, University of Illinois at Urbana-Champaign, Urbana, Illinois) coding program. Coder reliability was documented for the larger study.18
Postural Control
The complexity of center-of-pressure displacement (postural complexity) was assessed while the infant was in the supine position for 5 minutes in the toy condition and in the no-toy condition by use of a Conformat (Tekscan, South Boston, Massachusetts) pressure-sensitive mat sampling at 5 Hz.17 Behavioral data were used to identify continuous center-of-pressure time series of 500 data samples in which the infant was in the supine position, the infant was alert, no one was touching the infant, and the infant was not touching the toy.
As described in detail in our previous work, postural complexity was quantified as the approximate entropy in the caudal-cephalic and medial-lateral directions (ApENcc and ApENml).9,14,17 Postural complexity represents the repeatability of the postural control strategy or the pattern of variability within a time series. Repetitive center-of-pressure movement or the use of a limited number of postural control patterns is signified by low complexity or approximate entropy values. An expanded explanation of these terms and clinical examples are provided in the work of Dusing and Harbourne.9 The dependent variables of ApENcc and ApENml were transformed by use of a (natural) logarithmic transformation to more closely approximate a normal distribution, yielding Ln(ApENcc) and Ln(ApENml), respectively. Along with the postural complexity data for the infants in this case series, data from a cohort of 22 infants born full term with typical development are shown.14
Examination Outcomes
Infant 1
Infant 1 displayed developmental delays at all ages. Her TIMP z scores were greater than 1 standard deviation below the mean, and the raw score increased only 16 points during the 4 months of assessment using the TIMP (Fig. 2). She had delays in all developmental domains on the Bayley scales; however, her cognitive and language skills were areas of relative strength (Fig. 3A). At 12 months of age, she could hold her head up when her trunk was supported in the sitting position and roll from her side to her back. In a supported sitting position, she was able to reach for toys, pick up a cube, and engage her caregivers socially, and she showed interest in books.
Changes in Test of Infant Motor Development (TIMP) z scores before an adjusted age (AA) of 4 months. Not all infants were assessed in each age band. GA=gestational age.
Changes in scaled scores on the Bayley Scales of Infant and Toddler Development, 3rd edition, for infants 1 (A), 2 (B), and 3 (C). The mean for the normative data for the scaled scores was 10, and the standard deviation was 3.
Infant 1 did not learn to bring her head to the midline position in the no-toy condition or reach for a toy before an AA of 6 mo (Fig. 4A). She developed the ability to hold her head in the midline position in the toy condition at an AA of 6 mo (Fig. 5A).
Motor behaviors and corresponding postural complexity during the no-toy condition. (A) Head kept in the midline position. (B) Changes in approximate entropy of the center-of-pressure time series in the medial-lateral direction (ApENml) during the no-toy condition. Error bars represent the standard error of the mean for the group of infants born full term. (C) Change in approximate entropy of the center-of-pressure time series in the caudal-cephalic direction (ApENcc) during the no-toy condition. Error bars represent the standard error of the mean for the group of infants born full term.
Motor behaviors and corresponding postural complexity during the toy condition. (A) Head kept in the midline position. (B) Looking at the object. (C) Changes in approximate entropy of the center-of-pressure time series in the medial-lateral direction (ApENml) during the toy condition. Error bars represent the standard error of the mean for the group of infants born full term. (D) Changes in approximate entropy of the center-of-pressure time series in the caudal-cephalic direction (ApENcc) during the toy condition. Error bars represent the standard error of the mean for the group of infants born full term.
Infant 1 had postural complexity values that were higher than the reference values at the time of her initial assessments, and the values generally decreased over time in each condition (Figs. 4B and 4C). In the toy condition, infant 1's postural complexity rapidly declined from 1.5 to 3 months and then returned to values similar to the reference values when she began to keep her head in the midline position (Figs. 5C and 5D). We propose that the drastic reduction in postural complexity at 3 months may have resulted from this infant using her limited resources to visually focus on the toy, allowing for limited attention for active postural control.
Clinical Impression 2 for Infant 1
Infant 1 had developmental delays and atypical postural complexity at the first visit. However, her ability to adapt her postural complexity when focusing on a toy or learning to control her head is an example of adaptive motor behaviors enhancing cognitive development, that is, action (motor system) and perception (sensory system) leading to exploration and learning from the environment. It also is possible that this infant's cognitive strength enabled her to adapt her postural complexity and improve her motor behaviors. In either interpretation, this infant's characteristics highlight the strong interaction between the motor system and the cognitive system in early infancy. Infant 1 was diagnosed with quadriplegic cerebral palsy at 2 years of age.
Infant 2
Infant 2's developmental scores improved over the first months of life, with TIMP z scores within 0.5 standard deviation of the mean by an AA of 2 months (Fig. 2) and scores on all subtests of the Bayley scales within 1 standard deviation of the mean through an AA of 6 months (Fig. 3B). At an AA of 12 months, his motor skill and receptive communication scores were within 1 standard deviation of the mean, but his cognitive and expressive language scores were more than 1 standard deviation below the mean, suggesting developmental delays. At an AA of 12 months, infant 2 was able to pull himself to a standing position and cruise, bang blocks in the middle of his torso, pick up a small object using an immature pincer grasp, take blocks out of a cup, and vocalize socially.
Infant 2 learned to keep his head in the midline position by an AA of 4 months in the no-toy condition and by an AA of 3 months in the toy condition (Figs. 4A and 5A). He looked at toys once he could keep his head in the midline position (Fig. 5B). He was able to make contact with the toy in the toy condition starting with 1% of the toy condition at 4 months and increasing to 27.2% at 5 months and 79.7% by an AA of 6 months.
Interestingly, this infant had a pattern of postural complexity opposite that of the reference groups in the no-toy condition (Figs. 4B and 4C). He had high postural complexity at the first visit, lower complexity at 1 or 2 months, and transient increases at 2 and 4 months, just before he learned to keep his head in the midline position. In the toy condition, infant 2's postural complexity declined steadily, compared with the reference values, in both directions (Figs. 5C and 5D). This decline in complexity in the toy condition likely coincided with the infant's attempts to start reaching. Although the decline was faster than that shown by the reference values, it demonstrated a similar developmental pattern.
Clinical Impression 2 for Infant 2
We propose that infant 2's limited use of postural complexity early in development affected his learning through exploration and contributed to his short-term motor delay. Although his motor behaviors were age appropriate after an AA of 2 months, his use of repetitive postural control strategies limited his ability to learn from the perception-action cycle. Consistent with this idea, infant 2 had cognitive delays at 12 months of age, and did not appear to have cerebral palsy; however, no follow-up data are available beyond 14 months.
Infant 3
Infant 3's gross motor development in the first months of life was age appropriate, with TIMP z scores within 0.5 standard deviation of the mean on 4 of 5 assessments (Fig. 2). He began to develop delays, with gross motor development being most limited, at 5 months of age (Fig. 3C). A few weeks after his assessment at an AA of 6 months, he was admitted to the hospital with hydrocephalus and seizures, which were increasing in frequency. His seizures were difficult to control, and after a shunt was placed, ventriculomegaly took several months to resolve. This infant probably began developing hydrocephalus slowly in the months before the shunt placement, and the hydrocephalus altered his developmental progression during this case series. Although his developmental scores on the Bayley scales reflected significant decreases in all developmental areas at 9 months, these changes were transient, as his scaled scores increased in all domains—except for the fine motor domain—at 12 months (Fig. 3C). At an AA of 12 months, infant 3 was able to creep on his hands and knees, pull himself to a standing position, pick up a cube, and respond to his name. At an AA of 14 months, infant 3 was seen in the NICU follow-up clinic and was able to walk independently.
Infant 3 learned to keep his head in the midline position at an AA of 3 months in the no-toy condition and at an AA of 2 months in the toy condition (Figs. 4A and 5A). The length of time he spent looking at the toy gradually increased, but he did not reach for the toy consistently until an AA of 6 months (12.9% at 5 months and 29.4% at 6 months).
In both conditions, infant 3 demonstrated postural complexity similar to that of the reference infants through 3 months (Figs. 4B and 4C) and learned to keep his head in the midline position at a similar age (Fig. 4A). The higher-than-typical complexity in both conditions from 4 to 6 months as well as the long delay between achieving midline head control and reaching suggest that this infant's postural complexity was not typical after an AA of 4 months.
Clinical Impression 2 for Infant 3
Although infant 3 was at high risk for developmental delays, his development of postural complexity and motor behaviors (including head control) was age appropriate until an AA of 4 or 5 months, when his postural complexity began to deviate from that of his peers; this change was followed by gross motor delays. We propose that infant 3 was demonstrating excessive postural complexity related to developing hydrocephalus and that the excessive postural complexity reduced the efficiency of his movements and led to his developmental delays. The recovery of his motor skills was noted once his hydrocephalus and seizures were managed at 12 months of age, and he did not appear to have cerebral palsy; however, no records were available beyond 12 months of age.
Discussion
Postural complexity and developmental outcomes varied among the 3 infants in this case series, highlighting both the challenges of predicting disability and the relationship between the perception-action cycle and adaptive postural control.
Our previous work suggested that postural complexity in infants born full term is highest at birth and decreases during the development of new motor behaviors, including head control and reaching in the supine position.14 Although group differences exist, the general patterns of changes in postural complexity are similar for infants born full term and infants born preterm.18
All 3 infants in this case series demonstrated some changes in postural complexity over time, between conditions, or both, and these changes may have been related to their developmental outcomes. Infant 1's optimal postural complexity in early development supported her use of action and perception to learn, even with significant motor impairments. Her postural complexity was limited for a longer period in the toy condition when she was challenged by the additional task demands of visually focusing on the toy and attempting to reach it. Infant 2's limited early postural complexity and early motor delay limited his ability to learn through the perception-action cycle. He had different adaptive responses in the toy and no-toy conditions over time. Although he went on to have typical motor development, he had cognitive delays. It is unclear whether his improved postural complexity would eventually contribute to improved cognitive development. Infant 3's adequate early postural complexity in both conditions helped him learn how to adapt his postural complexity and use perceptual information. Even after a period of medical instability, motor delay, and limited complexity, he was able to adapt his motor abilities and achieve typical developmental outcomes.
Conclusion
The infants born preterm with white matter injury in this case series had altered postural complexity accompanying developmental delays. These findings highlight the importance of early complexity in learning from the environment, responding to different task demands, and overcoming medical status changes. Both excessive postural complexity and reduced postural complexity altered the infants' abilities to act on the world around them and use perceptual information to modify their actions. An optimal, intermediate level of postural complexity supports the use of a variety of postural control strategies and enhances the perception-action cycle. Further research is needed to determine whether there are critical periods in which altered postural complexity has the greatest impact on each developmental domain, how postural complexity relates to measures of movement variability (such as the General Movement Assessment20), and whether physical therapy intervention can modify postural complexity.
Footnotes
Dr Dusing and Dr Galloway provided concept/idea/project design, fund procurement, facilities/equipment, and institutional liaisons. All authors provided writing. Dr Dusing and Dr Izzo provided data collection. Dr Dusing and Dr Thacker provided data analysis. Dr Dusing provided project management and patients. Dr Thacker and Dr Galloway provided consultation (including review of manuscript before submission). The authors thank the families and research staff in the Motor Development Lab at Virginia Commonwealth University and the Neonatal Intensive Care Unit at the Children's Hospital of Richmond.
The Committee for the Protection of Human Subjects at Virginia Commonwealth University approved this project, and parents signed consent statements before project participation began.
An abstract of similar data has been submitted to the American Academy of Developmental Medicine and Child Neurology 68th Annual Meeting; September 10–13, 2014; San Diego, California.
This research was funded, in part, by the National Institutes of Health (1K12HD055931 and UL1TR000058) and the A.D. Williams Trust Fund.
- Received January 28, 2014.
- Accepted May 19, 2014.
- © 2014 American Physical Therapy Association
References
Issue highlights
