In-Home Kicking-Activated Mobile Task to Motivate Selective Motor Control of Infants at High Risk of Cerebral Palsy: A Feasibility Study

Abstract

Objective

Children with spastic cerebral palsy (CP) have gait impairments resulting from decreased selective motor control, an inability to move the leg joints independently of one another, relying on excessive flexion or extension coupling across the 3 joints. Infants with white matter injury are at high risk of CP and have decreased selective motor control as early as 1 month corrected age. An in-home kicking-activated mobile task was developed to motivate more selective hip-knee control of infants at high risk of CP. The purposes of this study were to determine the feasibility of the in-home mobile task and to determine whether infants at high risk of CP and infants with typical development (TD) learn the association between their leg movements and mobile activation.

Methods

Ten infants at high risk of CP based on neuroimaging and 11 infants with TD participated in this cohort study at 3.5 to 4.5 months corrected age. Each infant participated in the in-home kicking-activated mobile task for 8 to 10 min/d, 5 d/wk, for 6 weeks. Learning was assessed weekly based on an increase in the time that the infant demonstrated the reinforced leg actions when interacting with the kicking-activated mobile compared with spontaneous kicking.

Results

With regard to feasibility, participation averaged 92% for infants at high risk of CP and 99% for infants with TD. With regard to learning, the group at high risk of CP demonstrated learning of the task for 2 of 6 weeks, whereas the group with TD demonstrated learning for all 6 weeks.

Conclusions

Infants at high risk of CP demonstrated learning of the kicking-activated mobile task but at a reduced amount compared with infants with TD. Further research is necessary to determine whether the kicking-activated mobile task has potential as an intervention to motivate more selective hip-knee control and improve walking outcomes of infants at high risk of CP.

Impact

This study investigated the feasibility of an in-home kicking-activated mobile task, a discovery learning task designed to motivate infants at high risk of CP to engage in the intensive task practice necessary to promote their learning abilities and selective motor control.

Lay Summary

CP is a lifelong disorder of movement caused by abnormal development or early damage to the brain. If an in-home infant kicking-activated mobile task could be used to motivate certain types of age-appropriate leg movements of infants who are at high risk of CP, the task could help improve walking outcomes, which eventually could contribute to improving children’s ability to participate in daily life. This study showed that infants at high risk of CP did learn the infant kicking-activated mobile task but at a much reduced amount compared with infants who are developing typically; so, this is a first step in determining whether the task has potential to motivate more age-appropriate leg movements in infants at high risk of cerebral palsy.

Children with spastic cerebral palsy (CP) experience lifelong walking limitations resulting from an inability to move the hip, knee, and ankle joints independently of one another, relying on excessive flexion or extension coupling across the 3 joints.^1,2 This impairment in selective motor control is caused by an early lesion to the motor areas of the brain, which disrupts the formation of appropriate neural connections.^3,4 Animal studies support that the ability of the lesioned motor system to establish appropriate neural connections may be enhanced by activity-based interventions during the critical period of neuromotor development, before 6 to 12 months of age in humans.⁵ Thus, to optimize the neural connections that support selective motor control, it may be necessary to promote selective motor control of infants at high risk of CP before 6 months of age. However, existing interventions for infants at high risk of CP do not specifically include activities to promote selective motor control.^6,7 The objectives of this study were to determine the feasibility of a task designed to motivate more selective leg joint control during the critical period of neuromotor development and to investigate whether infants at high risk of CP can learn the task. A future study will investigate whether participating in the task motivated more selective motor control.

Preterm infants with white matter injury have decreased selective leg joint control as early as 1 month corrected age⁸ (CA; corrected for preterm birth) and are considered to be at high risk of CP, with estimates indicating that 24% to 67% will develop CP.⁹ We developed an innovative method of promoting selective leg joint control for infants at high risk of CP: positioning infants under an infant kicking-activated mobile and reinforcing vertical, antigravity leg movements with music and movement of the mobile.^10–12 We first trialed the infant kicking-activated mobile with infants born very or extremely preterm, with estimates indicating that 5% to 10% will develop CP.¹³ Our research supports that 3-month-old infants born full-term who learned the association between their leg movements and mobile activation changed their hip-knee coordination when interacting with the mobile from an immature, in-phase hip-knee coordination (hip and knee flex and extend in synchrony) to a more mature, selective hip-knee coordination (hip flexes while knee extends).¹⁰ In contrast, 3-month-old CA infants born preterm did not learn the association between their leg movements and mobile activation. When the infants born preterm participated again at 4 months CA, the infants learned the association, but the infants born preterm who learned the association did not change their hip-knee coordination.¹² These studies were limited to 2 days in a laboratory setting. Two days may not provide sufficient practice for 3-month-old infants born preterm to generate more selective hip-knee control.

For this study, we chose to enroll infants at high risk of CP. Infants at high risk of CP have known impairments in selective motor control⁸ and are at high risk for learning disabilities and intellectual disability.¹⁴ Therefore, infants at high risk of CP may require daily practice for several weeks to both learn the association between their leg movements and mobile activation and generate more selective hip-knee control. To support this amount of practice, we developed an in-home infant kicking-activated mobile system. In this study, infants at high risk of CP and infants with typical development (TD) at 3.5 to 4.5 months of age played with the infant mobile for 8 to 10 min/d, 5 d/wk, for 6 weeks.

This study had 2 objectives. The first objective was to determine the feasibility of the in-home mobile task based on participation, child’s arousal during the study, and parents’ perceptions of the task. We hypothesized that the in-home mobile task would be feasible for infants and their families based on participation of >80%, infants happy and alert >80% of the time when playing with the mobile, and task implementation averaging 20 to 30 minutes with parents rating it as “somewhat easy.” The second objective was to determine whether infants at high risk of CP and infants with TD learned the association between their leg movements and mobile activation. We hypothesized that infants at high risk of CP would learn the association between their leg movements and mobile activation during the last 3 weeks of the study, whereas infants with TD would learn the association each of the 6 weeks of the study.

Methods

Participants

Infants at high risk of CP were recruited from the Newborn and Infant Critical Care Unit at Children’s Hospital Los Angeles and the Neonatal Intensive Care Unit at Children’s Hospital Orange County from March 2016 to December 2018. Infants with TD were recruited from the Los Angeles and Orange County communities from July 2015 to March 2017.

Infants were included in the high risk of CP group if they had white matter injury on magnetic resonance imaging conducted at term equivalent age documented in medical reports and confirmed by a study author (J.L.W.). Infants with TD were included if they were born full-term without birth complications and were developing typically as per parent report and scores of ≥10th percentile on the Bayley Scales of Infant and Toddler Development, Third Edition (Bayley-3)¹⁵ administered at 3.5 months of age. Infants at high risk of CP and infants with TD were excluded for maternal substance abuse, congenital anomalies, and severe visual or hearing deficits that would hinder their ability to see and hear the mobile as reported by parents.

A sample size calculation was computed using G*Power 3.1¹⁶ (Dusseldorf, Germany) based on the learning outcome in previous studies.^10,12 An effect size of 1.23 was obtained using an F test and an analysis of variance for repeated measures within factors. With an alpha of .05 and a power of .95, a necessary sample size was estimated at 4 per group. We increased the sample size to 10 per group to raise statistical power to detect potential within-group differences in our new population, infants at high risk of CP.

Parents provided written informed consent prior to participation in the study, and families received a small gift for participation. The Institutional Review Board at Children’s Hospital Los Angeles approved the study (CHLA-15-00354), with ceded review by the University of Southern California.

Procedure

Experimental setup

Data collection took place in the infant’s home and occurred each time the infant participated in the in-home kicking-activated mobile task. The infant was placed supine on the floor under an infant kicking-activated mobile with a Microsoft Kinect (Microsoft Corp, Redmond, WA) sensor on each side (Fig. 1; Suppl. Video). The infant kicking-activated mobile consisted of an infant mobile (Fisher-Price, 2008) and 2 Microsoft Kinect sensors connected to a laptop computer (Lenovo) with custom software. Two custom, rigid arrays with 2 light-emitting diodes (LED) were attached bilaterally to the infant’s pelvis and shank using soft straps with the yellow LED over the lateral trunk, red LED over the lateral hip, blue LED over the lateral knee, and green LED over the lateral ankle. A video camera was overhead to capture facial expressions and eye gaze.

Figure 1.

In-home infant kicking-activated mobile task. A light-emitting diode (LED) placed on each foot moves vertically to cross the virtual threshold (black dashed line) to activate the mobile overhead.

The setup of the in-home kicking-activated mobile task included the following 6 steps: arranging the system as depicted in Supplementary Figure 1 with the mobile over the infant and the Kinect sensors on each side of the infant, turning on the mobile and computer, aligning the Kinect sensors using a single LED placed on a specific mark on the black towel, pushing a button to initiate the exposure time for the Kinect sensors (the Kinect adjusted the light entering the sensors so that the sensors only collected data from the colored LED lights placed on the infant and no other colors in the area), putting straps and sensors on the infant, and starting the video camera and computer program. Once the data collection was complete, the mobile, computer system, and video camera were turned off, and the straps and sensors were removed from the infant. Each time the infant participated in the infant kicking-activated mobile task, overhead video data were collected, and position data from the LED markers were automatically collected by the infant kicking-activated mobile system.

In-home kicking-activated mobile task

Each infant participated in the in-home kicking-activated mobile task for 8 to 10 min/d, 5 d/wk, for 6 weeks (a total of 30 days).

Day 1 of each week consisted of a 2-minute baseline condition. The infant kicked spontaneously, but leg actions did not cause the mobile to play music or rotate. The computer used position data from the green LED on each foot to compute an individualized threshold for mobile activation at a height that was 1 SD above the average height of both feet during the 2-minute baseline.^10–12 This was immediately followed by an 8-minute acquisition condition in which the mobile rotated and played 1 of 4 songs when either foot was above the threshold to a maximum of 3 seconds. After 3 seconds, the mobile only reactivated if the infant moved at least 1 foot below the threshold and then again moved 1 foot vertically to cross the threshold. This encouraged lower extremity movement rather than holding the feet above the threshold.

Days 2 through 5 of each week consisted of an 8-minute acquisition condition. The mobile rotated and played music when either foot was above the threshold that was computed on day 1 of the week. The threshold was recomputed on day 1 of each week to control for maturation.

For all days of week 1 of the study, parents were trained by a physical therapist researcher (B.S.) on the implementation of the in-home infant-kicking activated mobile task using demonstration and a manual. Day 1 of weeks 2 through 6, the researcher (B.S.) or a physical therapist research assistant changed the stuffed toys that hung from the mobile to increase infant engagement, set up the computer to include the 2-minute baseline immediately before the 8-minute acquisition, and observed the parent implementing the mobile task to confirm that it was being implemented correctly. For days 2 through 5 of weeks 2 through 6, the parents independently implemented the mobile task.

During the mobile task, parents were instructed to supervise the infant while remaining quiet and not touching the infant. If the infant became fussy or cried, parents were instructed to use soothing words or a pacifier; however, if the infant cried for >2 minutes, the session was ended because the data would not be used for that day.

Anthropometric and developmental measures

During the first week of the study, each infant was weighed on a digital electric scale (Health-O-meter, Sunbeam Products Inc., Boca Raton, FL, USA), and height was measured using a measuring tape. Each infant was also assessed using the cognitive, language, and motor subtests of the Bayley-3 by 1 researcher (B.S.).¹⁵ These 3 subtests of the Bayley-3 have good content and construct validity, a high level of internal consistency (r = 0.71–0.93) for ages 3 to 4 months, and good test–retest reliability (r = 0.67–0.80) for ages 2 to 4 months.¹⁵

Parent perception survey

At the conclusion of the in-home kicking-activated mobile task, parents were given a survey requesting their perceptions of the mobile task. The survey included questions on the amount of time required to perform the mobile task, the ease of using the mobile system, the child’s affect during setup and when playing the mobile task, and recommendations to improve the system. Supplementary Appendix shows a copy of the survey.

Follow-up developmental assessment

At 2 years of age, each infant at high risk of CP was assessed using the cognitive, language, and motor subtests of the Bayley-3 by 1 researcher (B.S.).¹⁵ These 3 subtests have good content and construct validity, a high level of internal consistency (r = 0.71–0.91), and good test–retest reliability (r = 0.71–0.88) for 2 years of age.¹⁵ Parents were also asked questions, and medical reports from high-risk infant clinic visits were examined to determine when the infant started walking and whether the infant had any diagnosed conditions, such as CP or developmental delay.

Data Reduction

Feasibility

Feasibility was assessed using measures of participation (computer recorded), arousal level (assessed via video tapes by evaluators masked to group), and parent perceptions (collected using survey).

Learning

The dependent measure for learning was defined as the reinforced lower extremity action (RLA) ratio. RLA during the mobile condition was defined as the amount of time that the infant activated the mobile. Since the mobile did not activate during baseline, RLA during the day 1 baseline condition was computed post hoc as the time the mobile would have been activated using the coordinates of the foot LEDs that crossed the threshold.^10–12 The RLA ratio was computed weekly for days 2 through 5 as a ratio of the RLA during the entire 8-minute acquisition condition divided by the RLA of the 2-minute day 1 baseline condition. By definition, the RLA ratio for day 1 baseline is 1.

Learning of each group each week and each day was measured statistically by determining whether the RLA ratio during the entire 8-minute acquisition condition on days 2 through 5 exceeded the RLA ratio during the day 1 baseline condition.^10–12 Individual infants were categorized as learners each day if the RLA ratio during the entire 8-minute acquisition condition of day 2 through 5 was ≥1.5 times the baseline condition of day 1.^10,12

Arousal

Video tapes were coded for arousal and rolling using Datavyu, a video coding tool (Datavyu Team; Databrary Project, New York University, New York, NY, USA), by 2 evaluators masked to group. Interrater reliability was assessed using 15 videos randomly selected from 15 infants participating in the study. The interrater reliability ICC was .94 (95% CI = 0.91–0.96) based on a single-rating, absolute-agreement, 2-way mixed-effects model using SPSS v.25 (IBM, Armonk, NY). Once reliability was established, the remaining videos were coded by 1 of the 2 evaluators. The arousal scale is sleeping, drowsy, alert, fussy, crying. If an infant slept, cried, or rolled for >2 minutes during data collection, then the infant was not participating in the task and the data for that day were not used.^10,12 If rolling for >2 minutes occurred for 3 data collections, then infants were excluded from the study.

Data Analysis

Feasibility

For each infant group, feasibility data were described using frequency counts for nominal data, median and range for ordinal data, and means and SDs for interval data.

Learning

For each infant group and for each week, learning was assessed on a weekly and daily basis. Mixed regression models using repeated measures with a compound symmetry covariance structure were used to test differences of RLA ratio each week between the day 1 baseline condition and the mobile condition. Mixed regression models using repeated measures with a heterogeneous autoregressive covariance structure were used to test differences of RLA ratio each day between the day 1 baseline condition and the mobile condition.

For all mixed regression models, the selected covariance structure was chosen because it was consistent with the study design and was the best fit when competing covariance structures were tested using Bayesian information criteria and Akaike information criteria. Statistical tests were completed using SAS (version 7.0; SAS Institute Inc, Cary, NC, USA) with an alpha value of .05. Preplanned comparisons were performed using a Bonferroni correction to adjust for multiple comparisons.

Role of the Funding Source

This study was supported by the National Institutes of Health (NIH) Eunice Kennedy Shriver National Institute of Child Health and Human Development under award number K12-HD055929 (principal investigator: K. Ottenbacher) to B. Sargent and under award number K23-HD099309 to J.L. Wisnowski. The content is solely the responsibility of the authors and does not necessarily represent the official views of NIH. In addition, the study was funded by the Academy of Pediatric Physical Therapy, a Southern California Clinical and Translational Science Institute (SC CTSI) Mentored Research Career Development Award, an SC CTSI grant, and the Office of the Provost at the University of Southern California to B. Sargent. The funders played no role in the design, conduct and reporting of this study.

Results

Participants

Ten infants at high risk of CP and 11 infants with TD participated in the study. All infants participated at 3.5 months CA, except for 2 infants at high risk who initiated the study at 4.5 months CA, within days of discharge from the neonatal intensive care unit. All infants completed the 6-week program, except for 1 infant at high risk of CP who stopped during week 2 because of neurosurgery and 1 infant with TD who stopped during week 5 because of constant rolling. In addition, 2 infants at high risk of CP did not have analyzable data for 1 week because of technology issues for 1 infant and a combination of illness and technology issues for the second infant. One child in each group did not complete data collection for 2 of 30 days because of illness.

Demographic and baseline developmental data for infants at high risk of CP and TD are in Table 1 and Supplementary Table 1, respectively. Locus of the brain injury and follow-up data for infants at high risk of CP is in Table 2. Three infants at high risk of CP were born full-term and 7 were born preterm (range = 23–32 weeks of gestation). Five infants at high risk of CP scored ≥10th percentile on the cognitive, language, and motor subtests of the Bayley-3, and 5 scored <10th percentile on 1 or more subtests. All infants with TD were born full-term and scored ≥10th percentile on the cognitive, language, and motor subtests of the Bayley-3.¹⁷

Table 1.

Demographics and Baseline Developmental Data for Infants at High Risk of Cerebral Palsy^a

Infant	Sex	Race	Ethnicity	Gestational Age (wk)	Birth Weight (kg)	Age at Initiation of Study	Ponderal Index (kg/m 3 )	Bayley-3 Cognitive (Percentile)	Bayley-3 Language (Percentile)	Bayley-3 Motor (Percentile)
1	F	W	H	24	0.75	3 mo 9 d	24.91	16	58	42
2	M	W	H	28	1.11	3 mo 10 d	22.51	2	2	16
3	F	W	H	39	3.42	3 mo 11 d	18.96	37	4	50
4	M	W	NH	32	1.39	3 mo 12 d	20.63	16	42	8
5	M	B	NH	25	0.89	3 mo 15 d	26.18	25	50	88
6	M	W	NH	40	5.22	3 mo 21 d	22.9	37	66	50
7	F	W	H	37	3.20	3 mo 22 d	29.01	16	66	58
8	F	B	NH	24	0.65	3 mo 22 d	NA	90	58	75
9	M	W	NH	29	1.28	4 mo 10 d	21.8	0.4	18	75
10	M	W	H	23	0.55	4 mo 15 d	24.78	1	42	1

^a B = Black or African American; Bayley-3 = Bayley Scales of Infant and Toddler Development, Third Edition; F = female; H = Hispanic/Latino; M = male; NA = not available; NH = not Hispanic/Latino; W = White.

Table 2.

Locus of Brain Injury, Individualized Learning Data, and Follow-up Data for Infants at High Risk of CP^a

Infant	Locus of Brain Injury by MRI	% of Days on Which Infant Met Learning Criteria	Motor Medical Conditions at 2 y	Corrected Age at Walking (mo)	Bayley-3 Cognitive at 2 y (Percentile)	Bayley-3 Language at 2 y (Percentile)	Bayley-3 Motor at 2 y (Percentile)
1	B thalamus L > R; B WMI L > R including L CST	0b	CP with SHc	UAc	1c	0.3c	0.3c
2	B grade IV IVH with WMI	50	CP with SQc	UAc	0.1c	<0.1c	<0.1c
3	B grade III IVH with WMI; reduced volume in caudate, putamen, and thalamus	60	None	12	9	27	16
4	B WMI L > R including B CST; B thalamus	47	CP with SQ	UA	NT	NT	NT
5	WMI R prefrontal and B parietal; injury to R lateral putamen and external capsule	44	Toe walking	15	50	23	16
6	HIE, after cooling; mild WMI with at least 1 punctate WML in L frontal	64	NT	13	NT	NT	NT
7	B WMI in frontal, parietal, occipital, and temporal	60	None	13	37	34	21
8	R WMI in centrum semiovale	46	None	18	9	4	4
9	B cerebellar hemorrhage, R > L; mildly enlarged ventricles with reduced WMV	58	DD	16	0.4	0.3	5
10	B grade III IVH with WMI	21	NA	NA	NA	NA	NA

^a B = bilateral; Bayley-3 = Bayley Scales of Infant and Toddler Development, Third Edition; CP = cerebral palsy; CST = corticospinal tracts; DD = developmental delay; HIE = hypoxic ischemic encephalopathy; IVH = intraventricular hemorrhage; L = left; MRI = magnetic resonance imaging; NA = lost to follow-up; NT = infant unable to be tested because of pandemic; R = right; SH = spastic hemiplegia; SQ = spastic quadriplegia; UA = unable to walk at 2 years of age; WMI = white matter injury; WML = white matter lesion; WMV = white matter volume.

^b Discontinued study after 2 weeks because of surgery.

^c From medical record.

Feasibility

Feasibility was assessed through participation (computer recorded), child’s arousal during the study (assessed via video tapes using masked assessors), and a parent perception form. The mean participation for infants at high risk of CP was 92% (SD = 9%) and for infants with TD was 99% (SD = 2%). The arousal for infants at high risk of CP was alert 93% (SD = 17%), fussy 7% (SD = 16%), and drowsy, sleeping, or crying 0.2% (SD = 1%) of the time. The arousal for infants with TD was alert 92% (SD = 20%), fussy 7% (SD = 19%), and drowsy, sleeping, or crying 0.2% (SD = 1%) of the time. The percentages of sessions that were discontinued because of crying or sleeping for ≥2 minutes averaged 2% (SD = 2%) for infants at high risk of CP and 2% (SD = 3%) for infants with TD. The percentage of sessions that could not be analyzed because of technical reasons averaged 4% (SD = 6%) for infants at high risk of CP and 2% (SD = 3%) for infants with TD. There were no adverse events.

Parent perceptions of the mobile task are in Supplementary Table 2. Nine parents from the group that was at high risk of CP and 10 parents from the TD group completed the survey. For the high risk of CP and TD groups, 89% and 90%, respectively, kept the mobile setup in the home. All families were trained and used a manual in English, except for 1 family in the high risk of CP group who was trained and used a manual in Spanish. In each group, parents’ median (range) rating for setting up and taking down the mobile system on a 5-point scale was 1 (1–3) with 1 as “easy” and 3 as “neutral.” This score included turning on the mobile and computer, aligning the sensors, performing the exposure time for the sensors, putting straps and sensors on the infant, starting the video camera and computer program, and shutting off the system. For the high risk of CP and TD groups, parents reported it required an average of 20 (SD = 10) minutes and 18 (SD = 4) minutes, respectively, to conduct the study each day. For both the high risk of CP and TD groups, parents’ median rating for their child’s affect when playing with the mobile on a 5-point scale was 2 (range = 1–4), with 1 as “happy 81% to 100% of the time,” 2 as “happy 61% to 81% of the time,” and 4 as “fussy 60% to 80% of the time.”

Common changes to the mobile task recommended by parents included shortening the time the infants interacted with the mobile to 5 to 6 minutes because some infants seemed to lose interest toward the end of the 8 minutes, using parent’s voices or pictures as audio or visual stimulation for the mobile, decreasing the size of the system, and simplifying technology to decrease setup time.

Learning

Infants at high risk of CP

The means and SEs of the RLA ratio each week are shown in Figure 2. The main effect of condition was significant for week 1 (F_1,9 = 5.55; P = .04) and week 5 (F_1,8 = 5.81; P = .04) but not for week 2 (F_1,9 = 0.98; P = .35), week 3 (F_1,7 = 3.75; P = .09), week 4 (F_1,8 = 3.21; P = .11), or week 6 (F_1,7 = 2.08; P = .19). The RLA ratio each week by day is in Table 3 and graphed in Supplementary Figures 2–7. Day 5 of week 5 was the only day that the RLA ratio reached statistical significance. The percentage of days each week that individual infants met the learning criteria ranged from 37% to 69% and is in Table 3. The percentage of days over 6 weeks that individual infants met the learning criteria is in Table 2.

Figure 2.

Mean reinforced leg action ratio during baseline and mobile conditions by week for infants at high risk of cerebral palsy. These infants (n = 10) demonstrated learning for 2 of 6 weeks. Error bars are SEs.

Table 3.

RLA Ratio of Infants at HRCP and Infants With TD by Day^a

Week	Group	Interaction Between Condition and Day	Mean Day 1 Baseline RLA Ratio	Mean (SE) Mobile RLA Ratio by Day:				Percentage of Days on Which Individual Infants Met Learning Criteria
				2	3	4	5
1	HRCP	F 4,29 = 2.03 (P = .12)	1	2.23 (0.56)	2.28 (0.56)	1.50 (0.58)	2.18 (0.60)	52
	TD	F 4,39  = 2.6 (P = .05)	1	3.64 (1.32)	5.43 (1.32) b	5.93 (1.32) b	4.90 (1.36)	74
2	HRCP	F 4,26 = 1.30 (P = .29)	1	1.70 (0.33)	1.23 (0.35)	0.91 (0.33)	1.49 (0.48)	37
	TD	F 4,36  = 2.86 (P < .05)	1	4.07 (0.99) b	2.56 (0.99)	3.19 (0.95) b	2.89 (0.95)	56
3	HRCP	F 4,21 = 1.04 (P = .41)	1	2.48 (0.95)	2.66 (0.95)	2.09 (0.99)	2.30 (1.56)	40
	TD	F 4,37  = 5.91 (P < .05)	1	2.66 (0.48) b	2.92 (0.49) b	3.14 (0.47) b	3.26 (0.47) b	81
4	HRCP	F 4,27 = 1.00 (P = .43)	1	1.96 (0.66)	1.85 (0.64)	1.55 (0.64)	2.07 (0.74)	41
	TD	F 4,37  = 3.74 (P < .05)	1	1.96 (0.35) b	2.17 (0.35) b	1.59 (0.35)	1.94 (0.37)	70
5	HRCP	F 4,27  = 2.94 (P < .05)	1	2.06 (0.74)	2.02 (0.73)	2.53 (0.78)	3.69 (0.73) b	68
	TD	F 4,33  = 2.94 (P < .05)	1	3.48 (0.94)	2.48 (0.96)	3.64 (0.99)	2.91 (1.00)	75
6	HRCP	F 4,18 = 0.93 (P = .47)	1	4.08 (2.40)	2.67 (3.02)	8.08 (3.02)	4.05 (3.37)	55
	TD	F 4,34  = 2.88 (P < .05)	1	1.91 (0.36)	2.12 (0.38) b	1.58 (0.36)	2.23 (0.38) b	63

^a Statistically significant learning results are shown in bold type. HRCP = high risk of cerebral palsy; RLA = reinforced leg action; TD = typical development.

^b Adjusted P < .05.

Infants with TD

The means and SEs of the RLA ratio each week are shown in Figure 3. The main effect of condition was significant for week 1 (F_1,10 = 12.8; P < .01), week 2 (F_1,10 = 10.02; P < .02), week 3 (F_1,10 = 31.94; P < .02), week 4 (F_1,10 = 9.67; P < .02), week 5 (F_1,10 = 8.59; P < .02), and week 6 (F_1,9 = 8.7; P < .02). The RLA ratio each week by day is in Table 3 and graphed in Supplementary Figures 2–7. The RLA reached statistical significance on at least 2 days of each week, except for week 5. The percentage of days each week that individual infants met the learning criteria ranged from 56% to 81% and is in Table 3.

Figure 3.

Mean reinforced leg action ratio during baseline and mobile conditions by week for infants with typical development. These infants (n = 11) demonstrated learning each week. Error bars are SEs.

These results can be interpreted as the infants with TD demonstrated more robust learning than the infants at high risk of CP, although individual infants at high risk of CP did demonstrate learning on some days.

Discussion

We hypothesized that the in-home mobile task would be feasible for infants and their families. We believe the in-home mobile task is feasible since participation in the group that was at high risk of CP was >90%, infants were happy and alert >90% of the time when playing with the mobile, and task implementation averaged 20 minutes with most parents rating it as “easy.” Our participation rates are higher than those reported in other studies of infants at high risk of CP using parent-implemented in-home programs.¹⁸ Many factors may have contributed to this: weekly visits by the researcher to support the parents in implementing the task correctly, an engaging play-based task, and the fact that the infants independently played with the kicking-activated mobile. A frequent comment by parents was that it was fun to watch their infant independently play with the mobile and see their learning process.

Many in-home therapeutic tasks for infants at high risk of CP rely on parents to set up an environment and provide some type of handling to support an infant’s participation in the task.^6,7 Infants at high risk of CP may have limited opportunities to independently problem solve a task, which is a common way for infants with TD to learn about the capabilities of their bodies and the effect their motor actions have on their environment. Both types of therapeutic experiences may be necessary to optimize function of infants at high risk of CP, but independent exploration of constructed environments using technology may provide a unique opportunity for infants at high risk of CP to independently learn about their bodies and their bodies’ effect on their environment. This is especially true for infants who will develop more severe motor disabilities because these infants may be at higher risk for “learned nonuse,” in which infants do not use motor actions within their capabilities because they have learned that their motor actions have little to no effect on their environment.¹⁹ Independent exploration of constructed environments using technology may be a means for these infants to learn that their motor actions can have very interesting and important effects on their environment.

We also hypothesized that infants at high risk of CP would learn the association between their leg movements and mobile activation during the last 3 weeks of the study, whereas infants with TD would learn the association each of the 6 weeks of the study. We found that the infants at high risk of CP demonstrated learning during 2 weeks of the study, whereas the infants with TD demonstrated learning during all 6 weeks of the study. The learning results of the infants at high risk of CP are not surprising since learning disabilities and intellectual disability are conditions commonly associated with CP.¹⁴ Discovery learning tasks, such as the one described here, afford a novel way of supporting the learning abilities of young infants. We define a discovery learning task as a task in which infants independently demonstrate a wide range of exploratory actions to generate information about possible outcomes of actions and then independently exploit actions that result in outcomes with adaptive value.¹⁰ In the in-home kicking-activated mobile task, an informal observation was that when first placed under the mobile, infants explored their environment by talking to the mobile, moving their hands toward the mobile, and kicking with their legs, but over time when placed under the mobile, the infants spent a lot more time kicking with their legs and less time moving other parts of their bodies. This discovery learning task afforded the infants a means to independently use the exploration-exploitation process to “learn how to learn.”

This study has 4 limitations. First, the sample size was small, which may have reduced our ability to detect learning for the group of infants at high risk of CP. Second, although all infants in the high risk of CP group were at high risk based on brain imaging results, not all infants in the high risk group developed CP. Third, this study did not investigate potential causes for the reduced amount of learning in the infants at high risk of CP, including attention deficits that may limit the ability to attend to the mobile activation, cognitive impairments that may impede cause and effect learning, tactile and proprioceptive deficits that may constrain the ability to monitor where the leg is in space, and force-generating capacity that may limit antigravity kicking movements. Last, sociodemographic characteristics of the family, such as parent age, education, language, socioeconomic status, and home environment, were not collected. The influence of these factors on infant learning or parental participation with the task may be important to investigate in future studies of the in-home kicking-activated mobile task.

Clinical Relevance

Increasing evidence supports that early therapeutic intervention improves motor skill acquisition of infants at high risk of CP.^6,7 However, there is minimal evidence on ways to maintain and improve the underlying capacities necessary to perform motor skills, for example, the capacities of selective motor control, bone mineral density, muscle flexibility, and force generation. Discovery learning tasks may afford a novel way of motivating infants to engage in the intensive task practice necessary to maintain and improve the underlying capacities necessary for functional motor skills. Specifically, therapeutic environments could be designed such that infants discover and practice targeted capacities, such as selective hip–knee control, as they explore the relation between their leg action and its effects in the constructed environment.

Discovery learning tasks, such as this one, are currently being developed and investigated as a means of supporting the capacity of infants with neuromotor disabilities. A discovery learning task was developed for infants with brachial plexus injury to reinforce biceps activation, a primary impairment of infants with brachial plexus injury.¹⁷ In addition, a discovery learning task was developed to motivate 3- to 6-month-old infants to lift their head in prone²⁰ and another to encourage 4-month-old infants born very preterm to generate more selective hip-knee movement.²¹ Discovery learning tasks have the potential to increase both the learning abilities and specific motor capacities of infants with a wide variety of medical conditions, such as CP, spina bifida, and Down syndrome.

The results of this study support that infants at high risk of CP can learn the in-home kicking-activated mobile task, although at a reduced amount compared with infants with TD. The results also support that the 6-week in-home mobile task was feasible for parents to implement. Further research is necessary to determine whether the in-home kicking-activated mobile task has potential as an intervention to motivate more selective hip-knee control and improve walking outcomes of infants at high risk of CP.

Author Contributions and Acknowledgments

Concept/idea/research design: B. Sargent, L. Fetters

Writing: B. Sargent, L. Fetters

Data collection: B. Sargent

Data analysis: B. Sargent, K.L. Havens, J.L. Wisnowski, M. Kubo, L. Fetters

Project management: B. Sargent

Fund procurement: B. Sargent, L. Fetters

Providing participants: B. Sargent, T-W. Wu

Providing facilities/equipment: B. Sargent

Providing institutional liaisons: B. Sargent

Consultation (including review of manuscript before submitting): K.L. Havens, J.L. Wisnowski, T-W. Wu, M. Kubo

The authors are grateful to Chien-Yen (Kevin) Chang, Kevin Feeley, Robert B. Fuchs, and Rachel Proffitt from the University of Southern California Institute for Creative Technologies for providing technical resources for the development of the in-home kicking-activated mobile system and to Nicole Marcione, Alieh Zamany, Alina Marrone Dancel, and Margaret Ridenhour for assisting with data collection and analysis. Special thanks are due to the parents and infants who participated in the study.

Ethics Approval

This study was approved by the Institutional Review Board at Children’s Hospital Los Angeles (CHLA-15-00354), with ceded review by the University of Southern California.

Funding

This study was supported by the National Institutes of Health (NIH) Eunice Kennedy Shriver National Institute of Child Health and Human Development under award number K12-HD055929 (principal investigator: K. Ottenbacher) to B. Sargent and under award number K23-HD099309 to J.L. Wisnowski. The content is solely the responsibility of the authors and does not necessarily represent the official views of NIH. In addition, the study was funded by the Academy of Pediatric Physical Therapy, a Southern California Clinical and Translational Science Institute (SC CTSI) Mentored Research Career Development Award, an SC CTSI grant, and the Office of the Provost at the University of Southern California to B. Sargent.

Disclosures

The authors completed the ICMJE Form for Disclosure of Potential Conflicts of Interest and reported no conflicts of interest.