Positioning and baby devices impact infant spinal muscle activity

Abstract

Infant positioning in daily life, particularly in relation to active neck and back muscles, may affect spinal development, psychosocial progression, and motor milestone achievement. Yet the impact of infant body position on muscle activity is unknown. The objective of this study was to evaluate neck and back muscle activity of healthy infants in common positions and baby devices. Healthy full-term infants (n=22, 2–6 months) participated in this experimental study. Daily sleep and positioning were reported by caregivers. Cervical paraspinal and erector spinae muscle activity was measured using surface electromyography (EMG) in five positions: lying prone, lying supine, held in-arms, held in a baby carrier, and buckled into a car seat. Mean filtered EMG signal and time that muscles were active were calculated. Paired t-tests were used to compare positions to the prone condition. Caregivers reported that infants spent 12% of daily awake time prone, 43% in supine-lying baby gear, and 44% held in-arms or upright in a baby carrier. Infants exhibited highest erector spinae activity when prone, and lowest cervical paraspinal muscle activity in the car seat. No differences were found between in-arms carrying and babywearing. This first evaluation of the muscle activity of healthy infants supports the importance of prone time in infants’ early spinal development because it promotes neck and back muscle activity. Carrying babies in-arms or in baby carriers may also be beneficial to neck muscle development, while prolonged time spent in car seats or containment devices may be detrimental to spinal development.

Proper development of the pediatric spine and the muscles that support it is important in relation to spinal pathologies, and in preventing spinal injuries in infancy and adolescence. Recent studies indicate the negative role of weak back muscles on spondylolisthesis (Zhu et al., 2016), and the positive role of neck circumference as a predictor of neck strength in children (Lavallee et al., 2013), particularly for injury prevention. Computational and in-vivo murine models demonstrate the positive role of muscle-induced mechanical loads in shaping developing fetal bones (Sharir et al., 2011). While the exact role of spinal muscle use in the developing spine is unknown, a recent study by Rolfe et. al. found a relationship between altered muscle forces and abnormal spinal development (hyperkyphosis and vertebral malformations) in paralyzed fetal chicks (Rolfe et al., 2017). Sustained static muscle loading was highly detrimental to early spine development, and paralysis at earlier developmental stages resulted in more severe spinal deformations. If muscle activity plays a critical role in the developing fetal skeleton, it is likely that active muscles also contribute to neonatal spinal development, especially in the early months of life when significant changes in spinal shape are exhibited.

The American Academy of Pediatrics (AAP) launched the Back-to-Sleep campaign in 1992 as a response to the high rate of Sudden Infant Death Syndrome (SIDS) in the United States, recommending that healthy infants be positioned on their back for sleep (AAP Task Force on Infant Positioning and SIDS, 1992). Recommendations against prone sleep position to prevent SIDS resulted in babies spending more time in the supine position, even during waking hours (Dudek-Shriber and Zelazny, 2007), which coincided with increased incidence of deformational plagiocephaly (Branch et al., 2015; Kane et al., 1996), and delayed motor development in infants who slept supine (Davis et al., 1998; Majnemer and Barr, 2006, 2005). Supine lying early in life has also been linked to idiopathic infantile scoliosis (Mau, 1981). Conversely, prone positioning during infants’ awake time correlates to proper motor development (Dudek-Shriber and Zelazny, 2007; Kuo et al., 2008), extensor muscle development (Russell DC, Kriel H, 2009), and cranial asymmetry prevention (Zachry et al., 2017). To help infants develop neck and head control, pediatricians advise parents to initiate supervised tummy time (i.e. spending time in the prone position) 2–3 times daily for 3–5 minutes (American Academy of Pediatrics, 2011). However, many parents are not aware of (Koren et al., 2010) or do not adhere to (Zachry and Kitzmann, 2011) this prone play guideline, resulting in limited opportunity for critical musculoskeletal development for infants.

Aside from awareness and implementation of prone play, parents are faced with several commercial baby gear options that often keep infants lying supine and buckled into containment devices such as car seats, strollers, swings, or baby seats. Despite their abundance in the current infant care environment, limited research has been conducted on the biomechanical impact of placing infants in supine-lying commercial products. However, excessive time in car seats has been linked to increased deformational plagiocephaly (Littlefield et al., 2003), and decreased oxygen saturation levels (Cerar et al., 2009), and leg movement (Jiang et al., 2016). Upright-carry gear includes carriers for babywearing (i.e. carrying infants on caregivers’ bodies), which offers many emotional, physiological, and physical benefits to the infant. Emotionally, babies carried by their mothers exhibit decreased crying (Hunziker and Barr, 1986), increased attachment (Anisfeld et al., 1990), and responsive breastfeeding (Little et al., 2018). Inward facing skin-to-skin contact of newborns (“kangaroo care”) through in-arms holding or aided by babywearing decreases infant mortality rates (Boundy et al., 2016), increases breastfeeding success and duration (Pisacane et al., 2012), reduces agitation (McCain et al., 2005) and sleep apnea (Messmer et al., 1997), and improves cardiorespiratory (Chi Luong et al., 2016) and temperature regulation (Christensson et al., 1992). Upright-carrying may also prevent ear infections and alleviate gastroesophageal reflux disease (Schön and Silvén, 2007).

Despite the growing body of evidence indicating that infant positioning in daily life and in commercially-available devices (including upright caregiver-worn baby carriers and supine lying containment carrying devices) may affect spinal development, psychosocial progression, and motor milestone achievement, no research has explored these effects from a biomechanics standpoint. The objective of this research was to evaluate the spinal muscle activity of healthy infants while: prone, supine, carried by a caregiver in-arms, held in an inward-facing soft-structured baby carrier, and buckled in a standard infant car seat. The primary hypothesis was that infants would exhibit the highest neck and back muscle activity during prone lying. The two secondary hypotheses were that muscle activity would be different (i) between the in-arms and baby carrier conditions, and (ii) between the car seat compared to supine lying.

Materials and Methods

Twenty-two infants (2–6 months) participated in this study, approved by the Institutional Review Board of the University of Arkansas for Medical Sciences. Infants were born full-term (>37 weeks of gestation), of normal birth weight (>2,500 g), between 5–95^th percentile in height and weight (CDC), and at least 1-week past vaccination and/or sickness. Participants were recruited from flyers and by word of mouth. Caregivers of the participants provided informed assent for infants to participate, signed a photo release, and completed a survey about infants’ demographics and features of daily life (feeding and sleeping tendencies).

Infants participated in five 30-second positional tasks (Figure 1): lying prone on a mat, lying supine on a mat, held in-arms by caregiver, held in a soft-structured inward-facing baby carrier (Boba, Inc.; Boulder, CO), and buckled in a standard 5-point harness car seat (Evenflo Inc.; Miamisburg, OH). Task order was randomized, and the data was considered usable if the infants were awake, and not crying. Between tasks, infant participants took a 60-second break. However, if they were agitated or tired, the break time was adjusted to accommodate comfort.

Figure 1.

(A) Infant fitted with bilateral electromyography (EMG) electrodes in (B) Prone position, (C) Caregiver’s arms, (D) Car seat, (E) Inward facing soft-structured baby carrier, and (F) Supine position. Motion capture sensors visible on the infants’ legs and diaper are part of a larger study, and not relevant to the current study.

During these tasks, surface electromyography (EMG) electrodes (Delsys Inc., Natick, MA) recorded the muscle activity (1000 Hz) of bilateral cervical paraspinal (semispinalis capitis and splenius capitis) and erector spinae (longissimus dorsi and spinalis dorsi) muscles. The cervical paraspinal and erector spinae muscle groups were specifically chosen for measurement since they are the prime muscle groups responsible for neck and trunk extension, and play an important role in head control, a major milestone for infants under 6 months of age (Bly, 1994). The infants’ torsos were wrapped with soft self-adherent tape to contain EMG sensor wires. Raw EMG waveforms were assessed for corrupted or missing data using power spectral analysis (Boxtel, 2001), and such data was excluded from analysis.

The raw EMG waveforms were band-pass filtered using a 4^th-order Butterworth filter between 35–500 Hz, to reduce contamination from movement artifacts, electrocardiogram signals (Drake and Callaghan, 2006), and high-frequency noise (Hermens et al., 1999). Additionally, EMG waveforms were notch-filtered at 60 Hz using a 4^th-order Butterworth filter to eliminate the effects of interference from nearby electronic sources. EMG waveforms were then full-wave rectified, demeaned, and subjected to a low-pass 4^th order Butterworth filter with a cutoff frequency of 50 Hz to obtain the EMG envelope (Hodges and Bui, 1996). The filtered and rectified EMG data was normalized to the prone condition using custom MATLAB code (The MathWorks, Natick, MA). Mean EMG signal and the percent time that muscles were active above the prone threshold (two standard deviations above the mean EMG signal of prone) were calculated.

Due to the paired samples design in the intended analysis of the collected data, the differences between the EMG outcome variables at the prone position and all other positions were evaluated for normality. Since most of these differences did not demonstrate normality on the Shapiro-Wilk test (i.e. p<0.05), the Friedman test was conducted. The dependent variables in the EMG data were the mean EMG signal and the percent time the muscles were active above the prone threshold. The independent variable was infant position, with five groups: prone, in-arms, car seat, baby carrier, and supine. If the Friedman test was significant, post-hoc pairwise Wilcoxon signed rank tests were used to compare the EMG data from all positions to the prone position with a Bonferroni correction (p<0.0125). For each comparison, the effect size (d) was calculated as the Z-statistic divided by the square root of the sum of the number of observations in the two compared positions (Pallant, 2007). Effect size was considered small if 0.1<d<0.3, medium if 0.3<d<0.5, and large if d>0.5 (Cohen J., 1988). Additional post-hoc pairwise comparisons (to evaluate the secondary hypotheses) were conducted for all EMG variables of the in-arms vs. the baby carrier conditions, and the car seat vs. the supine conditions. SPSS (IBM Corp., Armonk, NY) was used for all statistical analyses.

Results

Participants were well-distributed in age between 2–6 months, with a mean of 4.2 months (13 M, 9 F). Key information from the caregiver completed survey are presented in Table 1. Recorded and reported measurements of participants’ age, length, weight, and daily awake time positioning frequency are presented in Table 2.

Table 1.

Participant demographics, characteristics, and features of daily infant life

Gender
Male	13
Female	9
Ethnicity
White	18
Black or African American	2
Hispanic or Latino	1
Asian	1
Infant Sleeping Positions
Prone	3
Supine	11
Supine & Prone	6
Side	2
Infant Sleeping Arrangements at Night
Crib in Separate Room	10
Crib in Caregiver Room	7
In Caregiver Bed	3
Other	2
Infant Care Arrangements
Guardian at Home: Full-time	14
Guardian at Home: Part-time	2
Full-time Daycare	4
Other Caregiver at Home: Full-time	2
Feeding
Exclusively Breastfed	14
Breastfed and Formula Fed	4
Exclusively Formula Fed	4
Started Solids?
Yes	11
No	11

Table 2.

Participant measurements and daily positioning frequency: Mean ± SD, (Range)

Measurements
Age at testing (months)	4.2 ± 1.6 (2.0 – 6.9)
Length at testing (cm)	60.7 ± 5.1 (50.8 – 71.1)
Weight at testing (kg)	6.6 ± 1.5 (3.6 – 10.0)
Length at birth (cm)	50.8 ± 3.0 (45.7 – 55.9)
Weight at birth (kg)	3.5 ± 0.5 (2.5 – 4.2)
Awake time positioning
Caregiver’s arms/lap (%)	34.4 ± 15.2 (15.0 – 80.0)
Swing/ Car seat/ Stroller (%)	15.4 ± 9.6 (0.0 – 30.0)
Tummy (Prone) (%)	12.4 ± 10.7 (0.0 – 30.0)
Back (Supine) (%)	27.7 ± 16.4 (0.0 – 60.0)
Baby carrier (%)	9.1 ± 11.6 (0.0 – 50.0)
Other (%)	1.0 ± 3.0 (0.0 – 10.0)

Usable EMG data, after assessment for corrupted or missing data, was extracted for 19 participants in the prone position, 19 in-arms, 18 in the car seat, 17 in the baby carrier, and 18 in the supine position. Infants exhibited significantly higher erector spinae activity during prone positioning compared to all other tasks (p<0.0125), both in mean muscle activity and in percentage of active muscle time. Cervical paraspinal activity varied among positions. However, infants exhibited significantly lower cervical paraspinal activity in the car seat, when compared to the prone position (p<0.0125), in both mean muscle activity and percentage of active muscle time (Figure 2, Table 3).

Figure 2.

(Top) Mean erector spinae and cervical paraspinal muscle activity (%) at the different positions, normalized to the prone position, and (Bottom) Percentage of time that the erector spinae and cervical paraspinals are active above threshold (mean + 2 S.D. of prone position) for the different positions; Error bars are 95% CI; *p < 0.0125 and ^‡p < 0.05 vs. prone.

Table 3.

Electromyography (EMG) variables at the different positions and statistical analyses (Friedman test reported in line with the dependent variable; Pairwise comparisons in table).

Variable	Mean (95% Confidence Interval)	Asymptotic Significance (p-value)	Z-statistic	Effect size ($Z/\sqrt{N}$)
Mean muscle activity (mV) of Erector Spinae, χ2(4) =29.45, p < 0.001
Prone	21.52 (17.45 – 25.60)	-	-	-
In-arms	8.53 (6.54 – 10.94)	< 0.001	−3.910	0.63
Car seat	6.01 (3.73 – 8.29)	< 0.001	−3.920	0.65
Baby carrier	8.47 (4.92 – 11.42)	< 0.001	−3.622	0.62
Supine	8.35 (5.45 – 10.69)	< 0.001	−3.845	0.64
Mean muscle activity (mV) of Cervical Paraspinal, χ2(4) =17.44, p = 0.002
Prone	123.95 (47.32 – 200.57)	-	-	-
In-arms	108.53 (43.41 – 201.71)	0.17	−1.368	0.22
Car seat	69.61 (5.03 – 141.94)	< 0.001	−3.506	0.58
Baby carrier	105.85 (29.58 – 175.02)	0.10	−1.633	0.28
Supine	109.82 (38.42 – 191.31)	0.02	−2.417	0.40
Time (%) above Prone threshold for Erector Spinae, χ2(4) =30.87, p < 0.001
Prone	4.44 (3.78 – 4.69)	-	-	-
In-arms	1.14 (0.00 – 2.17)	0.001	−3.215	0.52
Car seat	0.39 (0.12 – 0.61)	< 0.001	−4.015	0.67
Baby carrier	1.19 (0.08 – 2.19)	< 0.001	−3.563	0.61
Supine	1.69 (0.40 – 2.82)	0.004	−2.902	0.48
Time (%) above Prone threshold for Cervical Paraspinal, χ2(4) =20.14, p < 0.001
Prone	3.87 (2.61 – 4.07)	-	-	-
In-arms	3.29 (1.23 – 4.76)	0.69	−0.402	0.07
Car seat	0.52 (0.00 – 0.90)	< 0.001	−3.549	0.59
Baby carrier	3.38 (0.38 – 5.78)	0.15	−1.449	0.25
Supine	2.88 (0.69 – 4.55)	0.14	−1.489	0.25

Post-hoc comparisons of both EMG variables for the erector spinae and cervical paraspinals did not show statistically significant differences between the in-arms and the baby carrier conditions, or between the car seat and supine conditions. However, small to medium effect sizes were observed between the car seat and supine conditions in mean EMG activity of the erector spinae (Z=−1.449, p=0.147, d=0.23) and cervical paraspinals (Z=−2.059, p=0.039, d=0.34), and in the percentage of active muscle time for the erector spinae (Z=−1.811, p=0.070, d=0.28) and cervical paraspinals (Z=−2.272, p=0.023, d=0.37).

Discussion

The purpose of the study was to evaluate spinal muscle activity in healthy infants while prone, supine, carried by a caregiver in-arms, held in an inward-facing soft-structured baby carrier, and buckled in a standard infant car seat. The primary hypothesis was that infants would exhibit the highest neck and back muscle activity during prone lying. The two secondary hypotheses were that neck and back muscle activity would be different between (i) the in-arms and baby carrier conditions, and (ii) the car seat compared to supine lying.

The EMG characteristics of the erector spinae and the cervical paraspinal muscles highlight the importance of prone time in the development of the thoracic, lumbar, and cervical spine in early infancy. Our primary hypothesis was accepted for the EMG results from the erector spinae but rejected partially for those from the cervical paraspinals. While all infants exhibited significantly lower erector spinae activity compared to prone for all conditions in our study, the EMG values for the in-arms, baby carrier, and supine conditions were similar in magnitude across all conditions. In contrast, the erector spinae EMG activity in the car seat was much lower, exhibiting the highest effect size in its difference to the prone condition. Cervical paraspinal EMG activity, while not significantly different from the prone condition for the in-arms, baby carrier and supine conditions, was also significantly lower in the car seat condition both in mean muscle activity (two times lower) and active muscle time (seven times lower). Additionally, specific pairwise comparisons between the car seat and supine conditions indicated that infants exhibited lower mean muscle activity and active muscle time in the car seat. The effect sizes for the differences between car seat and supine were medium, indicating clinical significance, though statistical significance was not reached (Kazis et al., 1989).

These are particularly impactful findings, especially in the context of the study by Rolfe et. al. linking prolonged fetal immobilization to spinal deformity (Rolfe et al., 2017). Furthermore, in conjunction with the recent evidence correlating car seat usage with high plagiocephaly incidence (Littlefield et al., 2003), it is likely that car seats, while critical to safely transport infants in automobiles, may negatively affect spine development in young infants if they spend prolonged periods of time buckled into them. Additionally, other infant gear with similar positioning to car seats (rockers, swings, strollers, etc.), are cause for concern if infants spend too much time in these devices.

Babywearing in an inward-facing soft-structured baby carrier exhibited a non-significant difference with small effect sizes in the EMG activity of the erector spinae and cervical paraspinals, when compared to carrying babies in-arms. Observationally, the position in which infants were carried in-arms varied greatly, with some caregivers supporting the natural C-shape of the infant spine and some who did not. In contrast, the infant position in the inward-facing soft structured baby carrier was consistently supportive of the infant’s spine. Our findings indicate that there are no attributable differences in the neck and back muscle activity of awake infants who are carried in-arms versus in an inward-facing soft-structured baby carrier, and that appropriate babywearing has the potential to consistently support infants’ spines, making babywearing a convenient addendum to infants’ daily positioning without compromising the biomechanical and musculoskeletal benefits attained from in-arms holding and prone time.

Sleeping positions of the infants highlighted a curious story, with 40.9% (9/22) infants sleeping either prone or in a combination of supine and prone, despite the AAP’s Back-to-Sleep recommendations. Awake time positioning indicated that the infants in our study received, on average, only 12.4% prone time, compared to 43.1% time lying supine on the floor or in a supine-lying containment device (15.4% in swings, car seats and strollers, and 27.7% lying on their back), and 43.5% time in active carry (34.4% in caregiver’s arms or lap, and 9.1% in baby carrier). The 12.4% daily prone time in infants’ awake time, especially since 54.5% (12/22) infants reported 10% or lower daily prone time, is concerning since associations between insufficient daily prone time and delayed motor development (Dudek-Shriber and Zelazny, 2007; Kuo et al., 2008), cranial asymmetry (Zachry et al., 2017), and extensor muscle development (Russell DC, Kriel H, 2009) have been consistently reported in developmental literature. Our survey findings are comparable to a recent survey study by Franchak et al. who quantified the frequency of daily infant positioning during awake time in 95 infants <1 year (Franchak, 2018). Three-month old infants spent 2.9% of their awake time prone, 42.7% lying supine and in supine/reclined containment (18.3% lying supine, 24.4% reclined in a car seat or swing), and 49.8% in active carry (in-arms, on lap, or in a baby carrier). Six-month old infants spent 9.2% of their awake time prone, 35.8% lying supine and in supine/reclined containment (16.6% lying supine, 19.2% reclined in a car seat or swing), and 34.0% in active carry.

The study was not without its limitations. The infants enrolled in this study were recruited from the same geographic location within a 20-mile radius of Little Rock, Arkansas. The survey questionnaire had a small sample size for survey data and relied on caregiver-reported outcomes, which may have low reliability and bias (Hnatiuk et al., 2013). In the future, inertial measurement units may be used to accurately quantify daily infant positioning. The EMG data for each task had a duration of 30-seconds (isolated from 60-second collections), and may not be ideally representative of real-life scenarios, particularly prolonged exposure (both in awake and asleep time) to the positions or products evaluated in this study. We recorded muscle activity using surface EMG electrodes, which have been demonstrated to adequately represent agonistic muscle group activity when compared to fine-wire indwelling electrodes (Perry et al., 1981). While the Delsys Trigno EMG sensors used in this study were designed for use on adult humans, previous usage has been documented in infants of a similar age range (Price et al., 2018). Additionally, the Delsys Trigno Mini EMG sensors used on infants in this study were designed to collect data from small muscles. Infant compliance was a challenge in this study, so data was only included in the analysis if the infant was awake, and not crying. While the muscle activity of the spine is critical to understand in the context of spinal development, infant positioning likely impacts the entire musculoskeletal system. Future studies should seek to find how other muscles and joints may be impacted by infant devices or positions.

Our novel biomechanical study evaluated infant spinal muscle activity in common positions and in upright and supine-lying baby devices. Overall, our findings indicate the crucial role of prone positioning in the muscle activity of the back and neck, the utility of babywearing in an inward-facing soft-structured baby carrier for neck muscle development, and the harmful implications of prolonged time spent by infants in a car seat. The caregiver-reported survey indicated that infants spent little daily awake time in the prone position, and a large amount of time in passive carry or containment positions. Our findings are particularly impactful for caregivers whose infants spend an increasing amount of time in commercially available baby gear.