The contractile patterns, anatomy and physiology of the hyoid musculature change longitudinally through infancy

Abstract

All mammalian infants suckle, a fundamentally different process than drinking in adults. Infant mammal oropharyngeal anatomy is also anteroposteriorly compressed and becomes more elongate postnatally. While suckling and drinking require different patterns of muscle use and kinematics, little insight exists into how the neuromotor and anatomical systems change through the time that infants suckle. We measured the orientation, activity and contractile patterns of five muscles active during infant feeding from early infancy until weaning using a pig model. Muscles not aligned with the long axis of the body became less mediolaterally orientated with age. However, the timing of activation and the contractile patterns of those muscles exhibited little change, although variation was larger in younger infants than older infants. At both ages, there were differences in contractile patterns within muscles active during both sucking and swallowing, as well as variation among muscles during swallowing. The changes in anatomy, coupled with less variation closer to weaning and little change in muscle firing and shortening patterns suggest that the neuromotor system may be optimized to transition to solid foods. The lesser consequences of aspiration during feeding on an all-liquid diet may not necessitate the evolution of variation in neuromotor function through infancy.

1. Introduction

Suckling evolved at or near the origin of mammals [1]. Coupled with maternal lactation, suckling enhances infant survival by enabling infants to acquire a ready source of tailor-made nutrients. The general mechanism of suckling is similar across therian mammals, whereby intra-oral suction is generated to transfer milk from the teat into the oral cavity [2]. The process of infant feeding requires a fundamentally different suite of behaviours, neuromotor patterns and anatomical configurations than drinking in adults [3], and growth patterns of the craniomandibular muscles, tendons, and bones contribute significantly to changes in morphology and oromotor function as mammals grow [4,5]. In particular, infant mammals are characterized by being anteroposteriorly compressed in their oropharyngeal anatomy, with craniofacial elongation occurring as they mature postnatally [6–8], and patterns of muscle activity and jaw movements change during the weaning process [4,9–11]. Although we have a general appreciation for how the musculoskeletal system functions during feeding before and after weaning, we lack insight into how it develops while the infant is still feeding on milk prior to transitioning to solid food.

Understanding how infants feed throughout the entire duration of suckling is critical, because despite the anatomical and neuromotor changes that occur through infancy, all mammals must suck, and continue to do so as they approach an age at which they would wean. During this time, infants are performing the same behaviours (sucking and swallowing) with the same structures and muscles. However, they are doing so under vastly different anatomic conditions as they grow, their craniofacial anatomy elongates and their neuromotor system matures [8,12]. This is especially true in the context of changes in the configurations and activity patterns of oropharyngeal and laryngeal muscles powering feeding, many of which are aligned oblique to the long axis of the body.

One component of muscle function is the angle at which it operates relative to a joint [13,14]. Because the muscles involved in swallowing to move the hyoid bone do not cross a bony joint, their orientation from the hyoid relative to the midline of the animal gives an indication of their function. However, muscle function is also determined by dynamic components [15,16], including the timing of muscle activity [17] and the shortening patterns of the muscle during activity [18]. While we know the general patterns of activity in much of the musculature driving feeding in infants [19,20], the contractile patterns of muscles have only been examined in two of over 25 paired muscles involved, and only at single age points [21,22]. Therefore, although we understand the general mechanics of how infants suckle, and that infant oropharyngeal and craniofacial anatomy elongate postnatally, we have a poor understanding of how muscle firing and shortening patterns interact with changes in anatomy throughout the nursing period.

We used a pig animal model to study the changes in orientation of five of the muscles powering infant feeding and swallowing through infancy, while simultaneously measuring muscle activation using fine-wire bipolar electromyography (EMG) and measuring length changes using fluoromicrometry [18]. Infant pigs are representative of a generalized mammalian infant oropharyngeal morphology and are also a validated model for understanding infant human feeding physiology [23]. We included two muscles that are parallel to the long axis of the body (geniohyoid and thyrohyoid), and three that run oblique to the long axis of the body (digastric, stylohyoid and omohyoid). We expected that the line of action of the geniohyoid and thyrohyoid would change minimally as infants grew, but that the line of action for the digastric, stylohyoid and omohyoid would become less mediolateral with age as their orientation changes with the pig's elongating craniofacial anatomy. Although we expect muscles to exhibit some combination of concentric, eccentric and isomeric contractions, very little is understood about the contractile patterns of the hyoid musculature during infant feeding. This results in a lack of a foundation for developing directional expectations of muscle functions during infant feeding. However, we predicted that muscles which are active during both sucking and swallowing (digastric, geniohyoid and stylohyoid) would have distinct patterns of activity and shortening for each behaviour.

2. Methods

(a). Animal housing and care

Infant pigs (Yorkshire/Landrace sows, Shoup Investments Ltd., Orrville, OH) were obtained at 1 day of age and trained to drink infant pig formula (Solustart Pig Milk Replacement, Land o’ Lakes, Arden Mills, MN) via a bottle and modified nipple (NASCO Farm & Ranch, Fort Atkinson, WI). Infant pig care followed standard procedure, including the administration of Excede as an antibiotic [12,24].

(b). Surgical procedures

Throughout the duration of the experiment, infant pigs underwent three separate procedures. At 5 days of age, pigs were anaesthetized using isoflurane anaesthesia. A custom bead injector needle was used to place 0.8 mm tantalum markers (X-Medics, Frederiksberg, Denmark) into the subdermal space of the dorsal surface of the snout, in the hard palate, tongue midline (anterior, middle and posterior locations), soft palate and palatopharyngeal arches, and a radiopaque hemoclip was placed on the anterior border of the epiglottis.

Between 6 and 8 days of age, pigs underwent the first of two sterile surgeries. In this surgery, an incision was made slightly to the left of the midline from the angle of the jaw anteriorly, to the posterior margin of the thyroid eminence. A bead was sutured between the bellies of the sternohyoid muscle at the location of their insertion to the hyoid, and another bead was sutured to the fascia overlying the thyroid cartilage to facilitate tracking of their movements in X-ray video. Following this, muscles for EMG and fluoromicrometry implants were identified. We placed beads and electrodes in a total of five muscles: geniohyoid, digastric, stylohyoid, thyrohyoid and omohyoid. Each pig received implants for three muscles, resulting in each muscle being recorded in three pigs (n = 5 pigs total). These muscles are of interest because they include muscles anterior to the hyoid that are involved in suction (digastric and geniohyoid) as well as muscles that are traditionally thought to be active only during swallowing (stylohyoid, thyrohyoid and omohyoid). Additionally, some of these muscles run parallel to the midline (geniohyoid and thyrohyoid), whereas others run at an angle that may change through infancy (digastric, stylohyoid and omohyoid). Depending on muscle visibility, beads were injected with a bead injector needle towards either end of the muscle, following standard fluoromicrometry procedures (figure 1) [18]. Following bead placement, two bipolar electrodes were inserted into the bellies of each muscle near the location of bead insertion (figure 1). Electrodes were sutured into connective tissue to provide an extra level of security. Electrodes were connected to a microconnector, and all exited the posterior margin of the incision, which was sutured following implantation of all electrodes. A standard 25-pin D-connector was connected to microconnectors outside of the body, and cables were secured with Vetwrap to minimize the risk of electrodes getting disconnected or causing animal injury. Prior to surgery, Buprenorphine and Meloxicam were administered subcutaneously as an analgesic and to reduce swelling, and Bupivacaine was administered post-operatively at the site of the incision as an analgesic.

Figure 1.

Muscle orientations for newborn (a) and older, pre-weaning (b) pigs in lateral (a,b) and ventral (c) views relative to a newborn skull. In (c), the newborn musculature is shown on the top half of the image, and the musculature of the older pig is on the bottom. White circles in (c) indicate location of bead placement for fluoromicrometry. White bolt symbols indicate location of electrodes for electromyography, and the medial ends of the hyoid bone are indicated in brown. Green, geniohyoid; blue, digastric; yellow, stylohyoid; red, omohyoid; purple, thyrohyoid. (Online version in colour.)

As infant pigs are very active, EMG wires implanted between 6 and 8 days old were functional for up to 5 days before failing. Thus, to obtain data at an older age, we placed new beads and electrodes on the right side of the body during a sterile surgery at an age of between 20 and 22 days. Surgical procedures follow those outlined above, albeit without the suturing of the hyoid and thyroid bead at the midline. Seven days of age is the earliest age where pigs are able to maintain stable body temperatures and be recorded (equivalent of a 1- to 2-month-old human infant), and 21 days of age is the approximate age where pigs transition to a solid diet (equivalent to a 10- to 12-month-old human infant) [11,25]. These age ranges thus represent the earliest and latest possible time points to record suckling in pigs.

(c). Data collection

We collected two modalities of data shortly after both EMG surgeries: videofluoroscopy and EMG, which were synchronized via a digital trigger signal.

Video data were collected using a high-speed camera (XC1 M, XCitex, Cambridge, MA) at 100 Hz during fluoroscopic exposure (GE 9400C-Arm, 70 kV, 5 mA) using standard XROMM data collection procedures [26]. Pigs were placed in a radiolucent box and fed from a bottle and nipple ad libitum. To better visualize milk during feeding, barium was mixed with infant formula. We recorded at least 20 swallows per pig per age, starting after the first 10 s of feeding, which occurs at a faster rate than normal feeding [27]. All EMG signals were amplified and recorded at 10 kHZ on a 16 channel Powerlab (30/16, AD Instruments, Colorado Springs, CO).

Anatomical data were collected using DiceCT [28]. We scanned two pigs of different ages: (i) an infant pig that was 1 day old at the time of death (hereafter, newborn) and (ii) an infant pig that was 21 days old at the time of death (hereafter, older pig). We scanned the newborn pig prior to staining to associate bone models with muscles during reconstruction. The older pig could not be scanned prior, as much of the cranium and the right half of the jaw and skull had to be bisected (slightly off axis to the midline and when the pig was frozen to ensure muscles of interest were not disturbed) to fit into the scanner. Both pigs were initially perfused using a 5% I2Ki solution through the common carotid artery to decrease diffusing time. Pigs were scanned to assess the breadth of perfusion, and were placed in a bath of 10% formalin for 14 days to fix the tissue. The newborn pig was fully perfused and scanned after fixation (35 µm, 70kVp, 22 µA). The older pig was placed in a 2.5% I2Ki solution following fixation, which was refreshed every 14 days along with a computerized tomography scan until the scan showed complete and even staining throughout the tissue.

(d). Data processing

The beginning of the swallow was identified by determining the frame where the bolus was accumulated in the valleculae prior to passing the epiglottis following published procedures [24,29,30]. The end of the swallow was identified as the frame that the epiglottis returned to a resting position. The beginning of suck was identified as the frame where the middle tongue made contact with the hard palate, and ended the frame prior to the next contact. Muscle on/off times were processed in PowerLab to calculate the time of onset and offset of muscle activity for sucks and swallows.

Kinematic data were processed in lateral and dorsoventral views using XMALab [31] using a combination of manual tracking and automatic tracking from DeepLabCut [32,33], which was then manually checked and corrected as necessary. Markers inserted into the nose and hard palate had intermarker distance (IMD) standard deviations of less than 0.03 mm and were assigned as a rigid body. Fluoromicrometry data exported from XMALab were processed using a custom Matlab script which calculated the IMD between two markers and synchronized it with muscle activity timing for each swallow. For each set of swallows, we exported raw IMD during muscle activity, as well as IMD interpolated to 101 frames to account for differences in length of muscle activation for both the entire swallow, as well as for only when the muscle was active. For muscles active during sucks, we performed the same analysis, but synchronized muscle length changes with suck timing. For each muscle during each activity, muscle length was set at zero at the frame the muscle began firing, and subsequent lengths reflected deviations from that initial length. We report IMDs from the side EMG data was collected from at both ages, although IMD is generally synchronous between sides, as previous work examining muscle firing patterns has demonstrated (electronic supplementary material, figure S1) [34].

DiceCT data were segmented using Avizo 9.4 (FEI Visualizations Science Group; Hillsboro, OR). Muscle fibres and endomysium were defined as high-density material and were surrounded by lower density material (perimysium), which was below the minimum grey-scale value threshold to be included in segmented volumes. We segmented the entire muscle for each muscle we recorded EMG data on. On the newborn infant, we segmented muscles on the right side of the body, and we segmented muscles on the left side of the body for the older infant so that they could be superimposed for angle measurements. Segmented muscles were exported as individual .objs and were imported into Autodesk Maya (Autodesk Inc., San Rafael, CA, USA), along with models of the skull, hyoid and left and right mandibles from the newborn infant. Each set of muscles (newborn or older piglet) were parented to each other and were then rotated and translated into position to be aligned with the skull. To do so, we positioned the hyoid between all segmented muscles (except digastric, which in pigs does not have a posterior belly and attaches directly to the mastoid process of the skull), with geniohyoid running parallel to the long axis of the body. To measure muscle angles, we created a set of axes with the anteroposterior (AP) axis aligned to be parallel to the long axis of the body, with the dorsoventral (DV) axis set as being perpendicular to the hard palate and the mediolateral (ML) axis being orthogonal to those two (electronic supplementary material, Movie 1). As AP and ML angles are perpendicular to one another and a given vector could be represented as angular deviation relative to either axis, here we focus on ML angles. We then translated the axis to the location of the insertion of each muscle at the hyoid and measured the two-dimensional angle for each axis direction relative to the orientation of the muscle vector.

(e). Statistical analyses

We used mixed effects models to test for the effect of age on the timing of muscle onset and offset relative to the beginning of the swallow, with individual as a random effect (variable ∼ age + (1|ID), lme4), [35]. We also performed Cohen's D analyses to estimate effects size [36]. We used Levene's tests for equality of variance to determine if age had an impact on variation in muscle onset and offset relative to the beginning of the swallow. We did not perform statistical analyses on the timing of muscles relative to sucks owing to regional heterogeneity in muscle activation relative to suck timing [21]. As different individuals had slight differences in the exact location of fluoromicrometry beads and electrode placement, we analysed all length change data on an individual basis. We did not compute statistics for muscle orientations between the two ages, as we had only one measure per age. Data used in statistical analyses are available on figshare [37].

3. Results

(a). Anatomy

We found substantive changes in muscle orientation through ontogeny. These changes were most prominent in the digastric, stylohyoid and omohyoid. We found little difference in muscle orientation for the geniohyoid and thyrohyoid, which are oriented parallel to the long axis of the body. Digastric, stylohyoid and omohyoid all exhibited decreased mediolateral angles throughout ontogeny, with the mediolateral angle in older pigs being approximately half that in younger infants (figure 1 and table 1). Similarly, dorsoventral angles decreased with age in digastric, stylohyoid and omohyoid (figure 1 and table 1).

Table 1.

Muscle orientations. (Bold values indicate muscles which exhibited approximately half the ML angle in older pigs, with a similar increase in AP angle, indicating muscles which became oriented more parallel to the long axis of the body through ontogeny. Italicized values indicate muscles in which the DV angle decreased by over 10 degrees.)

	newborn			weaning
	AP	ML	DV	AP	ML	DV
geniohyoid	91.93	−1.93	−17.1	91.06	−1.06	−11.9
digastric	70.47	19.53	22.45	86.01	3.99	13.35
stylohyoid	40.11	49.89	34.58	66.78	24.78	25.34
thyrohyoid	83.94	6.06	3.11	91.51	−1.51	3.52
omohyoid	70.27	19.73	18.67	79.6	10.4	2.78

(b). Muscle timing through infancy

Owing to electrode failure in some animals, we did not acquire data from all individuals for all instrumented muscles at both ages, but report all data collected and processed for completeness (figure 3). The timing of muscle activity during sucking in the geniohyoid was complex and indicated extensive regional heterogeneity, as previous work has identified [21]. We also observed EMG activity in the stylohyoid during sucking from electrodes placed in the ventral, but not dorsal, portion of the muscle (figure 3; electronic supplementary material, figure S2), in addition to its standard activity at the initiation of the swallow (electronic supplementary material, table S2).

Figure 3.

Interpolated muscle length changes during activity while swallowing (orange) and sucking (blue) in 7-day-old (left) and 21-day-old (right) pigs. Muscle length changes are generally consistent through ontogeny, but differ between sucking and swallowing. Mean standard deviation through the cycle per pig is provided in blue (sucking) or orange (swallowing), highlighting that for most muscles, standard deviation of length changes is less in older pigs than in younger pigs. (Online version in colour.)

Swallow duration (young = 0.21 ± 0.02 s, old = 0.20 ± 0.02 s, Cohen's D = 0.3) and muscle firing patterns relative to the onset of the swallow as identified by X-ray video did not differ substantively between younger and older infants. Of the muscles measured, the stylohyoid was the first muscle to fire, prior to the initiation of the swallow as identified by X-ray video, followed by the thyrohyoid and omohyoid, which began their activity at the beginning of the swallow identified in X-ray, similar to previous work [19,20]. A similar pattern of muscle offset was observed, with the stylohyoid turning off prior to the thyrohyoid and omohyoid. The only muscle that differed in its timing of firing was the offset in the thyrohyoid, which was an average of 0.05 s earlier in younger pigs than older pigs (figure 2; electronic supplementary material, table S2). By contrast, Levene's tests for homogeneity of variance found that younger infants had more variable timing in their muscle firing patterns for stylohyoid and omohyoid onset timing and thyrohyoid offset timing, while stylohyoid and omohyoid offset and thyrohyoid onset had similarly variable patterns in young and old infants (figure 2; electronic supplementary material, table S3).

Figure 2.

Muscle activity relative to the frame of the beginning of the swallow identified by X-ray video (time = 0, grey dashed line) for 7-day-old (red) and 21-day-old (blue) pigs when swallowing. Grey bars indicate ± standard deviation. There is very little variation in timing of activity between seven and 21 days, although 7-day-old pigs are more variable in their muscle firing timing than 21-day-old pigs for thyrohyoid and digastric. (Online version in colour.)

(c). Intermarker distance

The three muscles active during both sucking and swallowing behaviours show distinct muscle length changes during activity (figure 3). The stylohyoid was consistently isometric during sucking, whereas the geniohyoid showed variability across individuals. The digastric was isometric or weakly eccentric in the younger pig and isometric in the older pig. By contrast, all three muscles showed some combination of concentric and eccentric contractions during swallowing. The geniohyoid was primarily concentric for the first 50% of its activity and eccentric for the second half (figure 3a,b). In the younger pig, the digastric was concentric initially, and isometric for the second half of its activity, and in the older pig was primarily concentric throughout its activity (figure 3c,d). The stylohyoid was concentric for the first half of its activity, and then was isometric or slightly eccentric for the latter half.

The thyrohyoid and omohyoid only fired during swallowing and had different contractile patterns throughout the duration of their firing. The thyrohyoid was concentric initially, and by 50% of the cycle was eccentric (figure 3g,h; electronic supplementary material, figure S3). By contrast, the omohyoid was isometric or weakly eccentric throughout the duration of its activity (figure 3i,j; electronic supplementary material, figure S3).

(d). Intermarker distance through ontogeny

There were no discrete differences in muscle length change patterns between the younger and older infants. However, younger infants showed more variable patterns of length change than older infants for most instances of muscle length changes during activity (figure 3; electronic supplementary material, table S4).

4. Discussion

The orientation of the muscles powering feeding changes dramatically through ontogeny, with muscles exhibiting a more mediolateral orientation at birth than they do just prior to weaning. This is most likely a by-product of the well-known phenomenon of postnatal craniofacial elongation in mammalian infants [6,8]. In contrast with the substantive anatomical changes occurring throughout infancy, we found few differences in either the timing of muscle activity or the patterns of muscle shortening during activity. This suggests that even though the orientation of the feeding muscles may change, the firing and shortening patterns of their muscles powering feeding remains similar.

(a). The role of variation in muscle firing during feeding

Younger pigs had higher variation in both muscle firing patterns and length changes relative to older pigs. Many of the behaviours associated with infant sucking and swallowing undergo postnatal maturation [12,38,39]. Our results indicate that the neural control of those behaviours may become less variable with age. The higher variation in various components of feeding in young infants suggests that they are feeding with a still developing nervous system [12,40,41].

The higher variation in muscle firing and shortening in younger pigs could also have implications for the control of the hyoid bone. The hyoid is often assumed to move primarily anteroposteriorly and dorsoventrally, with little to no mediolateral movement owing to the precise coactivation of muscle pairs [42–45]. However, with more mediolaterally oriented muscles and increased variation in muscle firing and function, the hyoid may move mediolaterally in young infants compared to older infants, which would be expected to exhibit less mediolateral deviations in hyoid movement during swallowing. The functional consequences of this possibility are not immediately obvious. However, performance failures such as penetration and aspiration are more common in infant feeding than in adults [30,46]. The mediolaterally oriented muscles of infants may serve as an explanation for their increased rates of penetration and aspiration and warrants further study. Head orientation also differs across mammalian infants and may impact the orientation and firing patterns of the muscles involved in feeding. For example, infant humans often feed laying on their back with their chin tucked in, as opposed to most other mammals (including pigs) which feed standing up or lying on their belly. The impact of these differences in posture and position on function are relatively unknown.

(b). Regional heterogeneity in muscle firing

Similar to previous research [21], we found extensive variation in the timing of the geniohyoid depending on the location of electrode insertion. We also found regional heterogeneity in the firing patterns of the stylohyoid. Previously, low amplitude bursts of muscle activity in stylohyoid have been observed during sucking, with larger amplitude bursts during swallowing [20]. However, by systematically placing electrodes in ventral and dorsal locations, we found that only electrodes placed ventrally (close to the hyoid) registered muscle activity during sucking, with dorsally placed electrodes only showing activity during swallowing (electronic supplementary material, figure S2). These results suggest that the muscles of the hyoid that are active during both sucking and swallowing may exhibit functionally relevant regional heterogeneity, with relatively small motor units that can only be accurately assessed with intramuscular fine-wire EMG, rather than surface EMG that is often used in limb muscles and for studies of human populations [47–49].

The regional variation in patterns of activation of the stylohyoid suggests anatomical patterning of motor units, with a ventrally located population of motor units that are specifically involved in sucking. This in turn suggests a patterning of the motor neurons that supply the stylohyoid, which are located in the facial nucleus, with some of these neurons specifically activated during sucking. Sucking and swallowing are hypothesized to be controlled by different central pattern generators [50–52]. Our results for the stylohyoid are consistent with output signals from these two pattern generators projecting to different target neurons within the facial nucleus that target specific subpopulations of muscle fibres within the stylohyoid. Inferences about both normal and pathophysiologic muscle activity based on surface electrodes, particularly in fragile patients, such as infants or the elderly suffering from neurogenic dysphagia cannot provide the degree of resolution that reveals details of the neural control of oropharyngeal function.

(c). Muscle activity and length changes vary between behaviours and muscles

We also observed substantial variation in muscle shortening patterns, both within a muscle during sucking versus swallowing, and between muscles generally. Among the muscles active during both sucking and swallowing, the digastric and stylohyoid both exhibited isometric contractions during sucking, probably to stabilize the jaw (digastric) and to stabilize the hyoid bone (stylohyoid). The stabilization of the hyoid from posterior and ventral directions would enable the anterior muscles, such as the geniohyoid, genioglossus and mylohyoid more effectively to generate suction to acquire and transport milk. As geniohyoid firing patterns are variable and depend on the location of electrode placement [21], its shortening patterns during activity were more variable.

During swallowing, the concentric contractions of the digastric and stylohyoid during the initial 50% of their firing, and eccentric contractions during the latter 50% probably relate to their roles during swallowing. The digastric is primarily a jaw depressor in pigs and does not directly attach to the hyoid, but instead tracks straight from its origin on the mastoid process to the mandible. Its concentric activity probably relates to jaw depression, with eccentric activity in the latter 50% of muscle firing being coupled with jaw stabilization during its elevation associated with the subsequent suck. Similarly, the stylohyoid is generally thought to function primarily as a hyoid elevator [53,54]. Its activity and shortening at the initiation of the swallow probably coincides with hyoid elevation. The latter portion of its activity, when it lengthens, probably stabilizes or decelerates the hyoid as it moves posteriorly and depresses back to its resting position.

We also found variation between the thyrohyoid and omohyoid shortening patterns. As was the case for the stylohyoid, the thyrohyoid was concentric initially, and eccentric at the conclusion of its activity. As the movements of the thyroid cartilage and hyoid bone are tightly coordinated in order to safely swallow [29,44,55], thyrohyoid shortening probably functions to coordinate this movement at the beginning of the swallow. By contrast, the line of action of the omohyoid is primarily to pull the hyoid posterior, and it is often assumed that omohyoid EMG activity acts to move the hyoid posteriorly [20,56]. However, we found that the omohyoid was eccentric throughout the majority of the duration of its activity and is more likely functioning as a hyoid stabilizer during a swallow. Future work explicitly analysing the relationships between muscle shortening and hyoid movements would provide validation for these possibilities [54]. Furthermore, this study examined muscle function in the context of feeding on bottles, and how the muscles powering infant feeding function during breastfeeding remains an open and intriguing question. This is especially true given the greater pressure generation and muscle activation amplitudes exhibited by infant humans when breastfeeding [57,58], which suggests that the patterns observed here may be more strongly evident during feeding directly from the maternal teat.

(d). Ontogenetic and implications for feeding physiology

The lack of changes in muscle activity and shortening patterns despite drastic changes in anatomy also has ontogenetic implications. All mammalian infants feed on milk, and although frequent aspiration can have serious health impacts, some amount of aspiration during feeding on milk is probably normal and does not have detrimental effects on health [30,46,59]. By contrast, aspirating on solid food can have immediate life or death consequences, making the risk much greater for adults. This suggests that the feeding and swallowing system may be optimized for adult physiology, and that infants simply are making the best of their anatomical and neurological configurations. This pattern is also observed in locomotion, as children do not establish mature gait patterns until adolescence and exhibit reduced performance as they transition to new locomotor skills [60–62], although these differences have been argued to be best understood as appropriate for their scale, rather than immature or suboptimal [63]. The relatively high frequency of aspiration in healthy infants may therefore arise as the system adjusts to a constantly changing anatomy and physiology.

5. Conclusion

Our results highlight the complex interplay between anatomy, physiology and postnatal maturation in infant feeding. Sucking and swallowing requires the precise coordination of over 25 paired muscles [51,64]. Our study highlights that in the five muscles we examined, these muscles undergo extensive changes in anatomy through infancy, a pattern we expect to be similar in other muscles involved in infant feeding. However, despite these changing anatomical configurations, infants must be able to maintain an ability to suck and swallow milk until they have concluded the weaning process. We found that they do so by exhibiting high levels of regional heterogeneity in firing patterns, as well as extensive variation in muscle contractile characteristics among different muscles controlling sucking and swallowing. However, there was very little change in the timing of muscle activity and shortening patterns throughout infancy, although variation in timing and shortening patterns decreased with age. The changes in anatomy, coupled with less variation closer to weaning and little change in muscle firing and shortening patterns supports the evolutionary concept that the neuromotor system is optimized to transition to eating solid foods. The decreased risks associated with aspiration when feeding on a liquid diet may not necessitate the evolution of variation in neuromotor function through infancy.

Ethics

All animal care and procedures were approved by NEOMED IACUC no. 19-03-222.

Data accessibility

Data used in statistical analyses are available as part of the electronic supplementary material.

Authors' contributions

Study design was done by C.J.M. and R.Z.G. Data collection was done by all authors. Data processing was done by C.J.M., K.E.S., A.M.C. and L.E.B. Data analysis was done by C.J.M. Manuscript writing was done by C.J.M. Manuscript editing has been done by all authors. Approved manuscript for publication has been done by all authors

Competing interests

The authors declare that they have no conflict of interests.

Funding

This project was funded by an American Association for Anatomy postdoctoral fellowship to C.J.M.