Nipple Properties Affect Sensorimotor Integration During Bottle Feeding in an Infant Pig Model

Abstract

Infant feeding is a critical neurological milestone in development defined by the coordination of muscles, peripheral nerves, and brainstem nuclei. In infants, milk flow rate is often limited to improve feeding performance without treating the underlying deficiencies in the sucking and swallowing processes. Modification of the neuromotor response via sensory information from the nipple of bottle feeding is an unexplored avenue for physiology-based interventions. In this study, we assessed how differences in nipple hole size and nipple stiffness affect sucking muscle activation and subsequent movement. We fabricated four bottle nipples of varying hole size and stiffness to determine how variation in nipple properties affects the sucking behavior of infant pigs. Our results demonstrate that sensory information from the nipple affects sucking motor output. Nipple hole sizes and stiffnesses with a larger milk flow rate resulted in greater muscle activity and kinematic movement. Additionally, our results suggest that sensorimotor interventions are better directed towards modulating tongue function rather than the mandible due to a greater response to sensory information. Understanding how sensory information influences infant feeding is instrumental in promoting effective infant feeding.

Graphical Abstract

Introduction

Feeding is a complex process that requires the coordination of over 20 muscles, multiple nerves, brainstem nuclei, and central pattern generators (CPGs) (Thexton et al., 2007, Marder and Bucher, 2001, Barlow et al., 2014). One key component of feeding in infants is generating suction to acquire milk. Suction is primarily generated by movements of the tongue, jaws, and the CPG controlling their movements (German et al., 1997, Lau, 2015, Geddes et al., 2008). Disruptions to normal sucking behavior in infants can result in aspiration, malnourishment, and delayed development (Hadders-Algra, 2018, Dodrill, 2011, Satter, 1990). One mechanism that shows promise for improving feeding outcomes is varying sensory stimulation to modulate motor output during feeding (Mayerl et al., 2020, Gould et al., 2020, Tsuji et al., 2015, DeLozier et al., 2018, Ding et al., 2013). Oral stimulation via tactile stimulus and nonnutritive sucking (NNS) has been shown to improve feeding outcomes (Fucile et al., 2011, Aguilar-Rodríguez et al., 2020, Song et al., 2019, Barlow et al., 2014). Additionally, application of capsaicin to the soft palate and valleculae has shown to alter the timing of the posterior tongue and decrease the time of bolus travel through the oropharynx (Edmonds et al., 2022). Although we know that sensorimotor feedback plays an important role in infant feeding, we lack insight into the structures and mechanisms by which the motor system of infants responds to variation in sensory information. Identifying which structures involved in suckling can be altered by sensory manipulation will promote a better understanding of the neural control of feeding (Mayerl et al., 2022).

Modifying flow rate is commonly used to alter infant motor output (Nowak et al., 1995, Matthew, 1990). In infant bottle feeding, passive flow rate is controlled using varied hole sizes and cross cuts in the nipple hole (Pados et al., 2015, Pados et al., 2019, Walden and Prendergast, 2000). These modifications have been suggested to improve feeding performance (Nowak, 1994, Pados et al., 2015). However, the relationship between passive flow rates and the physiological deficits underlying feeding difficulties are poorly understood (Mayerl et al., 2022). Active flow rate during infant feeding results from a combination of passive flow rate and suction generation (German et al., 1997, Lau, 2015). Previous research illustrates the potential for suction generation modification depending on variation in sensation (Mayerl et al., 2021). For instance, suction generation can be modified through nipple material properties, where nipple stiffness impacts the force applied to the nipple during nutritive sucking (Adjerid et al., 2022). Continued research on the effect of nipple stiffness integrated with passive flow rate is essential to understand the neurophysiology of infant feeding.

In this study, we aim to assess the impacts of varying nipple stiffness and nipple hole size on the sucking response of bottle-fed infant pigs just prior to weaning. By using an infant pig model, we can study aspects of physiology not possible in humans due to health and ethical concerns, while controlling for comorbidities and confounding variables that can arise in human research (German et al., 2022). We hypothesize that if nipple properties are changed, infant feeding motor output will change due to the influences of altered oral sensory information. To test this hypothesis, we will analyze the effects of material stiffness and hole size on oropharyngeal kinematics and muscle activity. We predict that more compliant nipples with a larger hole size will increase motor output in a positive feedback loop.

Methods

Animal Housing and Care

Animal care and procedures were approved by NEOMED IACUC (19-03-222). We acquired full term infant pigs (Yorkshire/Landrace sows, Shoup Investments, LTD, Wooster, OH) at 1 day of age (N =1 litter, 6 infants). Pigs received an infant pig milk replacer formula (Solustart Pig Milk Replacement, Land o’ lakes, Arden Mills, MN) for the duration of the experiments. Additional housing and feeding followed care guidelines previously outlined but are standard for this animal model (Mayerl et al., 2021, Holman et al., 2013).

Nipple Design

Four experimental nipples and one training nipple of identical shape and size were designed to evaluate the effect of material stiffness and hole size on feeding physiology. Artificial nipples were shaped according to commercially available animal nipples (Nasco Agriculture Education, Fort Atkinson, WI). Nipples were modeled in Autodesk Inventor (Autodesk Inc., San Rafael, CA, USA), 3D printed, and then cast using food safe silicone. This enabled us to design nipples of uniform shape and size while only varying stiffness and hole size. We designed the training nipple with intermediate properties relative to experimental nipples for routine bottle feeding (Shore 40A hardness and a 1.0$\pi$ mm² cross-sectional area hole size (CSA, Table S1)). The use of the training nipple ensured that all pigs fed on an intermediate nipple prior to exposure to the experimental nipples, and minimized the probability of piglets rejecting experimental nipples. The two compliant nipples were cast with shore hardness silicone ratings of 20A (Cast-a-Mold 20A hardness, Specialty Resin and Chemical, Dowagiac, MI), and the two stiff nipples were cast with shore hardness silicone ratings of 55A (ReproRubber 55 A hardness, Flexbar, Islandia, NY). For each of the stiffness ratings, nipples were designed with a large hole size (2.0 π mm² CSA) and a small hole size (0.5π mm² CSA), corresponding to a four-fold difference in passive flow rates (2.44 ± 0.19 ml / s for the small hole and 8.33 ± 0.00 ml/s for the large hole). Young’s modulus values and passive flow rates were experimentally determined for each nipple type (Adjerid et al., 2023).

Surgical Procedures

Two surgical procedures were performed prior to data collection to facilitate x-ray tracking of oropharyngeal movements and electromyographic (EMG) data collection. Surgical procedures were done in two timelines to reduce extended anesthesia times and promote healthy recoveries. When the infants were five days old, aseptic surgery was performed under isoflurane anesthesia (2–4%) to suture 1.0mm laser-drilled tantalum beads into the fascia over the hyoid bone and thyroid cartilage. Hyoid markers were implanted between separated bellies of the sternohyoid toward the anterior insertion of the muscle onto the hyoid. Thyroid markers were sutured to the connective tissue over the thyroid eminence. We did not measure hyoid or thyroid kinematics in this manuscript, but mention the procedure for completeness. Following closure of the incision, a custom bead injector was used to implant 0.8mm tantalum beads into the hard palate, tongue midline (anterior, middle, and posterior locations), and sub-dermal spaces beneath the shout and chin.

A second aseptic surgery was performed on infants between 17–19 days of age. Additional 1.0mm beads were sutured to the connective tissue adjacent to the hyoid and mandible for rigid body reconstruction. Two beads were implanted bilaterally on the hyoid, approximately 6mm from the pre-existing hyoid marker. Mandible markers were placed on the lateral margin of the mandible by suturing a bead into the connective tissue on the mandible, adjacent to the masseter. We positioned two bipolar electromyographic (EMG) electrodes bilaterally into three of the muscles associated with infant feeding (Delozier et al., 2018, Thexton et al., 2007). We chose three muscles primarily associated with sucking: genioglossus (an extrinsic tongue muscle), digastric (a jaw depressor in pigs that attaches directly to the cranium), and masseter (a jaw elevating muscle). Micro connectors attached to the electrodes exited the animal at the posterior margin of the incision where they were connected to a 25-pin D-connector before being secured by vet wrap. Animals received buprenorphine (0.1–0.15 ml/kg) subcutaneously and bupivacaine (0.2 ml/kg) topically as analgesics for both surgical procedures.

Data Collection

We collected synchronized EMG and kinematic data following recovery from the second aseptic surgery just prior to weaning. With the small-hole stiff nipple, infant pigs were unable to produce consecutive sucks and swallows. Consequently, data includes samples collected on the small-hole compliant, large-hole compliant, and large-hole stiff nipples only. Fluoroscopic data was collected using bilateral x-ray video (GE 9400C-Arm, 71–73 kV, 6.3 mA) and high-speed cameras (XC1 M, XCitex, Cambridge, MA) at 100 FPS following grid undistortion and cube calibration using standard XROMM data collection procedures (Brainerd et al., 2010). We bottle fed infant pigs a combination of milk and barium for x-ray contrast (E-Z-Paque, Radtech Xray, Vassar, MI). Experimental nipples were tested in a randomized order with training nipple washouts of 10 swallows between trials. After the first 10 seconds of feeding, which comprises swallows with an uncharacteristically high frequency (Gierbolini-Norat et al., 2014), 20 swallows were recorded per pig per experimental nipple to ensure sample size was adequate for each nipple, and not impacted by feed duration or feeding efficiency depending on nipple type. EMG recordings were amplified and recorded at 10 kHz on a 16 channel Powerlab (30/16, AD Instruments, Colorado Springs, CO). Postmortem, electrode placement was assessed via dissection.

Data Processing

Due to issues with EMG signals, we extracted and processed kinematic and EMG data from four of the six pigs with a total of 863 suck cycles. Feeding issues for the large-hole stiff nipple in one of the infant pigs resulted in a sample size of three for the large-hole stiff data set, so the sample size for nipples is as follows: large-hole compliant nipple n=4 pigs, 318 sucks, small hole compliant n= 4 pigs, 308 sucks, and large-hole stiff n=3 pigs, and 239 sucks. Suck cycles were defined as the period from the frame of contact between the middle of the tongue and the hard palate to the frame prior to the next instance of contact (Mayerl et al., 2021). Kinematic data were tracked in lateral and dorsoventral views using XMALab (Knörlein et al., 2016). Markers associated with the hard palate and nose were utilized to produce a rigid body for skull reconstruction, as the intermarker distances had standard deviations of less than 2 mm. We were unable to produce a rigid body of the mandible because of high intermarker distances between beads, and therefore, we are using a chin marker as an approximation for mandible movements. The cranium rigid body was exported from XMALab in combination with XYZ points of individual markers.

Cranium, mandible, and hyoid bones from each animal were segmented out of CT scans postmortem using Avizo 9.4 (FEI Visualizations Science Group, Hillsboro, OR). Segmented bony structures were imported into Autodesk Maya (Autodesk Inc., San Rafael, CA, USA). To standardize the placement of the cranial anatomical coordinate system (ACS) across individuals, we utilized three planes. The first plane was centered at the midline of the animal at the temporomandibular joint. The second plane ran anteroposteriorly along the cranium parallel with the hard palate, and the last plane aligned perpendicular to the long axis of the cranium. We then placed an ACSat the center, with X translations measuring anteroposterior movement, Y translations measuring vertical movement, and Z translations measuring mediolateral movement. Because our markers failed to capture the movements of the mandible, we only report translations of markers relative to the skull. We imported XMALab output into Maya and calculated the intermarker distance between the tongue/chin markers and ACS of the cranium using the ORel feature within the XROMM_MayaTools package. XYZ translations of the chin and anterior tongue relative to the skull were exported from MAYA.

Following visual assessment of EMG signals, EMG data were run through a custom R script for baseline correction, rectification, integration, and threshold correction of the raw signal (Thexton, 1996). EMG and kinematic data were imported into a MATLAB script for synchronization (Fig. S1), summation of the XYZ translations, and interpolation of the data with the suck cycle as a percentage to account for differences in suck cycle duration.

For each of our three nipple types (because animals rejected the combination stiff/small hole nipple), we measured kinematic excursions (EXC, Table S1) of the tongue relative to the skull (AntRelSkull, Table S1, Movie 1) and chin (AntRelChin, Table S1). EXC were calculated from the range of the 3d anterior tongue displacement from the temporomandibular joint where the ACS is centered for AntRelSkull, and from the range of displacement between the anterior tongue and chin marker for AntRelChin. EXC of the chin relative to the skull (ChinRelSkull, Table S1) were also calculated. Movement of the tongue in the anterior region during suckling was driven primarily by dorsoventral pumping, with minimal anteroposterior or mediolateral movement. The timing of the maximal depression (Max Dep, Table S1) and maximal elevation (Max Elv, Table S1) were calculated for AntRelSkull, AntRelChin, and ChinRelSkull. Timings are represented in position graphs (Figures 3–5) where position (Table S1) during a suck cycle with positive values indicate depressions and negative values indicate elevation relative to the initial position in the suck cycle. Therefore, kinematic timing variables are not standardized by animal. For muscle activity, EMG area under the curve (AUC, Table S1) and duration were calculated. To standardize the response to individuals, Kinematic EXC, EMG AUC, and EMG duration were scaled to each individual, with a value of 1 representing the maximum value for an individual.

Figure 3.

Position of the anterior tongue relative to the skull during a suck cycle. The anterior tongue was maximally depressed earlier when feeding on stiff nipples (p < 0.001, D = 2.58) or nipples with smaller hole sizes (p < 0.001, D = 1.27), and maximally elevated earlier when feeding on stiff nipples (p < 0.001, D = 1.45).

Figure 5.

Position of the chin relative to the skull during a suck cycle. Stiffer nipples resulted in earlier timing of maximal elevation (p < 0.001, D = 0.53) and maximal depression (p < 0.00, 0.81), whereas smaller holes resulted only in earlier maximal depression of the chin (p < 0.001, D = 0.67).

Statistical Analysis

All statistical tests were performed in RStudio (v. 1.1.463, http://www.r-project.org/). We utilized a linear mixed effect model (Bates et al., 2015) with hole size and nipple stiffness as fixed effects and individuals as a random effect. Since the infant pigs were unable to feed on the small-holed, stiff nipple, we were unable to measure the interactive effect of the two fixed effects. Therefore, statistical comparisons were made between the small compliant and large compliant nipples to evaluate the effects of passive flow rate, and between the large stiff and large compliant nipples to evaluate the effects of nipple stiffness. Following linear mixed effect modeling (with individual as a random effect), Cohen’s D analyses were used to evaluate the effect size of each independent variable on feeding physiology (Cohen, 1992). Cohen’s D responses of medium and large effect size were considered significant for this study.

Results

EMG activity during sucking

Overall, we found that changes in nipple properties altered EMG activity in muscle associated with the tongue (genioglossus) but not the mandible (masseter and digastric, Table 1). In genioglossus (GG, Table S1), stiffer nipples and smaller holed nipples resulted in a greater duration of muscle activity, but a smaller AUC (Table 1, Figure 1). This was true despite the fact that the duration of the suck was similar in large holed compliant (mean = 0.24 s ± 0.05 s) and stiff (mean = 0.23 s ± 0.03 s) nipples, with the small holed compliant nipples having longer total cycle durations (mean = 0.30 s ± 0.06 s, Planned contrasts: Large compliant vs large stiff: t = 0.9, p = 0.39; large compliant vs small compliant: t = 14.5, p < 0.001, large stiff vs small compliant t = 14.3, p < 0.001).

Table 1.

EMG response to variation in nipple properties.

	Large Stiff Mean ± SD	Large Comp Mean ± SD	Small Comp Mean ± SD	Stiffness p, D	Flow p, D
GG AUC	44.5 ± 13.2%	58.2 ± 15.3%	45.9 ± 15.1%	<0.001, 0.96	<0.001, 0.81
Mass AUC	53.2 ± 13.2%	47.0 ± 13.4%	39.5 ± 17.2%	0.982, 0.44	<0.001, 0.46
Dig AUC	52.6 ± 15.7%	45.3 ± 21.3%	37.8 ± 20.4%	<0.001, 0.38	<0.001, 0.36
GG Duration	0.20 ± 0.06s	0.16 ± 0.05s	0.22 ± 0.08s	<0.001, 0.73	<0.001, −1.02
Mass Duration	0.10 ± 0.03s	0.13 ± 0.07s	0.15 ± 0.08s	0.01, 0.55	<0.001, −0.29
Dig Duration	0.15 ± 0.05s	0.13 ± 0.06s	0.13 ± 0.06s	0.002, 0.31	0.278, 0.02

Figure 1.

Genioglossus AUC (A) was higher with more compliant nipples, and with nipples with higher flow rates, but duration was lower when feeding on more compliant nipples with higher flow rates (B). Lines connecting box plots indicate statistically significant differences between groups.

However, neither change in nipple properties resulted in significant differences with medium or large effects sizes in duration of muscle firing or EMG AUC for digastric (Dig, Table S1) or masseter (Mass, Table S1, Table 1).

Kinematic patterns during sucking

Generally, nipple properties affected kinematic excursions (EXC, Table S1) of the tongue and chin in a consistent pattern with EMG activity (Table 1 and 2). Feeding on stiff nipples resulted in smaller EXC of the anterior tongue relative to both the skull (AntRelSkull) and the chin (AntRelChin) (Figure 2). However, there was no effect of nipple hole size on AntRelSkull EXC or AntRelChin EXC (Figure 2). Chin EXC per suck did not vary across nipple types, which is consistent with the lack of change in masseter or digastric activity (Table 2).

Table 2.

Kinematic response to variation in nipple properties.

	Large Stiff Mean ± Sd	Large Comp Mean ± SD	Small Comp Mean ± SD	Stiffness p, D	Flow p, D
AntRelSkull EXC	43.7 ± 17.8%	57.9 ± 14.3%	59.0 ± 16.9%	<0.001, 0.89	0.216, 0.06
ChinRelSkull EXC	63.2 ± 8.4%	67.3 ± 17.0%	66.1 ± 13.8%	0.953, 0.29	0.376, 0.08
AntRelChin EXC	42.7 ± 15.3%	71.0 ± 12.2%	66.1 ± 10.7%	<0.001, 2.08	<0.001, 0.43
AntRelSkull Max Dep	7.8 ± 12.6%	39.7 ± 12.6%	22.0 ± 15.1%	<0.001, 2.58	<0.001, 1.27
AntRelSkull Max Elv	57.6 ± 16.3%	81.7 ± 16.8%	74.7 ± 19.1%	<0.001, 1.45	<0.001, 0.38
ChinRelSkull Max Dep	19.1 ± 18.5%	31.4 ± 12.0%	20.1 ± 21.0%	<0.001, 0.81	<0.001, 0.67
ChinRelSkull Max Elv	60.6 ± 17.8%	68.8 ± 13.4%	61.1 ± 18.1%	<0.001, 0.53	<0.001, 0.48
AntRelChin Max Elv	57.6 ± 26.0%	51.5 ± 11.1%	37.2 ± 11.7%	<0.001, −0.32	<0.001, 1.25
AntRelChin Max Dep	23.9 ± 27.9%	16.5 ± 21.7%	2.0 ± 11.1%	<0.001, −0.30	<0.001, 0.84

Values highlighted in bold are statistically significant with a medium or large effects size.

Figure 2.

Anterior tongue excursion standardized to the largest excursion produced per pig, relative to the skull (A) and chin (B) was lowest when feeding on stiff nipples, with no impact of hole size on excursion. Lines connecting box plots indicate statistically significant differences between groups.

Variation in nipple properties also affected the timing of the tongue and chin movements (Table 2, Figure 3, 4, 5). Maximal anterior tongue depression relative to the skull (AntRelSkull Max Dep, Table S1) occurred earlier in the suck cycle when feeding on stiffer or smaller holed nipples (Table 2, Figure 3). Additionally, feeding on stiff nipples resulted in earlier maximal tongue elevation relative to the skull (AntRelSkull Max Elv, Table S1, Table 2, Figure 3). Relative to the chin (AntRelChin), smaller hole size resulted in an earlier maximal depression and elevation of the anterior tongue (Figure 4), whereas there was no effect of stiffness (Table 2). Despite the consistency in EMG activity and kinematic excursions across varying nipple properties, hole size and stiffness both impacted the temporal relationship between chin movements and the suck cycle (Table 2, Figure 5). Stiffer and smaller holed nipples resulted in earlier chin Max Dep (Table S1, Table 2, Figure 5), and chin Max Elv (Table S1) occurred earlier with stiff nipples (Table 2, Figure 5).

Figure 4.

Position of the anterior tongue relative to the chin during a suck cycle. The maximal elevation of the anterior tongue relative to the chin was earlier when feeding on smaller holed nipples than on larger holed nipples (p < 0.001, D=1.25).

Qualitative Results

When infant pigs fed on the small-holed stiff nipple, all 6 pigs were unable to produce consecutive sucks and swallows necessary for the analysis of rhythmic sucking. Additionally, 1 of the 6 pigs was unable to successfully feed on the large-hole stiff nipple.

Discussion

Sensorimotor Integration in Infant Feeding

Changes in the kinematic and neuromotor response to variation in nipple properties illustrate that mechanical sensation from the nipple affects nutritive feeding physiology. This conclusion builds upon the literature showing that oral and laryngeal sensory modification affects infant feeding physiology (Finan and Barlow, 1998, Fucile et al. 2002, Aguilar-Rodríguez et al., 2020, Song et al., 2019). Our controlled nipple design further explicates the mechanistic relationship between sensory information and motor output in sucking physiology. Our data demonstrate that infant suckling is generally powered through a ‘positive feedback loop’, in which tongue movements are greater when infants feed on a more compliant nipple that facilitates milk flow (Figure 2). Furthermore, EMG analyses demonstrated that increased tongue movement is accompanied by higher genioglossus activity over a shorter period (Figure 1), which may explain how suckling effort increases when milk flow is higher. We found a similar effect on varying hole size, in which genioglossus activity was higher in the large hole compliant nipple than the small hole compliant nipple, also over a shorter duration (Figure 1). The greater activity seen in the musculature when feeding on nipples that facilitate a higher milk flow demonstrates that infants sucking can be modulated by the properties of the nipple, and that this is not driven by increased suck duration.

Neuromotor Control of Infant Feeding

Altering nipple properties affected the temporal coordination between the mandible and the tongue during suckling (Figure 4, Table 2). When passive flow rate was higher, the timing of the maximum separation of the anterior tongue and chin was later in the suck cycle. Oropharyngeal kinematic timings have shown to affect downstream feeding outcomes (Gould et al., 2020, DeLozier et al., 2018). Consequently, future research should determine how delayed tongue and chin separation relate to performance.

The differences between the tongue and mandible responses to varying stiffness and hole size were striking. While genioglossus activity and tongue movements varied with stiffness and hole size, the muscles of the jaw (digastric and masseter) did not differ in muscle activity with changes in hole size or stiffness, and there were no changes in movements of the mandible across nipple properties (Table 1 and 2). One possible explanation for the differences between the tongue and jaw response is the ontogeny of the infant feeding system. Jaw muscles may retain their primary functions through weaning, while the muscles of the tongue undergo significant changes to facilitate the transition from milk to solid foods. Overall, the varied motor response suggests different susceptibility of the two pathways to sensory information, which supports previous research that the neuromotor control of sucking is more complex than a single, cohesive response generated by the sucking CPG (Marder and Bucher, 2001, Zimmerman et al., 2008).

Implications for the neuromotor control of infant feeding and future directions

Our data has implications for the neuromotor control of infant feeding. Extensive variation in both flow rates and nipple stiffness exist across and within different bottle manufacturers (McGrattan et al., 2017, Pados, 2021, Salisbury 1975, Pados et al., 2019, Nowak et al., 1995). However, while the neural basis of the motor changes in response to nipple type can be inferred they are not entirely clear. Although jaw movement is reduced in suckling relative to mastication (German et al 1992; Thexton et al 2012), studies investigating the neural networks that control the phasic relationships among trigeminal and hypoglossal motor outputs highlight the ability of the neonatal nervous system to coordinate tongue and jaw movements as a part of the overall sucking motor response (Ihara et al 2013). However, the majority of differences in motor pattern associated with the transition to solids are in the tongue muscles, and not jaw kinematics (German et al., 1992; Thexton et al., 2012). Sensorimotor integration in the tongue integrates several cranial nerves, as well as components within the muscles themselves. For example, there are links between trigeminal afferents and hypoglossal efferents (Miller, 2006), similar to our results demonstrating a change in hypoglossal output following a change in trigeminal input. Furthermore, spindles or stretch receptors exist in the tongue muscles (Mathew et al 1982; Smith, 1989; Aeba et al 2002; Steele and Miller, 2010). These receptors, as well as the general sensation of the tongue surface, through V3 and IX, are the likely respondents to our treatments. There is also debate as to whether other muscles critical for suckling and swallowing, for example the oral floor or the pharynx, have spindles (Maier 1979; Muhl and Kotov 1988). Work on sensorimotor integration using nerve lesions as a model have demonstrated that significant sensorimotor integration occurs within the brainstem (DeLozier et al 2018; Gould et al 2020), as well as through cortical pathways (Martin 2009; Steele and Miller 2010). Our results suggest that specific experiments to isolate these pathways would help clarify the exact mechanisms that produced our results.

Furthermore, the relationship among milk flow rate, sucking physiology and performance have yet to be described. Often, infants exhibiting feeding difficulties are prescribed nipples with lower flow rates (Pados et al., 2015, Matthew and Cowen, 1988). Alternatively, infants with feeding difficulties are often fed milk with increased viscosity, which reduces the frequency of aspiration (Mayerl et al., 2021, Newman et al., 2016, Cichero et al., 2016). One limitation of these interventions is that they do not necessarily treat the underlying etiologies that generate feeding difficulties (Mayerl et al., 2022). For example, many of the feeding issues common in preterm infants involve poor abilities to generate suction (Mayerl et al., 2019, Lau, 2015, Dodrill, 2011). The positive feedback result in our study suggests that prescribing nipples that enhance suction generation by increasing nipple compliance may improve milk acquisition and muscular development. Future research should examine this result with inquiry into development and swallow safety. Exploring muscle development and consequential kinematic output in an ontogenetic study can elucidate if the positive feedback system shown in our study results in improved development. Furthermore, as the maturation of sucking patterns is an important neurological milestone in infant development (Gewolb et al., 2001, Lau et al., 2000), modulating stiffness has the potential to improve neurological development. Importantly future work should also assess whether the increase in motor output results in altered swallow safety. Determining the balance between neuromotor development and swallow safety is critical in optimizing infant feeding. Finally, this work focuses on only one representative mammalian infant, and delving into the mechanics of suckling, and how infants respond to variation in nipple properties across mammals (such as those born singly, instead of in a litter) represents an important avenue for future research.

Limitations

Because infant pigs were unable to produce consecutive swallows on the small holed stiff nipple, we could not examine the quantitative, interactive effect of stiffness and hole size. Future research should continue examining how the two milk reducing factors of stiffness and small hole size interact to affect infant feeding. Another limitation was the use of the chin marker as a proxy for overall mandible movement is an imperfect approximation. Future research should attempt to quantify the three-dimensional movements of the mandible.

Conclusions

Our results demonstrate that modulating milk flow rate and nipple stiffness during bottle feeding impacts infant feeding physiology. In general, infants respond to variation in these properties through a positive feedback system in which they increase effort with increased milk flow. This result suggests that nipple properties that facilitate milk flow have the potential to improve neuromotor development. Our results also describe how the muscles (and subsequent movements) of the mandible do not vary in accordance with nipple properties, whereas the tongue responds to both variation in hole size and nipple stiffness. By demonstrating the physiological response to variation in nipple properties, our data provides a foundation for optimal infant feeding development.

Data availability

Data used in statistical analyses will be deposited in a public repository upon acceptance.