The Function of the Mammal Extrinsic Tongue Musculature in the Transition from Suckling to Drinking

Synopsis

The transition from suckling to drinking is a developmental pathway that all mammals take. In both behaviors, the tongue is the primary structure involved in acquiring, transporting, and swallowing the liquid. However, the two processes are fundamentally different: during suckling, the tongue must function as a pump to generate suction to move milk, whereas during drinking, the tongue moves backwards and forwards through the mouth to acquire and move water. Despite these fundamental differences, we have little understanding of how tongues role varies between these behaviors. We used an infant pig model to investigate the relationships between anatomy, physiology, and function of the tongue to examine how lingual function is modulated in the transition from infancy to adulthood. We found that while some muscles were proportionally largest at birth, others were proportionally larger at the time of weaning. Furthermore, we found variation in tongue movements between suckling and drinking along both the mediolateral and anteroposterior axes, resulting in differences in tongue deformation between the two behaviors. The extrinsic tongue muscles also changed in function differently between drinking and suckling. Genioglossus increased its activity and turned on and off earlier in the cycle during drinking, whereas hyoglossus fired at lower amplitudes during drinking, and turned on and off later in the cycle. Together, the data highlight the significant need for high neuroplasticity in the control of the tongue at a young age in mammals and suggest that the ability to do so is key in the ontogeny and evolution of feeding in these animals.

Introduction

The presence of a mammary gland that produces milk, and the subsequent consumption of that milk, infant suckling are key evolutionary innovations in mammals and even form the basis for their name. Yet, the evolution of mammals is often told through the evolution of teeth (Hiiemae and Crompton 1985; Thomason and Russell 1986; Gould et al. 1987; Smith 1992; Bhullar et al. 2019; Zhou et al. 2019). The importance of the multiple aspects of the mammalian feeding system in shaping evolution has been recognized even pre-evolutionary theory, with George Cuvier famously claiming “Show me your teeth and I’ll tell you who you are.” Thus, suckling and the transition to solid foods are key features that all mammals go through. While we have a general understanding of the mechanics of suckling (German et al. 1992, 1997; Geddes et al. 2012; Thexton et al. 2012), and a fairly detailed understanding of how adult mammals feed (Hiiemae and Crompton 1985; Ross et al. 2007; Taylor and Vinyard 2009; Bhullar et al. 2019; Feilich et al. 2021), our understanding of the transition from infant to adult behaviors is limited (Thexton et al. 1998; Bond et al. 2020). This transition is further complicated because the properties of food being consumed varies greatly between the two life stages: as infants, mammals consume milk, a liquid, but most adult mammals consume and process solid foods. This difference makes understanding maturational changes in the mechanisms of feeding challenging.

One potential avenue to explore the developmental changes in feeding mechanics between infancy and adulthood lies in comparing suckling with drinking, again a process that most adult mammals must perform. The mechanics of drinking vary across mammals, with some animals using suction (including humans and pigs) and others, like dogs, using lapping (Crompton and Musinsky 2011; Gart et al. 2015 ; Olson et al. 2021a), but unlike with processing food, there is minimal mediolateral deformation in the tongue during drinking, and there is little oral processing prior to swallowing (Olson et al. 2021b). Similarly, as with suckling, the teeth and jaw movements are not the primary structures associated with drinking, and the tongue plays the primary role of acquiring, transporting, and swallowing liquids (Thexton 1984 ; Olson et al. 2021b).

Consumption of liquids in mammals of any age is primarily a function of the tongue. The mammalian tongue is a complex structure comprised of both intrinsic (originating and inserting within the tongue) and extrinsic (originating outside of the tongue and inserting into it) muscles (Mu and Sanders 1999, 2010; Kayalioglu et al. 2007; Sokoloff and Burkholder 2012). The extrinsic tongue muscles (genioglossus, hyoglossus, styloglossus, and palatoglossus) are generally thought to be responsible for protruding, retracting, depressing, and elevating the tongue (e.g., Kayalioglu et al. 2007), whereas the intrinsic muscles are generally regarded to function to adjust the shape of the tongue (Lowe 1980), although this may not be as simple as historically represented (Liu et al. 2009; Orsbon et al. 2020). In infants, the extrinsic tongue muscles are necessary to form a seal around a nipple to generate suction, whereas during drinking for most mammalian adults, the tongue acts to acquire and transport water from the ground with minimal (or absent) suction (German et al. 1992, 1997; Geddes et al. 2012; Mayerl et al. 2020a; Olson et al. 2021a). Because of these differences in function, it has been hypothesized that the anatomy and physiology of the tongue must differ between infants and adults (Iskander and Sanders 2003). For example, it has been suggested that the extrinsic tongue muscles, as with the rest of the tongue, are proportionally larger during infancy, and that they act as a buttress to enable the peristaltic wave to generate suction and acquire food (Iskander and Sanders 2003), a function not necessary in adult drinking.

Understanding the differences in the anatomy, physiology, and function of the mammalian tongue would provide key insight into neuromotor plasticity and would reveal how anatomy and physiology are linked. Here, we use a pig model to investigate (1) how the muscles powering the feeding system change through infancy, (2) differences in how the tongue moves between suckling and drinking, and (3) differences in the activity two extrinsic tongue muscles during suckling and drinking. By integrating these approaches, we investigate the relationships between anatomy, physiology, and function of the tongue to examine how lingual function is modulated in the transition from infancy to adulthood, with the expectation that the reduced effort to drink relative to suckle will be reflected by changes in muscle activity and tongue movements.

Methods

We collected data over two separate experiments. In the first, we collected data on tongue kinematics, and in the second, we collected electromyographic (EMG) data. We additionally collected diceCT data from postmortem specimens.

Animal housing and care

For both experiments, infant pigs (Yorkshire/Landrace sows, Shoup Investments LTD, Wooster, OH, USA) were obtained at 1 day old and housed in the NEOMED Comparative Medicine Unit (CMU). They were trained to feed on infant milk replacer (Solustart Pig Milk Replacement, Land o’ lakes, Arden Mills, MN, USA) from custom-built nipples (Stiffness: Durometer 20A), using a sow’s nipple as a base to determine shape and size. Infant care followed standard procedure approved by NEOMED IACUC Protocol # 19-03-222 (Mayerl et al. 2019, 2020b, 2020c, 2021a, 2021b).

Surgical procedures

To evaluate tongue kinematics, we used an 18-gauge needle with a metal plunger to implant tantalum markers (0.8 mm) at five days of age in several structures of interest in three pigs while maintaining isoflurane anesthesia (2–4%) (Mayerl et al. 2020b; Edmonds et al. 2020). Markers were inserted into the hard palate (four markers), subcutaneously in the rostrum (one marker), in the soft palate (one marker), in the palatopharyngeal arches (two markers), and in the tongue (nine markers). We placed tongue markers at the midline in anterior, middle, and posterior locations, with markers at the lateral margins of the tongue on both sides at the same position along the anteroposterior axis.

To record muscle function via EMG, we performed a sterile surgery between 18 and 20 days of age in three pigs to implant bipolar electrodes into muscles of interest. We used a 19-gauge needle to place electrodes in the left and right bellies of the genioglossus near the chin, as well as in the left belly of the hyoglossus in anterior and posterior locations to measure their activity. We also implanted electrodes in the thyrohyoid, a muscle known for its stereotyped activity during swallowing (Thexton et al. 2007), as a marker for identifying swallows in the EMG signals. We identified the genioglossus by parting the two bellies of the mylohyoid (Mayerl et al. 2022) with blunt dissection, then parting the two bellies of the geniohyoid at the midline. Electrodes were inserted into the muscle belly of the genioglossus, pointing anteriorly on both sides of the animal. Hyoglossus was identified by gently reflecting the posterior belly of the mylohyoid anteriorly (for posterior electrode implantation) or posteriorly (for anterior electrode implantation). Thyrohyoid was identified as the muscle lying on the ventral surface of the larynx between the thyroid eminence and the hyoid bone.

Electrodes were sutured to connective tissue as a means of strain relief, collected and tied together, and then exited the body through the posterior area of the incision. Electrodes were pre-operatively soldered to a 6-pin microconnector, which was attached to an HDMI connector postoperatively, which was secured using self-adhering wrap. Following data collection, animals were euthanized, and electrode positions were confirmed with dissection.

Data collection

Tongue kinematics

To collect data on tongue kinematics, we filmed infant pigs at approximately 20 days of age (after pigs have begun to transition from bottles to bowls) using high-speed bilateral videofluoroscopy (GE 9400C-Arm, 71–73 kV, 6.3–7.1 mA) with high-speed cameras (XC1 M, XCitex, Cambridge, MA, USA, 100 fps) when pigs fed either on a bottle, or drank from a 3-D printed bowl with a funnel and tube used to maintain milk volume throughout a feed. Data were collected on the same day to minimize variation in neuromotor maturation and anatomy, so that the only variables of interest would relate to the physiology underlying the behavior. Prior to data collection, we calibrated and undistorted the image following standard XROMM data collection procedures (Brainerd et al. 2010). Pigs were fed on milk replacer mixed with barium (E-Z-Paque, Radtech X-ray, Vassar, MI, USA) in a radiolucent plexiglass box. We collected at least 20 swallows per pig per feeding condition on the same day, in a randomized order per pig.

EMG

We followed similar data collection procedures for EMG pigs to collect biplanar x-ray video. In addition, we synchronously collected EMG data during feeding using a 16-channel Powerlab (16/35, ADInstruments, Colorado Springs, CO, USA) at 10 KHz following amplification (MA-300, Motion Laboratory Systems, Baton Rouge, LA, USA).

Anatomy

We collected contrast-enhanced CT (diceCT, [Mayerl et al. 2021b; Gignac et al. 2016]) data from two pigs: a newborn infant and a newly weaned, 25-day-old juvenile pig. Prior to staining, we collected a CT scan of the animals to acquire skeletal data. To facilitate scanning, we removed the cranium and one side of the lower jaw. We fixed pigs in formalin (10%) for 14 days, and then transferred them to a 2.5% solution that was refreshed every 10 days until staining was complete (approximately 1 month total). Following staining, we microCT-scanned the pigs (30 µm, 70 kVp, and 22 µA).

Data processing

Tongue kinematics

We identified periods of consistent feeding for both bottle feeding and drinking from the x-ray video. Within this sequence, we isolated individual sucks by defining the start of the suck as the frame that the tongue made contact with the hard palate to form a seal, and the end of the suck as the frame before the next suck began. We tracked radio-opaque markers in lateral and dorsoventral views in XMALab (Knörlein et al. 2016). Markers in the hard palate and nose had intermarker distance standard deviations of less than 0.03 mm, and were used as a single rigid body. We exported rigid body transformations and 3D positions of all markers with a 10 Hz low-pass filter as previously described (Mayerl et al., 2020b; Edmonds et al. 2022). Additionally, we exported undistorted video and Mayacams from both views. We imported these into Autodesk Maya (Autodesk, Inc. San Rafael, CA, USA). To calculate the kinematic movements of the tongue, we created an object in Maya, and imported the rigid body transformations of the cranium to that object. We then created an axis and placed it at the antero-most point of the hard palate, and aligned it at the midline of the animal. We positioned the axis such that the x-axis aligned anteroposteriorly, the y-axis aligned mediolaterally, and the z-axis aligned dorsoventrally relative to the hard palate. We then parented the axis to movements of the cranium transformations, and measured translations of the tongue markers relative to this skull axis (oRel, XROMM toolshelf). Doing so enabled us to create a reference that neutralizes the movements of the pigs head and standardizes movements across pigs (Mayerl et al. 2020b).

XYZ translations of these movements were exported from Maya and imported into a custom MATLAB App that calculated variables of interest for individual feeding cycles (sucks, or drink cycles). In this app, we calculated: (1) the movement of the tongue markers in each dimension throughout the cycle (raw and interpolated to 101 points to account for variation across cycles in duration), (2) the excursion of tongue movement in each duration within a single cycle (e.g., the distance between the most elevated and depressed point in the cycle for dorsoventral excursion), and (3) the three-dimensional vector distance between two markers along the anteroposterior axis (either anterior to middle, or middle to posterior) throughout the cycle. For each cycle, we standardized the start location as being zero to account for variation in marker placement between pigs.

EMG

As with kinematic data, we identified periods of consistent feeding for both bottle feeding and drinking from x-ray video and identified sucks within this range, synchronized with EMG data with a TTL trigger. For each feeding sequence within this range, we applied a bandpass filter on EMG data to reduce apparent noise (150 Hz low-pass, 300 Hz high-pass) and exported data from Labchart at 10 kHz for each electrode. Hyoglossus displayed regional heterogeneity in firing patterns, with anterior electrodes recording acquisition behaviors (suck and drink cycles), and posterior electrodes recording only swallow behaviors (Supplementary Fig. S1). As such, we present data only from anterior electrode positions in the hyoglossus.

Exported EMG data were rectified, integrated, and threshold noise levels were determined in R (V 4.0.3) based on published procedures for cyclical data (Thexton 1996; Mayerl et al. 2022). Thresholded data were imported into a custom MATLAB script that calculated the onset, offset, firing duration, and area under the curve (AUC) of a given electrode per suck. Onset was determined as the time where the electrode first recorded a signal above the threshold noise value, and offset was determined as the last instance at which a signal was above the threshold noise value within a given suck. Duration was the time between onset and offset, and AUC was calculated as the summed amplitude of the electrode firing during that duration, standardized to the maximum AUC of a given electrode across both behaviors to allow for comparisons across individuals and muscles.

Anatomy

Whole-muscle anatomical data were segmented in Avizo 9.4 (FEI Visualizations Science Group, Hillsboro, OR, USA). We identified endomysium and muscle fascicles as high-density material surrounded by lower-density material (perimysium) below the minimum grayscale threshold in segmented volumes. We identified six muscles associated with feeding (genioglossus, hyoglossus, styloglossus, intrinsic tongue, digastric, geniohyoid), and exported muscles as .obj files, which were subsequently imported into Autodesk Maya (Maya 2022, Autodesk, Inc. San Rafael, CA, USA), along with one half of the lower jaw (isolated from the CT scan acquired prior to staining). We measured the 3-D volume of each muscle, and the jaw in Maya, and then standardized muscle volume to jaw volume to account for variation in animal size through ontogeny. We did not conduct statistical analyses on anatomical data due to the small sample size.

Statistical analyses

Tongue kinematics

We collected 1000 cycles for bottle feeding, and 853 cycles for drinking across three pigs. To evaluate differences in total movements between lateral and medial markers in each feeding condition along the anteroposterior axis, we used a linear mixed effects model (Bates et al. 2015), with feeding condition, location along the anteroposterior axis, and their interaction as fixed effects and individual as a random effect. We evaluated the statistical significance of this model using the Anova() function in R. We then performed planned contrast (R package emmeans, [Length et al. 2018]) and Cohen’s D (Cohen 1992) analyses to evaluate differences between lateral and medial markers during bottle feeding and drinking at a single location along the anteroposterior axis. To evaluate differences in movement between bottle feeding and drinking along the anteroposterior and dorsoventral planes within a single marker at the midline, we used mixed effects models, with the plane of excursion and the marker location as the variable of interest (i.e., anterior medial marker in dorsoventral movement), feeding condition as the fixed effect, and individual as a random effect (Bates et al. 2015). We also calculated a metric of effects size using Cohen’s D (Cohen 1992).

EMG

We collected EMG data from 127 cycles of bottle feeding and 191 cycles of drinking across three individuals. We used linear mixed effects models (Bates et al. 2015) to evaluate differences between feeding conditions within the variable of interest, using feeding condition as a fixed effect and individual as a random effect. We evaluated the statistical significance of this model using the Anova() function in R, and tested for the effects size using Cohen’s D (Cohen 1992) analyses.

Results

Tongue anatomy through infancy

There was substantive variation in how the muscles associated with feeding increased in size as infants grew into juveniles (Supplementary Table S1). Digastric and hyoglossus were both proportionally larger as juveniles than as infants (Table 1, Movie 1). Geniohyoid and intrinsic tongue were both larger in the infant sample than the juvenile sample (Table 1, Movie 1). Genioglossus and styloglossus were proportionally the same size in the infant and juvenile samples (Table 1, Movie 1).

Table 1.

Muscle volume data from DiceCT, scaled by jaw volume

	Infant scaled volume	Juvenile scaled volume	Juvenile: infant
Digastric	0.077	0.126	1.636
Genioglossus	0.126	0.103	0.817
Geniohyoid	0.039	0.019	0.487
Hyoglossus	0.016	0.030	1.875
Styloglossus	0.021	0.022	1.048
Intrinsic tongue	2.151	0.887	0.412

Bolded values indicate a muscle that is proportionally larger in the infant sample; italicized values indicate a muscle that is proportionally larger in the juvenile sample.

Tongue kinematics during suckling and drinking are minimal in the mediolateral axis

Mediolateral movement was minimal for all markers across both behaviors (∼0.17 cm, 11% of total movement across all markers). As such, we analyzed the movements of both lateral and medial tongue markers only in the anteroposterior and dorsoventral planes.

Variation in anteroposterior and dorsoventral tongue kinematics across a mediolateral axis

The lateral margins of the tongue moved little during suckling in anterior, middle, and posterior locations, and increased dramatically in their movement during drinking. This was especially notable in the anterior and middle tongue markers, for which the total sum of movement during a suck cycle resulted in less than 0.5 cm of movement. In contrast, the total movement of the lateral margins of the tongue during drinking was double those during bottle feeding in the anterior and middle parts of the tongue (more than 1.0 cm, Fig. 1). We found no difference in the excursion of lateral tongue markers in the posterior tongue between feeding conditions (Fig. 1, Supplementary Table S2). During drinking, the lateral margins of the tongue mirrored the movements of the midline of the tongue, and we focused all further analyses on the midline of the tongue.

Fig. 1.

Total excursion of lateral tongue markers during bottle feeding (gold) and drinking (blue). Black lines indicate statistically significant differences between feeding modalities, gold lines indicate statistically significant differences between tongue regions during bottle feeding, blue lines indicate statistically significant differences between tongue regions during drinking. Solid lines indicate statistically significant differences with large effect sizes, dashed lines indicate statistically significant differences with medium effect sizes.

Variation in tongue kinematics across an anteroposterior axis at the midline

The tongue also varied in how it moved during suckling and drinking along an anteroposterior axis. The anterior tongue primarily moved dorsoventrally during suckling, with a slight anteroposterior component. In contrast, during drinking, the anterior tongue underwent substantive anteroposterior motion (Fig. 2A, X² = 88.6, P < 0.001; D = −1.09), and was depressed relative to its start position rather than elevated (Fig. 2A). A similar pattern was observed in the middle tongue. During suckling, the middle tongue marker functioned primarily as a dorsoventral pump, whereas during drinking the tongue exhibited substantial movement along both the dorsoventral and anteroposterior axes. It therefore follows that both anteroposterior and dorsoventral excursions of the tongue were greater during drinking than during suckling (Fig. 2B, Supplementary Table S3). However, the posterior tongue altered its kinematics differently, whereby it increased in its dorsoventral, but not anteroposterior movement during drinking (Supplementary Table S3). Additionally, the position of the posterior tongue at the start of the cycle differed between suckling and drinking. During suckling, the tongue was at its most dorsal point at the initiation of the suck, whereas during drinking the tongue was at its most ventral position at the beginning of tongue protrusion (Fig. 2C).

Fig. 2.

Movement of the medial tongue in anterior (A, top), middle (B, middle), and posterior (C, bottom) locations. Left side indicates the mean movement of each marker in each condition ± standard error, with the start position at zero. Lines on the right plot indicate statistically significant differences between feeding conditions (solid = large effect size, dashed = medium effect size).

As a result of these differences in tongue movements, we observed variation in how the different parts of the tongue moved relative to each other along an anteroposterior axis. During suckling, the anterior and middle tongue exhibited relatively little change in their three-dimensional distance from one another (Fig. 3A). In contrast, their distance increased greatly during a drinking cycle when drinking from a bowl (Fig. 3A). However, the distance from the middle to posterior tongue during suckling vs. drinking did not differ as drastically throughout a cycle (Fig. 3B).

Fig. 3.

Three-dimensional distance between the anterior and middle (A) and middle to posterior (B) medial tongue markers ± standard error, illustrating that deformation along the anteroposterior axis varies with feeding modality.

Variation in muscle activity during suckling and drinking

We found that cycle duration during suckling was shorter than during drinking (suckling mean = 0.229 ± 0.04 s, drinking mean = 0.266 ± 0.05 s, X² = 49.9, P < 0.001, D = −0.70). Similarly, hyoglossus and genioglossus both increased firing duration during drinking compared to suckling (Fig. 4, Supplementary Table S4). Hyoglossus firing began and ended earlier in the cycle during suckling when compared to drinking (Fig. 5, Supplementary Table S4), whereas genioglossus firing began and ended later in the cycle during suckling than drinking (Fig. 5, Supplementary Table S4). Hyoglossus AUC showed no difference between drinking and suckling, whereas in genioglossus AUC was higher during drinking than suckling (Fig. 5C).

Fig. 4.

Boxplots demonstrating cycle duration (left), hyoglossus firing duration (middle), and genioglossus firing duration (right) during bottle feeding (gold) and drinking (blue). Dashed lines indicate statistically significant differences with medium effect sizes.

Fig. 5.

Muscle activity of the hyoglossus (left) and genioglossus (right) during bottle feeding (gold) and drinking (blue). (A) Onset relative to the beginning of the cycle, (B) offset relative to the beginning of the cycle, and (C) AUC standardized to the maximum AUC per electrode for each cycle. Lines on the right plot indicate statistically significant differences between feeding conditions (solid = large effect size, dashed = medium effect size).

Discussion

The anatomy, movements, and motor control of the tongue changed through ontogeny in complex ways. For example, while some muscles were largest at infancy (intrinsic tongue and geniohyoid), others did not change in their relative volume (genioglossus and styloglossus), and still others were proportionally larger following weaning (digastric and hyoglossus). Similarly, we found variation in the movements of the tongue during suckling relative to drinking across the mediolateral and anteroposterior axes. The lateral tongue displayed minimal movement during suckling, but comparatively larger movements during drinking that mirrored the change in movements of the midline of the tongue. Furthermore, there was variation in muscle activity depending on the mode of consumption, as genioglossus and hyoglossus both changed when they fired within a cycle, which is likely reflected by changes in their contractile characteristics.

Regional variation in tongue movements

There was variation in tongue function along both the mediolateral and anteroposterior axes and between suckling and drinking, which resulted in differences in tongue function depending on tongue location. During suckling, the tongue is responsible (in part) for making contact with a nipple, curving up around it to latch onto it, and to then generate suction to acquire food (Thexton et al. 1998 ; Geddes et al. 2008 ; Mayerl et al. 2020a). How this is accomplished depends on a multifaceted function of the tongue in the mediolateral axis: the lateral margins function to generate a seal, while the medial portion of the tongue is depressed to increase suction. It is therefore not surprising that the lateral margins of the tongue that make contact with the nipple move little during suckling. However, what is surprising is that this lack of movement at the lateral margins of the tongue continues all the way to the posterior portion of the tongue, which does not contact the nipple at all. In contrast, during drinking in most mammals, the tongue is protruded from the mouth to make contact with water on the ground and then is manipulated to pull that water into the mouth (Liu et al. 2009; Olson et al. 2021b). As such, it follows that during this behavior, the lateral margins of the tongue match those of the medial tongue, with extensive movements. Although this pattern of movement is clear, the modulation of the neural control of the medial and lateral parts of the tongue across the transition from suckling to drinking and eating remains poorly understood.

We also found extensive differences in tongue function along the anteroposterior axis of the tongue, indicative of variation in how the different areas of the tongue are functioning during feeding. For example, similar to several other studies, we found that the anterior and middle tongue markers functioned primarily as a dorsoventral pump during suckling (German et al. 1992; Thexton et al. 2004; Geddes et al. 2008, 2012; Mayerl et al. 2021c; Genna et al. 2021). However, during drinking those markers increased their anteroposterior movement (protrusion and retraction) in order to move the tongue in and out of the mouth to make contact with milk on the ground or in a bowl lower than the animals head, although the tip of the tongue is not always retracted into the oral cavity (Olson et al. 2021a; Thexton et al. 1998). These data suggest that the anterior and middle parts of the tongue are critical for the acquisition of liquids during both behaviors, and that the modulation of their movement is essential to successfully acquire food. In contrast, we saw a comparatively smaller change in the movements of the posterior tongue, which primarily changed in the magnitude of its movement, as well as in where it was positioned at the initiation of a cycle. This likely reflects the posterior region of the tongue’s function in transporting and swallowing food, rather than in acquiring it (Hiiemae and Crompton 1985; German et al. 1989; German and Franks 1991; Orsbon et al. 2018; Feilich et al. 2021). Too often drinking is seen as one behavior. However, drinking is a process that comprises multiple behaviors, including acquisition, intra-oral transport, and transporting liquid through the pharynx, that may have different mechanical requirements even if the liquid being ingested does not vary. Furthermore, the differences in tongue function during suckling and drinking are likely to be further increased in species that utilize lapping during drinking such as dogs, as the tongue is generally protruded even further out of the oral cavity during this behavior (Crompton and Musinsky 2011; Gart et al. 2015).

These differences in tongue movements across the anteroposterior axis produced variation in how the tongue is deformed in the transition from suckling to drinking and eating. For example, during suckling, there is little change in the distance between the anterior and middle parts of the tongue, likely due to their concerted role to pump the tongue dorsoventrally to generate suction in an antero-posterior wave (German et al. 1992; Thexton et al. 2004; Geddes et al. 2008, 2012; Mayerl et al. 2020a; Genna et al. 2021). However, this changes drastically during drinking, where the distance between markers increases for the first 70% of the cycle before coming back to resting distances. This is likely due to the anterior tongue being protruded prior to the protrusion of the middle tongue during drinking, increasing the distance between them. Thus, the movements of the tongue shift from being a dorsoventral pump to an anteroposterior protruder, highlighting the complexity of neuromotor control in the mammalian feeding apparatus (Thexton et al. 1998; Kayalioglu et al. 2007; Mayerl et al. 2021a, 2021b).

Linking tongue anatomy with kinematics and muscle firing timing

The function of the extrinsic tongue musculature in the transition from suckling to drinking relies on a changing anatomical, physiological, and functional system. However, the muscles of the tongue are not necessarily changing in similar ways, and likely have different functions. Although we had a limited sample size, we found that hyoglossus was proportionally larger in juvenile pigs than in newborn infants. In the transition from suckling to drinking, the muscle turns on and off later, increases in duration in proportion to cycle duration, but fires at a lower amplitude. We hypothesize that the lower-amplitude firing patterns during drinking likely relate to the decreased effort associated with drinking. Hyoglossus extends from the base of the hyoid into the tongue, and thus its orientation suggests a role primarily in retracting the tongue and moving it ventrally if contracting concentrically (Iskander and Sanders 2003; Orsbon et al. 2020). However, the shift in muscle firing timing to be later in the cycle during drinking suggests that the muscle likely shifts in its contractile characteristics to be primarily eccentric, rather than concentric, although data on muscle length changes would be necessary to confirm this hypothesis (Mayerl et al. 2021a, 2021b, 2022). Another hypothesis about the changing function of hyoglossus in the transition to drinking could be that the increased relative size of hyoglossus in juveniles may be related to its importance in oral processing during chewing, although we note that in humans, hyoglossus is proportionally larger during infancy than adulthood (Iskander and Sanders 2003), though this pattern may differ if juveniles are examined.

In contrast to hyoglossus, genioglossus was active earlier during drinking than suckling. Furthermore, the AUC of the genioglossus increased during drinking, likely due to the longer total firing time during drinking rather than increasing in amplitude. We anticipate that as the genioglossus is the primary muscle involved in tongue protrusion, that the increased need to protrude the tongue during drinking makes up for the reduced need to generate suction during bottle feeding (Thexton et al. 1998; Kayalioglu et al. 2007). Essentially, genioglossus function changes from powering a dorsoventral pump at the midline of the animal during suckling, to facilitating tongue protrusion during drinking (McClung and Goldberg 2000), and that in concert with this change, we see an earlier firing time coinciding with anterior movement of the middle and posterior parts of the tongue (Thexton et al. 1998). Further experiments quantifying genioglossus deformation with matching fluoromicrometry studies would confirm how this fan-shaped muscle functions in the transition from suckling to drinking (Camp et al. 2016; Mayerl et al. 2021a, 2021b, 2022). Furthermore, examining how the extrinsic muscles of the tongue function during suckling and drinking in animals that lap, such as dogs, where tongue deformation is even more disparate during the two behaviors (Crompton and Musinsky 2011; Gart et al. 2015).

Limitations and future directions

Our work represents an extension of previous work exploring how the anatomy and physiology of the feeding system in mammals change in the transition from suckling to drinking (and chewing). But this system is complex, and understanding function in a system that is more than a set of levers is challenging, and thus several questions remain unexplored. We acquired data on only two of the four extrinsic tongue muscles, and none of the intrinsic muscles. This is especially interesting because the styloglossus has been hypothesized to be responsible for curling the lateral margins of the tongue during suckling, and would therefore be expected to show strong variation in function across the transition from suckling to drinking (Iskander and Sanders 2003). Furthermore, we only collected muscle activity data, and understanding muscle function requires insight into both activity and contractile characteristics (Mayerl et al. 2022). Additionally, our data are in one species, and we expect to see variation in the transition from suckling to drinking across mammals as they transition to diverse diets, especially in regards to variation in how adults are acquiring fluids (Crompton and Musinsky 2011 ; Gart et al. 2015; Olson et al. 2021a ). Additionally, our data on the anatomical changes occurring within the feeding apparatus through ontogeny is limited by sample size. Future work should both expand on the number of individuals being analyzed, as well as expand on the possibility for disparate patterns of anatomical change to occur in different species. Finally, we explored these relationships only in healthy individuals, yet feeding difficulties are common in infants. Exploring how infants with feeding challenges, such as those born prematurely, deal with the transition to drinking and eating, and what the long-term consequences of this transition may be, remains completely unexplored.

Conclusions

The tongue must dramatically change its function in the transition from suckling to drinking in mammals. While the overall patterns underlying these differences have been holistically described, most previous research has focused on the tongue as a whole structure, rather than by investigating the components that make it up. Here, we show that while the overall patterns of tongue movement are predictable, with an increase in the anteroposterior movement of the tongue, the way that those movements change is variable, as is the anatomy and neuromotor control of the tongue. Furthermore, we demonstrate that, at least in pigs, the anatomy of the feeding system does not scale in a simple manner, but instead, different muscles scale differently. Understanding how patterns in anatomy, physiology, and movements are reflected by changes in function represents a promising avenue for future research. Our data highlight the significant need for high neuroplasticity in the control of the tongue at a young age in mammals and suggest that the ability to do so is key in the ontogeny and evolution of feeding in these animals.

Notes

From the symposium “Biology at birth: the role of infancy in providing the foundation for lifetime success” presented at the annual meeting of the Society for Integrative and Comparative Biology virtual annual meeting, January 16–March 31, 2023.

Author contributions

Conceptualization: K.E.S., C.J.M.; Experimental design: K.E.S., R.Z.G., S.P.H., K.A., and C.J.M.; data collection: All authors; Data processing: K.E.S., M.L.J., L.E.B., A.K., K.C.N., R.Z.G., and C.J.M.; Manuscript writing: K.E.S., C.J.M.; Manuscript editing: all authors.

Funding

This work was supported by NIH NICHD K99HD101588 to CJM and NIH NICHD R01HD09688102 to RZG.

Conflict of interest

We declare no competing interests.

Data availability

All data used in statistical analyses are available on figshare: https://doi.org/10.6084/m9.figshare.22151978.v1.