Prolonged use of a soft diet during early growth and development alters feeding behavior and chewing kinematics in a young animal model

Abstract

In infants and children with feeding and swallowing issues, modifying solid foods to form a liquid or puree is used to ensure adequate growth and nutrition. However, the behavioral and neurophysiological effects of prolonged use of this intervention during critical periods of postnatal oral skill development have not been systematically examined, although substantial anecdotal evidence suggests it negatively impacts downstream feeding motor and coordination skills, possibly due to immature sensorimotor development. Using an established animal model for infant and juvenile feeding physiology, we leverage X-ray Reconstruction Of Moving Morphology to compare feeding behavior and kinematics between 12-week old pigs reared on solid chow (control) and an age- and sex-matched cohort raised on the same chow softened to a liquid. When feeding on two novel foods, almond and apple, maintenance on a soft diet decreases gape cycle duration resulting in a higher chewing frequency. When feeding on almonds, pigs in this group spent less time ingesting foods compared to controls, and chewing cycles were characterized by less jaw rotation about a dorsoventral axis (yaw) necessary for food reduction. There was also a reduced tendency to alternate chewing side with every chew during almond chewing, a behavioral pattern typical of pigs. These more pronounced impacts on behavior and kinematics during feeding on almonds, a tougher and stiffer food than apples, suggest that food properties mediate the behavioral and physiological impacts of early texture modification, and that the ability to adapt to different food properties may be underdeveloped. In contrast, the limited effects of food texture modification on apple chewing indicate that such intervention/treatment does not alter feeding behavior of less challenging foods. Observed differences cannot be attributed to morphology because texture modification over the treatment period had limited impact on craniodental growth. Short-term impacts of soft-texture modification during postweaning development on feeding dynamics should be considered as potential negative outcomes of this treatment strategy.

Graphical Abstract

The prolonged use of softened or texture modified foods during postweaning oromotor maturation is often utilized in children with feeding and swallowing difficulties. Using an animal model, we evaluate the impacts of this treatment on feeding dynamics when eating novel foods. Compared to pigs reared on a normal diet, pigs reared on the same diet in liquidized form chewed faster, utilized less jaw yaw, spent less time ingesting, and reduced the frequency of side-switching between chews.

INTRODUCTION

Postweaning growth and development in most mammals involves a transitional period during which behavioral, physiological, functional, and biomechanical feeding dynamics change. Sue Herring’s extensive foundational work on the development of mammalian feeding provides an important lens into these changes. In particular, this work emphasizes the relationship between morphology and the maturation of feeding motor systems, neuromuscular control of jaw muscles, and at a gross level, the movements they produce during chewing (Herring & Scapino, 1973; Herring, 1977, 1985b; Herring & Wineski, 1986; Huang et al., 1994; Lakars & Herring, 1980). An important finding of this work is that increasing complexity of behavior and jaw muscle function is associated more so with neural development than musculoskeletal development (Herring, 1985a; Herring & Wineski, 1986; Lakars & Herring, 1980). Additional insights from other species demonstrate species-specific changes in the motor control of jaw movements (Green et al., 1997; Huang et al., 1994; Iinuma et al., 1991; Lakars & Herring, 1980; Langenbach et al., 2001; Weijs & Brugman, 1989; Westneat & Hall, 1992; Williams et al., 2010; Yamada et al., 2023), either through assessment of jaw muscle activity or its output as kinematics or behavior.

Inherent in this complexity of behavior and motor control is the development of the ability to modulate the response of the jaw closing muscles producing force and movement during feeding, particularly as animals shift from liquids to solids and more complex foods. Modulation is an important component of effective and safe feeding as the bolus is prepared for swallowing. For feeding on solid foods, the oral phases are especially complex and involve constantly changing interactions between the bolus and the teeth, cheeks, and tongue to produce a swallowable bolus (e.g., Dutra et al., 2010; Hiiemae et al., 1995; Liu et al., 2007; Olson, Montuelle, Chadwell, et al., 2021; Olson, Montuelle, Curtis, et al., 2021). Furthermore, the timing of bolus transport out of the oral cavity through the pharynx to the esophagus during the swallow must be coordinated with laryngeal closure and respiration (Martin-Harris, 2006; Martin-Harris et al., 2003). In adult feeding systems, modulation of these processes from oral stages to pharyngeal swallowing occurs through complex sensorimotor integration pathways within and between stages of feeding (Humbert et al., 2012; Martin, 2009; Pouderoux & Kahrilas, 1995; Steele & Miller, 2010; Trulsson, 2006).

The coordination of these stages is influenced by food texture, which has become an increasingly important property for understanding physiological processes related to feeding (Koc et al., 2013, Nishinari & Fang, 2018). Food texture refers to “all the mechanical, geometrical and surface attributes of [food] perceptible by means of mechanical, tactile and, where appropriate, visual and auditory responses” (ISO 1992). In essence, it reflects sensory properties of foods (Koc et al., 2013).

The foundations for sensorimotor pathways that enable the complex interaction between food texture and oromotor control are laid during early critical periods of development. Furthermore, oral sensory stimulation is thought to be essential during this time (Arvedson, 2006; Delaney & Arvedson, 2008). Critical periods are an important window of opportunity for normal development resulting from the interplay between emerging behaviors, the maturing central nervous system, and environmental influences (Cioni & Sgandurra, 2013; Purves et al., 2001; Sengpiel, 2007). In human infants, this critical period seems to start around 6 months, when rapid anatomical, neuromotor, and sensorimotor maturation occurs related to the shift from suckling to solid foods (Delaney & Arvedson, 2008). Hallmarks of this development are multifactorial, observed as changes in the kinematic of jaw movements, improved chewing performance, and shifts in behavior related to food handling acceptance. For example, infants as early as 7 months respond to different food textures by changing the rate of jaw movement (Sheppard & Mysak, 1984; Wilson & Green, 2009). At 8 months, exposure to larger foods and foods of different textures improves chewing ability by 4 weeks (Costa et al., 2017). On the other hand, delays in introducing heterogeneous food textures during this critical period seem to impact motor and behavioral development. Infants who cannot feed on solid foods fail to develop lateral tongue movements (Mason et al., 2005; Reilly et al., 1995), and children with delayed exposure to texturally heterogeneous (i.e., lumpy) foods after 10 months of age incorporate fewer solids at 15 months (Northstone et al., 2001) and demonstrate continued feeding challenges through 7 years (Coulthard et al., 2009).

Because maintaining adequate nutritional and caloric intake is essential to overall growth and development, infants facing difficulties with the transition to solid foods may be prescribed a texture modified diet of thickened liquids, or mashed or blended liquidized foods requiring little or no oral stage bolus preparation. Texture modification typically results in softened or liquidized foods from a sensory perspective. When prescribed during or for prolonged periods after the critical period, there may be negative impacts in as little as 3 weeks on developing oromotor skills and oral sensitivity, with additional long-term impacts, even in otherwise low-risk infants (Arvedson, 2008; Delaney & Arvedson, 2008; Dodrill et al., 2004; Mason et al., 2005). Whereas many studies have used soft and hard diets to understand how loading environments during growth impact the musculoskeletal system in select animal models (e.g., Ciochon et al., 1997; He & Kiliaridis, 2003; Larsson et al., 2005; Menegaz et al., 2008, 2009; Tanaka et al., 2007), including pigs, we know far less about the impact of texture modification, hereafter referred to as soft diet, on the development of oromotor behavior and function.

In the present study, we use the pig, a model well-established by Herring and others for understanding postweaning feeding dynamics (e.g., Huang et al., 1994; Herring & Scapino, 1973; Herring, 1985, 1977), to investigate how long-term use of a soft diet during a critical period of oromotor development impacts the dynamic response to novel foods that have different textures due to differences in initial mechanical properties. We compare feeding in juvenile pigs reared on a soft diet to pigs reared on a standard pellet diet to better understand this dynamic response on two levels: 1) the duration, frequency, and types of feeding cycles throughout the series of chewing cycles leading to a swallow (albeit not the terminal swallow), and 2) the kinematics of jaw movements within each cycle (i.e., opening-closing and mediolateral deviation during tooth-food-tooth contact). Our overarching hypothesis is that rearing conditions will result in differences in behavior and kinematics across both levels when pigs are tasked with feeding on novel solid foods.

METHODS

Study design.

Eight pigs were used for this study that lasted from when they were 3–16 weeks old. Behaviorally, pigs at 2–3 weeks are comparable to human infants 6–12 months. By 8 weeks, pigs have multiple bunodont teeth in occlusion and are typically fully weaned (Herring, 1985; Huang et al., 1994). This is dentally and behaviorally comparable to a 23–31 month-old child. At this stage of oromotor development, both humans and pigs should be independently eating a variety of foods with different textures (Delaney and Arvedson, 2008; Herring, 1985b).

Throughout the experiment, all animals were provisioned the same amount of food per kg body weight daily, had ad libitum access to water, and were offered the same daily enrichment opportunities. Body weight was monitored throughout the study to determine feeding requirements and to ensure consistent weight gain. Initially, all 3-week old pigs were fed a LabDiet Starter pellet (St Louis, MO, USA) pulverized and mixed with milk-replacer for 2 weeks. At the start of week 6, pigs were split into a control and treatment (soft diet) group, each with two males and two females. Up until data collection at week 12, the control group was maintained on standard Labdiet Grower pellets (St. Louis, MO, USA). The treatment group was maintained on pulverized Labdiet Grower pellets at ratio of 100g of pellet to 240mL water to obtain a homogenous “liquidized” texture reflecting level 3 of the International Dysphagia Diet Standardization Initiative (IDDSI) (Cichero et al., 2017; Complete IDDSI Framework Detailed Definitions 2.0, www.Iddsi.Org, 2019). IDDSI levels are mainly defined qualitatively, with level 3 referring to “moderately thick” but soft texture that (i) drips slowly in dollops through the prongs of a fork, or (ii) no less than 8mL should remain in a 10mL syringe after 10 seconds of flow.

Marker implantation.

Following marker-based XROMM protocols (Brainerd et al., 2010; Montuelle et al., 2018), 5 to 6 1.6 mm radiopaque markers were surgically implanted in the skull and lower jaw of all individuals while they were under general isoflurane anesthesia. After surgery, the animals were given a single intramuscular injection of ketoprofen (2 mg/kg), and were allowed a week-long recovery period before data collection.

Experiments.

Apple and almond were selected as experimental foods because of their difference in initial hardness and toughness; i.e., apples: low stiffness and low toughness versus almonds: high stiffness and high toughness (Williams et al., 2005). At the start of week 12, almonds and cubes of apple of similar size (approximately 2 cm x 2 cm x 2cm) were presented to the animals in a bowl positioned in the field of view of two synchronized Oqus 310 cameras (Qualisys, Goteborg, Sweden) mounted on the output ports of two synchronized OEC-9000 fluoroscopes (General Electrics, Boston, MA, USA). Sequences were recorded at 250 frames per second for a total duration of 25 seconds, with radiation doses averaging 90 kVp and 4.5mA across trials. All feeding sequences for each animal were recorded in a single day across 5 or fewer sequences per food. Recording all the data for an animal within a single day limits the potential for motor learning to ensure a novel response to foods. Prior to and following recording sessions, a distortion grid was imaged to correct distortion inherent to x-ray imaging, and a custom calibration cube was imaged to calculate the 3D space in the field of view. After data collection, the animals were CT-scanned while under isoflurane anesthesia at The Ohio State University College of Veterinary Medicine (Columbus, OH, USA) on a GE Lightspeed Ultra CT scanner (General Electric, Boston, MA, USA). The CT scans were processed in VGSTUDIO Max (version 3.3, Volume Graphics GmbH, Heidelberg, Germany) to generate virtual 3D meshes for each of the skull and jaw and for determining the 3D coordinates of each marker centroids. All procedures involving live animals were approved by the Ohio University Institutional Animal Care and Use Committee (protocol #19-U-002).

Morphological measurements.

Skull and jaw measurements were acquired in VGSTUDIO Max from the reconstructed CT scans taken at week 12. These measurements are illustrated in Figure 1 and summarized in Supplemental Table 1. In addition, from the reconstructed CT scans, we determined the occlusal stage as the number of cheek teeth in occlusion per quadrant for each animal following (Huang et al., 1994): Stage 1, 1 in occlusion; Stage 2, 2–3 in occlusion; Stage 3; 4–5 in occlusion. During these scans, animals were masked and jaws were placed in centric relation, but confirmation of occlusal stage occurred during visualization of the reconstructed jaw kinematics as well (see below). Given their ages, we expect all animals to have 2–3 cheek teeth in occlusion (Stage 2).

Figure 1. Skull and jaw measurements.

A) Skull length (SL); skull width (SW); and bizygomatic width (BW). B) Palate length (PL). C) Skull height (SH). D) Jaw length (JL); bicondylar breadth (BB); condyle width (CW). E) Tooth row length (TRL); corpus depth (CD); ramus height (RH). F) Symphysis width (SyW) and length (SyL) in cross section. All scale bars are 1cm.

Behavioral data.

From the fluoroscopy videos, we identified swallows by observing the video frame at which the epiglottis folds backwards over the airway. Maximum jaw opening was used to approximate the start and end of each gape cycle. Each gape cycle was categorized as “ingestion” if the tongue protracted outside the mouth to contact a food item inside the bowl, or as a “chew” if the tongue did not contact a food item within the bowl. Because pigs transport and chew food particles within the same gape cycle, separating between these two intra-oral behaviors was not possible in a standardized process, so cycles coded as “chews” may include strict chewing cycles, strict transport cycles, as well as cycles combining chewing and transport. We extracted the mean ± standard deviation for the following behavioral variables using only sequences where a swallow was observed: number of cycles before a swallow; number of ingestion cycles before a swallow; number of chews before a swallow; and ingestion:chew cycle percentage ratio. We also calculated the time between two consecutive swallows; the total time spent ingesting between two swallows; the total time spent chewing ingesting between two swallows; the percentage of time spent chewing or ingesting between two swallows; all cycle frequency; chewing cycle frequency; and ingestion cycle frequency.

X-ray reconstruction Of Moving Morphology (XROMM).

Fluoroscopy videos were processed in XMALab (Knorlein et al., 2016) to generate filtered (low-pass Butterworth, 25 Hz cut-off frequency) rigid body transformations of the skull and the jaw. Note that only the chewing cycles were digitized, and thus the XROMM analysis focuses on chewing. The rigid body transformations were used to animate the corresponding 3D bone models in Maya (Autodesk Inc., San Rafael, CA, USA). Based on the XROMM workflow (Brainerd et al., 2010; Menegaz et al., 2015) and our previous studies (Montuelle et al., 2018, 2020), a 3-axis Joint Coordinate System (JCS) was set up to quantify jaw movements with respect to the skull. The JCS measures rotational and translational degrees of freedom according to three anatomical axes: (x) anteroposterior, (y) dorsoventral, and (z) mediolateral. Previous XROMM studies of pig feeding demonstrate that they only use two rotational degrees of freedom: jaw pitch (Rz), representing jaw opening-closing, and jaw yaw (Ry), representing the mediolateral deviation of the jaw during the power stroke. Note that the third degree of freedom of jaw movements during pig chewing is condylar protraction-retraction (Tx), but because of its close anatomical correlation with jaw pitch (Brainerd et al., 2010; Menegaz et al., 2015), it can be expected to follow a similar pattern as jaw opening-closing, and thus was not investigated further. Figure 2 shows a representative set of kinematic waves corresponding to jaw pitch and jaw yaw, as well as the two kinematic variables extracted: amplitude of jaw opening and jaw yaw (i.e., difference between minimum and maximum jaw Rz and Ry positions, respectively).

Figure 2. Jaw pitch (Rz, blue) and jaw yaw (Ry, green) during a single chewing cycle.

Arrows represent the maximum amplitude of jaw pitch and of jaw yaw (blue and green, respectively) measured in that cycle. Decrease in Rz (blue arrow) indicates jaw depression, thus gape opening. Increase in Ry (green arrow) reflects jaw deviating to the left, thus a right chew, whereas a decrease in Ry represents a left chew.

In addition, chewing side was determined from the direction of jaw yaw during each cycle. An increase in Ry reflects jaw deviating to the left, thus a right chew whereas a decrease in Ry represents a left chew. Because pigs tend to alternate chewing side with every chew (Herring & Scapino, 1973), the frequency of chewing side switches was quantified from portions of each sequence in which rhythmic chews were observed following the initial chew. Alternating chewing side at every chew would translate to 100% of chewing cycles being side switches, and a score near 0% would indicate the individual rarely switches chewing side.

Statistical analysis.

Morphological variables were compared between groups using 2-tailed t-tests. Differences between groups in behavioral and kinematic variables were tested using a one-way ANOVA coupled with F-ratios with the treatment entered as a fixed factor. The individual factor was also assessed within each group separately, and reported in Supplemental Tables 2–3. Finally, the frequency of side switches was compared using a chi-square test. All analyses were performed in R (version 4–3-3). Sample sizes for behavioral and kinematic datasets are provided in Table 1. The data that supports the findings of this study are available in the supplementary material of this article.

Table 1.

Sample sizes for behavioral and kinematic datasets

		$\frac{\mathbf{\text{Control diet}}}{(\text{4 individuals})}$	$\frac{\mathbf{\text{Soft diet}}}{(\text{4 individuals})}$
Behavior analysis
Almond	Nsequences Nswallows Ncycles	9 (2, 4, 2, 1) 15 (2, 6, 5, 2) 477 (114, 204, 120, 39)	7 (2, 3, 1, 1) 14 (5, 3, 1, 5) 381 (118, 119, 84, 60)
Apple	Nsequences Nswallows Ncycles	8 (2, 2, 2, 2) 30 (8, 7, 9, 6) 463 (119, 121, 139, 84)	7 (2, 2, 2, 1) 29 (8, 7, 8, 6) 502 (152, 152, 123, 75)
Kinematic analysis
Almond	Ncycles	169 (41, 74, 35, 19)	160 (47, 37, 33, 43)
Apple	Ncycles	150 (45, 19, 31, 55)	258 (87, 61, 57, 53)

Numbers in parentheses represent the sample for each individual.

RESULTS

Morphology

There were no significant differences in any of the morphological measurements, absolute or scaled, between the two groups (Table 2). The soft diet group was overall characterized by a taller and wider skull, with a longer palate, as well as a shorter and wider jaw with a shorter tooth row; but these differences were subtle and nominal. Occlusal stage was similar in both groups, but in the control group, all animals had 3 cheek teeth in occlusion, whereas in the soft diet group, 2 of the 4 animals only had 2 cheek teeth in occlusion. This likely drives the (non-significant) differences in tooth row length between the groups, as only occluding cheek teeth were measured.

Table 2. Summary of morphological measurements.

Absolute dimensions (top) and relative dimensions scaled to jaw length (bottom) are presented in each cell (except jaw length and body weight).

	Control diet	Soft diet	T-test p-value
Skull length (mm)	146.64 ± 5.68 1.19 ± 0.05	146.48 ± 6.20 1.21 ± 0.05	0.968
Skull height (mm)	40.50 ± 0.98 0.33 ± 0.01	41.40 ± 4.39 0.34 ± 0.01	0.703
Skull width (mm)	57.61 ± 2.29 0.47 ± 0.02	58.69 ± 0.70 0.48 ± 0.03	0.406
Bizygomatic width (mm)	84.63 ± 2.02 0.69 ± 0.02	84.91 ± 2.52 0.70 ± 0.03	0.868
Palate length (mm)	84.84 ± 1.71 0.69 ± 0.02	86.49 ± 4.37 0.71 ± 0.02	0.507
Jaw length (mm)	123.31 ± 2.43	121.73 ± 8.69	0.738
Bicondylar breadth (mm)	85.82 ± 4.69 0.70 ± 0.04	87.52 ± 2.81 0.72 ± 0.03	0.557
Condyle width (mm)	16.73 ± 1.22 0.14 ± 0.01	17.51 ± 1.74 0.14 ± 4e−3	0.492
Tooth row length (mm)	36.27 ± 0.83 0.29 ± 0.10	32.40 ± 6.06 0.27 ± 0.04	0.253
Corpus depth (mm)	22.24 ± 1.11 0.18 ± 0.01	22.70 ± 2.37 0.19 ± 0.01	0.736
Coronoid height (mm)	60.09 ± 3.49 0.49 ± 0.02	59.94 ± 5.83 0.49 ± 0.02	0.967
Symphysis length (mm)	35.26 ± 0.90 0.29 ± 0.01	34.69 ± 3.63 0.28 ± 0.01	0.770
Symphysis width (mm)	11.90 ± 0.58 0.10 ± 3e−3	11.98 ± 1.30 0.10 ± 4e−3	0.917
Body Weight (kg)	13.89 ± 4.57	15.32 ± 7.54	0.502
Occlusal status	Stage 2	Stage 2	-

Table entries are mean ± standard deviation.

Behavior

Overall the effects of the soft diet on feeding behavior were limited (Table 3). Compared to the control group, there was no significant difference in the time between 2 consecutive swallows. In addition, the number and frequency of gape cycles between 2 swallows were also not significantly different between the groups. The relative time allocation (i.e., % of time spent ingesting or chewing) was also similar between the groups. The only difference observed during both apple and almond feeding was a significant increase in chewing cycle frequency (p = 0.03 for almond chewing, p < 0.001 for apple chewing); albeit with a large effect size (see Table 3). All other differences between groups were observed for almond feeding only. Compared to controls, soft diet pigs spent less time ingesting during almond feeding (4.11 sec vs 2.11 sec, respectively; p = 0.017), but the time spent chewing was similar between groups. Finally, pigs reared on soft diets significantly decreased the frequency of side switching during almond chewing (69.0% in control vs 56.5% in treatment; p = 0.02), with no differences for apple chewing (Table 3).

Table 3.

Summary of behavioral data.

	Food	Control diet	Modified diet	Treatment F-ratio and Significance	Effect size (eta2, Cohen’s f)
Time between 2 swallows (sec)	almond	9.10 ± 3.68	7.44 ± 4.11	F1,26 = 1.271, p NS	η2 = 0.13, f = 0.39
	apple	4.55 ± 1.13	4.98 ± 1.63	F1,57 = 1.373, p NS	η2 = 0.09, f = 0.31
# of cycles between 2 swallows	almond	31.8 ± 15.0	22.8 ± 11.9	F1,26 = 3.007, p NS	η2 = 0.11, f = 0.35
	apple	15.4 ± 4.4	17.3 ± 5.9	F1,57 = 1.933, p NS	η2 = 0.13, f = 0.38
Chew:ingestion cycle ratio (%)	almond	54.5 / 45.5 ± 22.2	65.1 / 34.9 ± 18.1	F1,27 = 1.958, p NS	η2 = 0.07, f = 0.27
	apple	64.2 / 35.8 ± 25.6	67.3 / 32.7 ± 22.0	F1,57 = 0.247, p NS	η2 = 4e−3, f = 0.07
Time spent ingesting (sec)	almond	4.11 ± 2.49	2.11 ± 1.47	F1,26 = 6.472, p = 0.017	η2 = 0.20, f = 0.50
	apple	1.56 ± 1.21	1.62 ± 1.18	F1,57 = 0.046, p NS	η2 = 8e−4, f = 0.03
Time spent chewing (sec)	almond	5.59 ± 3.46	5.05 ± 2.92	F1,27 = 0.207, p NS	η2 = 2.5e−4, f = 5e−3
	apple	3.00 ± 1.27	3.36 ± 1.47	F1,57 = 1.015, p NS	η2 = 0.02, f = 0.13
Chew:ingestion time (% time)	almond	56.1 / 43.9 ± 22.6	64.6 / 35.4 ± 18.9	F1,27 = 1.176, p NS	η2 = 0.04, f = 0.21
	apple	66.2 / 33.8 ± 25.0	67.0 / 33.0 ± 21.5	F1,57 = 0.019, p NS	η2 = 3.4e−4, f = 0.02
All cycle frequency (cycles/sec)	almond	3.24 ± 0.20	3.32 ± 0.26	F1,27 = 0.915, p NS	η2 = 0.06, f = 0.25
	apple	3.37 ± 0.26	3.47 ± 0.24	F1,57 = 2.441, p NS	η2 = 0.04, f = 0.21
Ingestion cycle frequency (cycles/sec)	almond	3.45 ± 0.34	3.33 ± 0.36	F1,25 = 0.812, p NS	η2 = 0.04, f = 0.20
	apple	3.65 ± 0.36	3.43 ± 0.46	F1,52 = 3.829, p NS	η2 = 0.33, f = 0.71
Chewing frequency (cycles/sec)	almond	3.14 ± 0.19	3.37 ± 0.33	F1,27 = 5.267, p = 0.03	η2 = 0.16, f = 0.44
	apple	3.23 ± 0.26	3.48 ± 0.25	F1,56 = 13.770, p < 0.001	η2 = 0.59, f = 1.21
Side switch (% chews)	almond	69.0 % ± 0.2	56.5 % ± 0.1	𝝌 = 5.24, p = 0.02	-
	apple	58.0 % ± 0.1	66.0 % ± 0.1	𝝌 = 2.28, p NS	-

Table entries are mean ± standard deviation. Bold entries highlight significant differences between groups. Effect size can be considered large at η2 > 0.14.

Jaw kinematics

In the soft diet group, chewing cycle duration was significantly shorter compared to the control pigs (Table 4; Figure 3). This effect was observed during both almond (F_1,327 = 23.22, p < 0.001) and apple chewing (F_1,406 = 4.34, p = 0.04), although there was also a large effect size. The amplitude of jaw opening was not significantly different between control and soft diet pigs for both foods (Table 4). However, in the soft diet group, Rz was more variable throughout the gape cycle during almond chewing (sd_control = 1.45 vs sd_treatment = 2.49) and apple chewing (sd_control = 1.82 vs sd_treatment = 2.90) (Figure 4). Finally, soft diet significantly alters the amplitude of jaw yaw during the power stroke for almond but not apple chewing (Table 4; Figure 5).

Table 4.

Summary of kinematic data.

	Food	Control diet	Soft diet	Treatment F-ratio and Significance	Effect size (eta2, Cohen’s f)
Cycle duration (msec)	almond	333 ± 39	311 ± 43	F1,327 = 23.22 P < 0.001	η2 = 0.18, f = 0.46
	apple	316 ± 38	308 ± 36	F1,406 = 4.34 P = 0.04	η2 = 0.11, f = 0.35
Opening amplitude (degrees)	almond	15.5 ± 4.2	15.4 ± 5.1	NS	η2 = 1e-3, f = 0.04
	apple	17.9 ± 3.7	17.7 ± 4.3	NS	η2 = 0.03 f = 0.19
Yaw amplitude (degrees)	almond	2.57 ± 1.53	2.16 ± 1.44	F1,327 = 6.348 P = 0.01	η2 = 0.02, f = 0.13
	apple	2.19 ± 1.09	2.41 ± 1.35	NS	η2 = 0.09 f = 0.31

Table entries are mean ± standard deviation. Bold entries highlight significant differences between groups. Effect size can be considered large at η2 > 0.14.

Figure 3. Chewing cycle duration is shorter in soft diet pigs.

Asterisks indicate significant differences at p<0.05.

Figure 4. Jaw pitch (Rz) throughout the gape cycle is more variable in soft diet pigs during almond chewing (sd_control = 1.45 vs sd_treatment = 2.49) and apple chewing (sd_control = 1.82 vs sd_treatment = 2.90).

Average change in Rz (± standard deviation) over the time-standardized gape cycle are presented for almond (A) and apple (B) for the control (grey) and soft diet (red) datasets.

Figure 5. Jaw yaw amplitude in soft diet pigs is reduced during almond chewing.

Asterisks indicate significant differences at p<0.05.

DISCUSSION

Soft-texture modification is often implemented during a critical period of development of oromotor skills and behaviors to address issues with infant or child feeding of known and unknown etiology. While soft-texture modification is essential to safely ensuring adequate growth and nutrition in the short term, less is known about the behavioral and neurophysiological effects of this intervention on the dynamics and development of feeding behaviors. Our data show that during a limited period of growth, soft-texture modification alters some but not all aspects of feeding behavior and chewing kinematics in an animal model, and thus should be considered as potential outcomes for children as well.

The observed effects of prolonged maintenance on a soft diet occur in the absence of differences in craniodental or body size (see Table 2). Size differences would suggest more systemic and perhaps nutritional differences that could impact metabolic regulation and potentially skeletal form (Labouré et al., 2001). However, their absence in our sample makes functional or biomechanical differences that could alter load generation or resistance, unlikely factors contributing to our behavioral observations. This is one potential benefit of the relatively limited treatment window used in the present study as morphological differences associated with soft diets seem to be more pronounced after longer periods of growth across species, which could then exacerbate the effects on feeding (e.g., Burn et al., 2010; Ciochon et al., 1997; Larsson et al., 2005; Menegaz et al., 2009).

The observed changes, instead, highlight the possibility that relatively short-term postweaning use of a soft diet alters behavioral and functional characteristics of ingestion and chewing. Compared to the control group, the soft diet group chewed, on average, at a higher frequency due to shorter chewing cycles when feeding on both novel foods. Increase in chewing frequency is a factor contributing to the development of chewing post-weaning in pigs (Huang et al., 1994), as well as in rats (Westneat & Hall, 1992) and rabbits (Langenbach et al., 2001), and our data shows that food texture modification amplifies that trend. Although the differences appear subtle, they suggest that animals who do not undergo timely suckling-chewing transition to hard or textured foods, as in the soft diet group, may engage in more cursory bolus preparation. Whether this is due to neurobehavioral differences related to central pattern generation or to neurophysiological differences related to long-term effects on muscle properties is unclear. In the latter case, the results align with soft and hard diet comparisons in peri- and post-weaning rats and rabbits (e.g., Kawai et al., 2010; Kiliaridis et al., 1988; Langenbach et al., 2003; Taylor et al., 2006; Saito et al., 2002). These studies show that prolonged reduction in masticatory loading alters masseter muscle contractile properties, including fiber and physiological cross-sectional area and fiber type composition. Furthermore, consumption of a soft diet through weaning in these animal models results in jaw closing muscles with a higher proportion of fast-contracting fibers. If the same occurs in pigs, this may explain why the soft diet group had higher chewing frequencies.

The behavioral effects of a soft diet in pigs are overall very limited during apple chewing, but somewhat more pronounced during almond chewing. Given that almonds are stiffer and tougher food than apples (Williams et al., 2005), the fact that apple chewing is virtually identical between the control and the soft diet group (except for an increase in chewing frequency) shows that soft-diet intervention does not affect feeding on less challenging foods. In comparison, almond chewing in pigs reared on a soft diet is characterized by about a 50% reduction of the time allocated to ingestion, but not chewing, between two consecutive swallows. This suggests potentially more proficient handling of almonds during Stage 1 transport, or a reduced capability to create a manageable bolus. This may also be due to physical differences between almond and apple, especially their density which may impact ingestion, or their mechanical differences which may impact reduction and bolus formation.

In addition, alternating chewing sides with every chew is a key behavioral component of chewing established in stage-2 pigs (Herring & Scapino, 1973; Huang et al., 1994). Here, we observed that the frequency of side-switching behavior is significantly reduced during almond chewing in the soft diet group (69.0% in controls versus 56.5% in the modified diet group). Our results show that texture modification delays the development of this prominent chewing pattern in pigs. This further highlights potential deficits in intraoral bolus management and formation of tougher and stiffer foods and also potential impacts on higher level pattern generation to coordinate jaw-tongue movements as well.

The effects of soft-texture modification are also observed in jaw kinematics contributing to the gape cycle. Of the two rotational degrees of freedom used during chewing in pigs, pitch and yaw (Brainerd et al., 2010; Menegaz et al., 2015; Montuelle et al., 2018, 2020), only the amplitude of jaw yaw is significantly impacted by a prolonged soft diet at the cycle level, albeit only during almond chewing. A decrease in yaw during the chewing cycle suggests a more orthal gape cycle in the texture modified group that could negatively affect occlusal dynamics by reducing the grinding motion that is necessary to break down food. On the other hand, it may be that by not routinely being exposed to foods that need to be chewed, soft diet pigs have not adequately developed the wear facets to engage teeth in this way. Several studies have shown that prolonged use of a soft diet alters occlusal relations, morphology and wear (e.g., Ciochon et al., 1997, 1997; Renaud & Ledevin, 2017; Teaford & Oyen, 1989), all which can affect tooth-tooth interactions and thus, jaw movements (Laird, 2017; Teaford & Byrd, 1989). A cursory examination of the teeth of our animals shows minimal differences in occlusal relations between groups, likely due to the relatively short time period for the modified diet. Indeed, the marked changes that occurred in a previous comparison in pigs (Ciochon et al., 1997) was the result of 7 months of a softened diet. Nevertheless, a more focused analysis of occlusal relations, topography, and wear in relation to occlusal dynamics is needed to assess the implications for food reduction.

On the other hand, jaw opening-closing amplitude is similar between the control pigs and those reared on a soft diet, which mirrors the limited changes in jaw opening amplitude observed during pig growth and development (Huang et al., 1994). Jaw opening-closing movements are complex and composed of multiple intra-cycle phases based on the velocity and acceleration of jaw pitch position. These within-cycle kinematics contribute to important aspects of mammalian chewing including responding to food properties, maturation of chewing behavior, and maintaining cycle frequency (e.g., Montuelle et al., 2018, 2020; Reed & Ross, 2010). The fact that our analysis did not detect any effect of the soft-diet intervention on cycle-level amplitudes of jaw opening does not exclude treatment effects on other kinematic characteristics within the gape cycle (e.g., phase duration, relative amplitude, average velocity, and their respective variability).

Both groups of pigs chewed on average slower (longer cycle duration) on the same foods than reported previously (e.g., Montuelle et al., 2018). This could be due to the younger age of the present study: data for this study were collected right at 12 weeks, compared to previous data recorded closer to 16 weeks. Additionally, both experimental foods (almonds and apples) were novel foods to all pigs in the present study, whereas pigs in other studies were exposed to these food types routinely. Previously, chewing movements in pigs have been documented to include more jaw yaw when they chew on brittle and tough foods like almonds, than when chewing on apples (Montuelle et al., 2020), emphasizing that this kinematic parameter in particular is important when responding to increase in food toughness and stiffness. Therefore, reduction of jaw yaw during almond chewing in pigs reared on a soft diet may reflect a reduced ability to respond to changes in food properties which is critical to safe and efficient mastication. While a within-group analysis of between-foods differences would answer this point directly, the notable differences between the results for apple and almond suggest that food properties, such as toughness and stiffness, mediate the behavioral and physiological impacts of a soft diet intervention when animals consume novel foods. As the ability to adjust behavior and movements to novel foods is critical for chewing, ensuring that this skill is developed is essential for increasing dietary breath. This supports the idea that early sensorimotor learning opportunities are essential to enable adaptive responses to novel oral challenges necessary to effectively wean and transition to solid foods.

The next step is to understand whether, and if so how, perturbations to oral function as observed here alter bolus transport into the pharynx and beyond. In doing so, we may be able to identify behavioral and kinematic outcomes in the oral phase that are potential predictors of downstream performance issues on swallowing. Nevertheless, the observed effects in oral phase dynamics over a relatively short treatment time period in the absence of an impact on morphology suggest that understanding oral sensorimotor development and designing sensorimotor interventions are integral to treatment strategies for pediatric feeding.

Research highlight.

Sustained maintenance on a liquidized diet during early postweaning stages alters chewing cycle frequency, ingestion time, jaw kinematics and the frequency of side switching during chewing, despite having no impact on craniodental morphology.