Food Transport and Bolus Formation during Complete FeedingSequences on Foods of Different Initial Consistency
Karen M. Hiiemae, PhD, BDS1 and J.B. Palmer, MD21Department of Bioengineering and Neuroscience, and Institute for Sensory Research, Syracuse University, Syracuse, New York; and2Departments of Physical Medicine and Rehabilitation and of Otolaryngology-Head and Neck Surgery, The Johns Hopkins University MedicalSchool and Good Samaritan Hospital, Baltimore, Maryland, USA
Abstract. Food movements during complete feeding se-quences on soft and hard foods (8 g of chicken spread,banana, and hard cookie) were investigated in 10 normalsubjects; 6 of these subjects also ate 8 g peanuts. Foodswere coated with barium sulfate. Lateral projection vid-eofluorographic tapes were analyzed, and jaw and hyoidmovements were established after digitization of recordsfor 6 subjects. Sequences were divided into phases, eachinvolving different food management behaviors. Afteringestion, the bite was moved to the postcanines by apull-back tongue movement (Stage I transport) and pro-cessed for different times depending on initial consis-tency. Stage II transport of chewed food through thefauces to the oropharyngeal surface of the tongue oc-curred intermittently during jaw motion cycles. Thismovement, squeeze-back, depended on tongue–palatecontact. The bolus accumulated on the oropharyngealsurface of the tongue distal to the fauces, below the softpalate, but was cycled upward and forward on the tonguesurface, returning through the fauces into the oral cavity.The accumulating bolus spread into the valleculae. Thetotal oropharyngeal accumulation time differed with ini-tial food consistency but could be as long as 8–10 sec forthe hard foods. There was no predictable tongue–palatecontact at any time in the sequence. A new model forbolus formation and deglutition is proposed.Key words: Bolus formation — Food transport —VFG — Oropharynx — Valleculae — Soft palate —Tongue — Normal feeding — Deglutition — Deglutitiondisorders.
Although the pioneering observations of Ardran andKemp [1] laid the foundation for our understanding ofthe mechanism of swallowing, relatively little attentionhas since been given to the mechanisms involved in theintraoral management of ingested solid foods leading tobolus formation and deglutition in normal human sub-jects. The goal of this study was to describe the move-ments of food and those of the tongue and hyoid duringcomplete feeding sequences during normal feeding onsoft and hard conventional foods.
The mechanics of chewing have been extensivelystudied, with a primary focus on the interaction betweenthe upper and lower occlusal surfaces of the teeth, themovements of the jaws, the cyclical nature of the chew-ing cycle, and the ‘‘efficiency’’ with which solid food isactually reduced to a ‘‘swallowable’’ condition [2]. In-gestion and deglutition are not addressed in these studies.When jaw movement patterns in complete feeding se-quences (ingestion to terminal swallow for a single natu-ral bite) of normal foods of different consistencies wereexamined by using an electromagnetic jaw movementtransducer [2], it was found that (a) sequences beganwith a short period of irregular jaw movement in whichthe bite was transferred to the postcanine area; (b) therewere a variable number of ‘‘chewing’’ cycles, dependingon initial food consistency, before the first swallow; (c)sequences normally involved multiple swallows; and (d)a period of irregular jaw movement, termed ‘‘clear-ance,’’ usually preceded the terminal swallow and, insome cases, earlier swallows. Although the feeding pro-cess in man is a behavioral continuum, the actual move-ments of the tongue responsible for food managementwithin the oral cavity could not be described given thetechnique used [3]. Nevertheless, because multiple swal-lows occur within sequences, it was inferred that twoprocesses must be occurring concurrently during the
Correspondence to:Karen M. Hiiemae, Ph.D., B.D.S., Institute forSensory Research, Syracuse University, Merrill Lane, Syracuse, NY13244-5290, USA
chewing cycles preceding a swallow: first, inadequatelytriturated food continues to be processed; second, ad-equately triturated food is collected for bolus formation.The tongue has long been assumed to be largely respon-sible for these segregation and aggregation activities(e.g. [4]). However, the mechanisms of tongue–jaw–hyoid movements involved have not been described.
Studies on the role of the tongue in feeding onsolid foods have focused on tongue behavior in chewingcycles. El Malik [5], Tomura et al. [6], and Imai et al. [7],by using direct observation with cinephotography, cine-fluorography and ultrasonography, respectively, de-scribed tongue movement during this stage in the feedingsequence. Unfortunately, their results are not readily in-terpretable in the context of either the jaw movementcycle or the state and position of the food. Videofluo-rography (VFG) is the only available technique in whichthe movements of tongue, jaw, hyoid, and the food (ifradiopaque) can be recorded. Palmer et al. [8] examinedfeeding sequences in normal human subjects by focusingon the pattern of jaw and hyoid movement in the cyclespreceding the swallow. The associated levels of electro-myographic (EMG) activity in the masseter, digastric,geniohyoid, and sternohyoid muscles were recorded withfine-wire electrodes. A second study with radiopaquetongue markers followed [9] and demonstrated that thechanging cyclical patterns of tongue, jaw, and hyoidmovement from ingestion to terminal swallow could berecorded and analyzed within the constraints for radia-tion exposure dictated by the applicable institutional re-view board (5 min lifetime exposure per subject). Theresults confirm that (a) the hyoid is in continuous motionthroughout the feeding sequence, and (b) tongue markers‘‘cycle’’ in a predictable pattern relative to jaw move-ment.
In man, the processes involved in complete feed-ing sequences with solid foods have the following com-ponents:
Stage I transport: transport of ingested material from theincisal area to the molar region of the oral cavity;Processing: the reduction of food within the oral cavityto a ‘‘swallowable’’ condition;Stage II transport: movement of triturated food from theoral cavity through the pillars of the fauces to the oro-pharyngeal surface of the tongue;Pharyngeal swallow [2,8].
It must be emphasized that Stage II transport can occur atthe same time as processing. The sequence of processing,processing with concurrent Stage II transport, and thenpharyngeal swallow can be repeated several times in asingle complete sequence. Each repeat is termed a ‘‘sub-sequence.’’ The number of subsequences is correlated
with the initial consistency of the food [2]. Regular cy-clical jaw and tongue movements may be replaced byirregular jaw movements (clearance) for short periodsbefore a swallow [2,9].
This study was designed to examine the follow-ing predictions: (a) all foods would be transferred fromthe anterior oral cavity to the molar region by using adistinct pattern of tongue and jaw movements, i.e., StageI transport as described in macaque and other mammals[9–11]; (b) there would be a period in which the food wasreduced to a swallowable consistency and mixed withsaliva (usually called mastication); followed by (c) a pe-riod in which adequately triturated (swallowable) foodwas aggregated on the back of the tongue to form abolus; (d) bolus formation would proceed as further pro-cessing/ chewing of inadequately triturated food contin-ued; (e) the first bolus would be swallowed after trans-port through the fauces; and (f) the sequence b–e wouldbe repeated until all food was cleared from the mouth.These expectations were based on ‘‘process’’ rather thanon temporal considerations given in previous studies[2,9–11].
A protocol in which subjects were their own con-trols was used, so that feeding behavior in sequencecould be recorded for constant weights of foods of dif-ferent consistencies: first, with barium to delineate foodposition and condition; and second, without barium andwith tongue markers to describe tongue surface move-ment more accurately. This approach minimized the con-founding factor of substantial interindividual variation insequence behavior between subjects when feeding on thesame foods [2]. This report describes our findings withtwo soft and two hard foods, all with added barium. Jawand hyoid movements were correlated with food positionand tongue shape. Quantitative analysis of tongue move-ment is technically irreconcilable with documentation offood position and movement (see a–f above) because theaddition of barium to the food obscures the position ofmarkers on the tongue surface. Further reports will quan-titatively describe jaw, tongue, and hyoid movementsand address the biomechanics of the orofacial–pharyngeal complex during the feeding sequence basedon the present data and the analysis of the tongue markerdata.
Methods
Ten healthy adults (5 men, 5 women, aged 20–30 years; 7 Caucasians,1 Latino male, 1 Asian-American male, and 1 African-American male)participated in this study. All gave verbal informed consent beforedental and physical assessments to determine whether any of the ex-clusion criteria applied and, after confirmation of eligibility, writteninformed consent. The protocol was approved by the applicable insti-tutional review boards (Johns Hopkins University and Syracuse Uni-
32 K.M. Hiiemae and J.B. Palmer: Food Transport and Bolus Formation
versity). All subjects were in excellent health, with no history of majormedical problems or dysphagia. Swallowing was judged to be normalbased on slow-motion analysis of standard anteroposterior and lateralprojection VFG (barium/water, 50/50% weight/volume ratio; E-Z-HD,E-Z-EM Inc., Westbury, NY). All had normal occlusions (Class I molarrelationships, with overbite and overjet within normal range), with acomplete postcanine series (premolars, first and second molars) andfew or no dental restorations.
For the purposes of the overall study, after full-mouth exami-nation and acceptance into the protocol, upper and lower impressionswere taken (Hydrosil, Dentsply International Inc, Milford, DE), cast inextrahard dental stone (Whippmix, Louisville, KY), and articulated(Eastern Prosthetic Associates, Catonsville, MD) for future referencegiven the possibility that intraoral behaviors are affected by theanatomy of the oral cavity. Radiopaque markers, 4 mm in diameter cutfrom a 0.5-mm lead sheet (Goodfellow, Malvern, PA) with a standardhole punch, were then glued to the buccal surface of the left upper andlower canines with dental adhesive (Ketac, ESPE-Premier Sales Corp.,Norristown, PA).
The approved protocol provides for a maximum 5 min of VFGexposure for each subject. The time constraints restricted recordings totwo complete records for each test food, one without and one withtongue markers. Most records included at least two subsequences (i.e.,at least two swallows). In a few cases, incomplete records were madeof second subsequences, usually because the subject signaled ‘‘mouthclear’’ before all material had been eliminated from the oral cavity orvalleculae. For the present report, only the data for the first record set(food with barium) have been used.
Data Collection
VFG recordings (30 frames/sec, 60 videofields/sec) were made for eachsubject as follows: (a) drinking 10 cc of barium water mix (normalspontaneous swallow); (b) 8 g of chicken spread (Underwood ChunkyChicken, Pet Co., St. Louis, MO) mixed with a little barium paste(Esophotrast, Rhone-Poulenc, France); (c) 8 g of banana, purchased‘‘just ripe’’ as needed from a local supermarket, with a light superficialcoating of barium paste; and (d) 8 g of hard cookie, i.e., shortbreadfingers (Walker’s Shortbread Ltd., Aberlour-on-Spey, Scotland) with alight coating of barium paste. Six subjects also consumed 8 g of un-salted dried peanuts mixed with barium paste. Foods were not fed inrandom order, given previous experience [9]. Subjects were seatedcomfortably in a chair and were asked to minimize head movement (nocranial constraints were used). They were instructed to take the food,provided on a spoon (chicken spread, peanuts) and, when ready, toingest the material. The pieces of banana and biscuit were taken fromthe spoon and placed into the mouth with fingers. VFG recording beganas the spoon/food approached the lips (operator direct vision). Subjectswere also instructed to use a ‘‘hand-up’’ signal when they had clearedall the ingested food (bite) from the mouth. They were also warned thatthey were to signal if they sensed any discomfort, e.g., likely to cough,so the recording could be aborted. A full description of the VFG re-cording technique is given in Palmer et al. [8] and is expanded inPalmer et al. [9].
Immediately after each recording, the tape was viewed in slowmotion. The recording was deemed acceptable if there was good imagequality and minimal head movement and included the complete se-quence from ingestion to terminal swallow. In some cases, records ofhard foods were accepted if they included all behavior to at least thefirst swallow and for some period thereafter. Where possible, anaborted recording was repeated immediately. Given the concern to limitVFG time to the minimum per trial and the instruction to subjects tosignal immediately when they ‘‘had cleared the mouth’’ of all ingested
food, the few trials where the VFG shut-off inadvertently occurred lateare of some interest (see Discussion).
Data Analysis
Each VFG record was examined field by field (a single experiencedobserver), by using a stop-frame/slow-motion VCR. Detailed noteswere made of jaw position and direction of movement: e.g., start maxi-mum gape, end close, start open, end maximum gape [9], hyoid posi-tion, dimensions of the oropharynx at the level of the epiglottis, softpalate position relative to the tongue, and the position/condition of thefood. Preliminary review of the tapes also showed that the relationshipof the tongue surface changed relative to the palate such that its ante-rior, middle, and posterior oral surfaces changed their spatial relation-ship to the hard palate in a patterned manner. Figure 1 illustrates theworking definitions of anterior, middle, and posterior surfaces of thetongue anterior to the fauces. It must be emphasized that these are‘‘geographical relationships’’: the results show that the tongue surfaceis so dynamic that the conventional descriptions based on postmortemmaterial are of little value. These notes were then transferred to amanually generated (graph paper) time-event plot for all subjects. Theduration of in-sequence events (Stage I transport, processing, Stage IItransport, and bolus formation) were calculated from the behavioralnotes and compared with digitized data. The time during which oro-pharyngeal bolus formation occurred was measured against standardanatomical criteria: i.e., the total time elapsed between the start of StageII transport, i.e., transit through the pillars of the fauces (FP), and themoment at which the bolus began its transit across the hypopharynx.That period can be considered in two parts (Figs. 2, 3): (a) the timebetween the start of Stage II transport and the time at which the leadingedge of the forming bolus reached the lower border of the mandible(FP-LBM) and (b) the time between arrival of the leading edge at thevalleculae and the start of the swallow (LBM-Vall). For the purposes ofthis analysis (see Results), swallows are very narrowly defined as ‘‘hy-popharyngeal transit time’’ (HTT), i.e., the time elapsed from the mo-ment the leading edge of the bolus began to move across the hypo-pharynx until the time the trailing edge entered the esophagus, definedas reaching the level of the vocal cords.
Complete sequences (all foods) for five subjects (3 men, 2women) were digitized (SIP software [12]) after parallax correction[8,9]. Cartesian coordinates for the upper and lower canine markers, theapex of the distal cusp on M2 (upper and lower), and the anterosuperiormargin of the ossified image of the hyoid were obtained. These coor-dinates were then manipulated (Microsoft Excel, Redmond, WA) togive the position of the lower canine marker and the hyoid relative tothe upper occlusal plane (defined as the line between the upper canineand upper molar reference points). Event–time plots were prepared, andthe behavioral data from the manual plots were entered onto the coor-dinate plots to allow analysis of the intraoral events with reference tohyoid and mandibular movements.
The accuracy of the visual analysis was monitored by compar-ing the manual plots of jaw movement events (e.g., maximum gape,and start and finish of particular events, e.g., retraction of the body ofthe mandible or hyoid, with the computer-generated movement dataplots). The two showed very close correspondence (0.03–0.05 sec).
The duration of sequences and of included stages were mea-sured by using the following algorithm:
Total sequence duration: the start of Stage I transport (initial ingestion)to the end of the second swallow (defined as end HTT). A completesequence with only one swallow occurred in only two trials, one withchicken spread, one with banana. Conversely, third and fourth swal-lows were not unusual in sequences with the harder foods. A subse-
K.M. Hiiemae and J.B. Palmer: Food Transport and Bolus Formation 33
quence includes all behavior from ingestion to the end of the firstswallow or completion of one swallow to completion of the next.Stage I transport: the time the food crossed the incisors (start maximumgape) until hard foods (the first tooth–food–tooth contact occurred,determined visually and from rate changes in the jaw movement pro-file) or soft foods (the integrity of the image of the bite of food, theingested material) was disrupted.Processing: from the end of Stage I until the initiation of Stage IItransport. This is the period in the sequence in which food is brokendown by chewing, processed by the tongue acting against the hardpalate, or both.Stage II transport: defined as beginning at the time food was clearlydetected distal to the fauces, i.e., between the soft palate and the pha-ryngeal surface of the tongue. It is important to reemphasize that pro-cessing and Stage II can occur concurrently [8], i.e., food is processedas triturated food accumulates to form a bolus.HTT: the time elapsed from the moment the leading edge of the bolusleaves the valleculae to the time the trailing edge enters the esophagus.
Given the patterns of behavior recorded, it became clear that it wouldbe useful to develop some measures by which the movement of theaccumulating bolus on the pharyngeal surface of the tongue could bedescribed. The oropharyngeal aggregation time (OPAT) is the elapsedtime between the first appearance of food on the pharyngeal surface ofthe tongue and the start of HTT. Within OPAT, there are two distincttime periods: (a) time elapsed from first distinct passage of foodthrough the fauces (FP) onto the pharyngeal surface of the tongue andthe time the leading edge of that bolus reached the lower border of themandible (FP-LBM); (b) time elapsed between the time the leadingedge of the bolus filled the valleculae and the time of onset of HTT(LBM-Vall).
Where indicated in Results, data are presented based on theanalysis of the first subsequence (ingestion to first HTT) given thatbehavior in the second subsequence (S-S2) was so irregular that com-
parisons and analysis beyond simple behavioral observation (chew/clearance) would have had little utility.
Results
Although the purpose of this report is to describe themechanisms by which food is managed within the oralcavity and bolus formation occurs, the characteristics ofthe feeding sequence for the food types used precedes thedescriptions of food transport and bolus formation.
Sequence Behavior
As expected [2], total sequence durations differed withfood type, with peanuts (mean4 22.80 sec, SD4 6.19sec) and cookie (mean4 23.61 sec, SD4 3.75 sec)being significantly longer (p < 0.0001) than banana(mean4 9.74 sec, SD4 2.63 sec) or chicken spread(mean 4 9.56 sec, SD4 2.63 sec). There were nosignificant differences (t test) between either the hardfood or soft food sequence durations. However, the val-ues for the standard deviations reflect interindividual dif-ferences in feeding behavior.
Ingestion and Stage I Transport
Stage I transport occurred in all trials and was of similarduration for all food types (Fig. 2). At the start of StageI, the lips were separated, the jaws were wide apart, andthe tongue was depressed anteriorly but domed posteri-orly. There was a wide gap between the tongue dome andthe soft palate. Hard foods were deposited on the de-pressed anterior tongue. Soft foods were scraped fromthe spoon by upper incisors and collected on the anteriortongue as the spoon was withdrawn. The food was car-ried distally from the canine area to the level of the lastmolars by a rapid retraction of the tongue as the jawswere held at a wide gap. This was associated with a sharpdepression and retraction of the hyoid and marked nar-rowing of the oropharynx at the level of the epiglottis.The entire process took about 280 msec. We are describ-ing this form of food transport as pull-back because nochange in tongue shape occurred; rather, the tongue baselengthened (Fig. 1). Without EMG data from the poste-rior suprahyoid and infrahyoid muscles, it is not possibleto determine how this movement is produced. Suffice tosay that the euclidean distance between the mandibularsymphysis and the hyoid is rapidly increased. As the jawbegan to close and the tongue rise, some small forwarddisplacement of the tongue surface relative to the toothrow positioned the still intact bite in proximity to the firstupper molar. Processing then followed. The same pull-back mechanism was seen intermittently during process-
Fig. 1. Diagrammatic mid-line sagittal section through the oral cavityand oropharynx showing parts of the hard palate and the apposing orrelated tongue surfaces described in the text. The plasticity of tonguesurface shape observed in the recordings precludes anything other thana relational definition of each part. The pharyngeal surface of thetongue is that part of the tongue facing either the soft palate or, belowthe uvula, the oropharynx. The anatomical correlates of the temporalevents in bolus accumulation and deglutition are shown. HTT, hypo-pharyngeal transit time; PFAT, postfaucial aggregation time; VAT,vallecular aggregation time.
34 K.M. Hiiemae and J.B. Palmer: Food Transport and Bolus Formation
ing, returning food to the postcanine teeth after accumu-lation on the anterior hard palate.
Processing
The time spent in processing was greater for the harderfoods (Fig. 2). When processing durations for the hardfoods (peanuts and cookie) were compared, no signifi-cant difference was found. However, when the two softfoods (banana and chicken spread) were compared (ttest), they did differ significantly (p < 0.004). This dif-ference is minor when the durations for processing ofcookie and banana (representing hard and soft foods) arecompared:p < 0.000001. As predicted, there was a dis-tinct consistency effect in the duration of processing de-pendent on initial food consistency. Behavior in the sec-ond subsequence (S-S2, Fig. 2) was highly irregular, in-volving both clearance and chewing movements as wellas Stage II transport.
Food Movement
During early chewing cycles on the harder foods, thechanging profile of the tongue markers indicated rotationof the tongue surface about its anteroposterior long axis[9]. The details of this movement and the tongue shapechanges involved cannot be established from lateral pro-jection VFG. However, the surface of the tongue alsocycles continuously in the sagittal and coronal planesthroughout the sequence [9]. Forward movement of apoint on the anterior surface of the tongue begins late inclosing, continues as the surface continues to rise duringlate close and into the intercuspal phase (IP), then con-tinues its forward, but now downward, movement duringopening. In most cycles, the upward movement of theanterior tongue against the hard palate ‘‘squeezed’’ thefood, pushing it distally toward the molars. However, theforward tongue surface movement associated with IP andearly opening can be very powerful, visibly sweeping thebariumised food anteriorly (Fig. 4). This rhythmic action
Fig. 2. The duration (mean ± SD)for each component of the feedingsequence for all subjects and eachfood. Initial food consistency affectsthe duration of processing (Process),Stage II transport with processing(oropharyngeal aggregation time;OPAT) and the second subsequence(S-S2) but with neither Stage Itransport nor the duration ofhypopharyngeal transit time (HTT).First subsequence:n 4 10 forchicken spread, banana and cookie,6 for peanuts; second subsequence:n 4 9 for chicken spread andcookie, 5 for peanuts. No attemptwas made to distinguish processingor OPAT in S-S2 given thevariability of behavior observed.
Fig. 3. Oropharyngeal aggregation time (OPAT)for all trials and all foods (mean ± SD.)postfaucial aggregation time (PFAT) is the timewithin OPAT in which food accumulates on thepharyngeal surface of the tongue between thefauces and the lower border of the mandible;vallecular aggregation time (VAT) is the time inwhich food moves to and fills the valleculae fromthe lower border of the mandible. VAT ends withthe start of hypopharyngeal transit time (HTT).Considerable variation for the actual time thevalleculae were filled was noted in the records: insome trials, HTT followed immediately onvallecular filling; in others, HTT occurred aftersome elapsed time. The effect of food consistencyis clearly shown for OPAT; no effect is seen forHTT.
K.M. Hiiemae and J.B. Palmer: Food Transport and Bolus Formation 35
gradually resulted in a distinct accumulation of partiallytriturated material on the anterior hard palate. At irregu-lar intervals, this material was collected by the tonguewith an exaggerated upward movement of the tip of thetongue followed by its middle part and then moved dis-tally for recycling (Fig. 5).
Movement of food within the oral cavity duringprocessing might best be described as achieved by a‘‘push-pull’’ system: food is pushed backward by thesqueeze-back mechanism (described below) during IPand is intermittently pulled back by bodily movement ofthe tongue as in Stage I. Such repositioning of the foodmass during processing of hard foods was seen in allrecords.
There were distinct differences between the over-all intraoral management behaviors for the soft and hardfoods. As shown in Figure 2, the time spent in processingthe chicken spread and banana was very short. Absentcomparable VFG records in the anteroposterior projec-tion, it was not possible to be sure of the extent to whichthe chicken spread and banana were chewed, i.e.,squashed between the teeth, as opposed to being com-pressed between the tongue and hard palate with thesame net result. However the soft foods were rapidlyrendered swallowable, thus initiating Stage II transport
(Fig. 6). Movement of the squashed soft foods toward thefauces/soft palate was accomplished by a combination ofthe squeeze-back and pull-back mechanisms. In one sub-ject, both were used continually, although the move-ments associated with the latter were somewhat attenu-ated as compared with Stage I. In other subjects, thepull-back mechanism was used intermittently and mostoften during Stage II (see Stage II Transport and BolusFormation).
Stage II Transport and Bolus Formation
The finding that Stage II transport always began duringprocessing was expected, as was the observation thatprocessing continued as the triturated (swallowable) ma-terial was moved distally toward the fauces. What wasnot expected was the finding that Stage II transport oc-curred over such long periods, i.e., substantial propor-tions of the time spent in the first subsequence producedmean values of 50% for peanuts, 42% for cookie, 41%for banana, and 41% for chicken spread but with wideranges between individuals. However, the total durationof Stage II with processing did differ between hard and
Fig. 4.Early Stage II transport with tongue cycling.A–F Six sequentialvideofields from the record shown in Figure 6 taken at approximately0.1-sec intervals starting at 39:31. The straight arrows indicate thedirection of jaw movement; the curved arrows indicate the direction oftongue surface movement. The movement of the forming bolus towardand away from the oral cavity as the tongue cycles is shown, as is thechanging shape of the soft palate.
Fig. 5. Late Stage II transport and bolus formation. Four synchronousevents are shown:(A) the anterior sweep of the tongue collectingresidual food from the incisal area of the palate as the jaw closes (fullocclusion is not reached; see heavy arrow in Fig. 6);(B) the appearanceof an inclined plane on the tongue, which first appears in videofield B;(C) the presence of aggregated food on the pharyngeal surface of thetongue from the vibrating line (hard palate–soft palate junction) fillingthe valleculae;(D–F) distal movement of food from the anterior oralcavity to the pharyngeal surface of the tongue to form a single bolus.Hypopharyngeal transit time began at 45.06, i.e., 70 ms afterF.
36 K.M. Hiiemae and J.B. Palmer: Food Transport and Bolus Formation
soft foods (p < 0.005). We did not expect to find thataliquots of material were moved intermittently throughthe fauces. Stage II transport occurred as a discrete eventin the IP and early opening phase of masticatory jawmotion cycles. The behavioral analysis showed thatcycles in which substantial aliquots of material movedthrough the fauces were typically associated with achange in jaw-opening profile (Fig. 6). In all but tworecords (both for soft food), bolus formation did not oc-cur anterior to the fauces but rather on the pharyngealsurface of the tongue (Figs. 4, 5). Figure 2 shows that theduration of HTT differed little between the first and sec-ond swallows. Triturated food was therefore accumulat-ing on the pharyngeal surface of the tongue for a con-siderable percentage of the total time between the start ofprocessing and the initiation of the first HTT for a bolus.
Food Movement
To identify the start of Stage II, the observation criterionthat radioopaque material had to be clearly visible be-tween the pharyngeal surface of the tongue and the softpalate was used. This meant that small amounts of ma-
terial could well have passed through the fauces beforethe time at which a distinct mass could be identified.Stage II transport was not a one-time event with a largemass accumulated on the posterior oral surface of thetongue propelled through the fauces and promptly swal-lowed (though this behavior was seen in two soft foodrecords). Rather, it was highly incremental. Aliquots ofmaterial were squeezed through the fauces, with the totalvolume on the pharyngeal surface of the tongue buildinggradually and progressively. However, there was no ab-solute correlation between the visible movement of foodthrough the fauces and an immediate expansion of thebolus shadow as seen in lateral projection VFG.
The squeeze-back mechanism used to move foodthrough the fauces (FP) depended on contact between thetongue surface and the hard palate. The anterior part ofthe tongue first contacted the hard palate, followed by themiddle, and then the posterior surface as the tongue sur-face traveled first forward and then upward as the jawsbegan to separate in early opening. The net effect was tosqueeze the food backward as the area of contact be-tween the tongue and hard palate expanded backward.Triturated food then passed through the fauces onto thepharyngeal surface of the tongue.
Fig. 6. Digitized sequence of one trial with the subject eating chickenspread. The stages in sequence (see Fig. 2) are identified: Stage I beginsat the initial maximum gape and lasts until the fine dotted line. Thereare only two jaw closing movements (1.5 cycles) before the start ofStage II transport (heavy dots). Processing combined with Stage IItransport continues until the attentuated cycle marked by the heavyarrow. The fine arrows mark the rate changes in jaw opening associatedwith significant Stage II transport through the fauces. The heavy arrow
marking the attentuated cycle before the swallow marks the jaw move-ment profile for the tongue behavior shown in Figure 5. After oneregular cycle, jaw and hyoid movements in the second subsequence areirregular (see horizontal dashed line) as is characteristic of clearancebehavior. The x-axis shows time elapsed (1/100th sec). Figures 4 and5 show sections of this record. The time of each videoframe is shownin parentheses in those figures for comparison.
K.M. Hiiemae and J.B. Palmer: Food Transport and Bolus Formation 37
Documentation of the pattern of food movementin Stage II is complicated by the observation that thefauces are not a one-way passage. Figures 4 and 5 illus-trate the problem. The figures were drawn from vid-eofields for the same sequence for a subject eating softfood (Fig. 6) and are typical of the data obtained. Forclarity, the radiographic shadow of the hard palate andthe occlusal surfaces are drawn. Where the tongue sur-face profile or the upper surface of the food profileclosely approximated and paralleled the hard palate,tongue–food–palatal mucosal contact occurred. As thejaw started to open (Fig. 4A), there was a small butdistinct mass of material on the pharyngeal surface of thetongue below the soft palate. There was also a smallmass at the junction of the anterior and middle hardpalate (Fig. 1). As the tongue began to sweep forward inearly opening, these two masses merged and their vol-ume appeared to increase (this may simply be a functionof mediolateral tongue surface profile change as seen inlateral projection). With further jaw opening (Fig. 4C),the tongue surface traveled backward, and the shape ofthe food shadow changed again; but as the movementcontinued, the leading edge of the triturated materialpassed through the fauces onto the pharyngeal surface ofthe tongue. Eighty milliseconds later (Fig. 4E), the lead-ing edge of the forming bolus approached the lower bor-der of the mandible. However, and importantly, as thetongue surface rose in the next closing movement (Fig.4F), the bolus on its pharyngeal surface also rose, so thatits trailing edge (most anterior) border lay below thevibrating line marking the functional hard palate–softpalate junction. The cycling shown in Figure 4 continuedfor the period during which both processing and Stage IItransport occurred in synchrony. However, as the se-quence progressed, the volume of food ‘‘held’’ on thepharyngeal surface of the tongue steadily increased untilthe leading edge of the enlarged bolus had reached andfilled the valleculae (Fig. 5).
Toward the end of the Stage II period, an exag-gerated form of the squeeze-back mechanism often oc-curred, especially for the harder foods (Fig. 5). Residualmaterial had accumulated on the anterior hard palate.From a very depressed position, the tongue tip rose (Fig.5A) and then expanded upward to contact the distal sur-face of the incisors, then the anterior hard palate while‘‘collecting’’ the accumulated material on its surface(Fig. 5C). As the tongue surface continued to rise, itscontact with the hard palate traveled distally (Fig. 5C–E)as the jaws continued to open. The process continuedthrough the first part of the following jaw-closing move-ment (Fig. 5F). Figure 5 shows the changing profile ofthe bolus as the tongue was moving: it was near hori-zontal and highly angulated Figure 5B and nearly verti-cal in Figure 5E.
Figures 4 and 5 exemplify our results. Figure 3(10 subjects) shows the mean durations for the time foodaccumulated on the pharyngeal surface of the tongue.The observation that a bolus of triturated food could‘‘ride’’ on the pharyngeal surface of the tongue for sev-eral seconds before HTT made it very difficult to de-scribe pharyngeal bolus position in relation to traditionalanatomical landmarks such as the lower border of themandible (LBM) or the valleculae. In the earlier part ofthe Stage II transport stage of the sequence, the bolusmay have ridden on the upper part of the pharyngealsurface in relation to the soft palate (Fig. 4), whereas inthe last stages preceding the swallow, its leading edgecould not only reach the valleculae but could also remainfor some time before the onset of HTT. The total time thechewed food–saliva mix was retained on the pharyngealsurface of the tongue is clearly related to initial foodconsistency (Fig. 3).
DeglutitionAlthough not reported in the present study (other than inMethods), VFG records were obtained for all subjectswhile swallowing (no command) 10 cc of 50/50 barium/water mix. Those tapes were carefully reviewed and, in6 cases, digitized. No variation from expected behaviorwas seen: the barium suspension was contained in theoral cavity, propelled through the fauces, and swallowed.
Given these results, HTT can be measured andappropriately used to define the actual swallow, i.e., thepassage of the bolus from the lower oropharynx to theesophagus. Nevertheless, if the traditional description isused (i.e., from bolus accumulation in the oral cavity topassage through the upper esophageal sphincter), consid-erable variation was observed in the records. In the twocases (soft foods only) where Stage II and the swallowwere a seamless event, i.e., a bolus was swallowed di-rectly from the oral cavity, a comparatively small foodshadow (as seen in lateral projection) suddenly expandedas it passed through the fauces and passed directly intoand across the oropharynx. In rare cases, a bolus accu-mulated on the posterior oral surface of the tongue whilematerial was already riding on its pharyngeal surface.Material still in the mouth was then swept through thefauces, collecting the pharyngeal material as it passed,and the totality was swallowed. In the most frequentpattern, the bolus formed on the pharyngeal surface ofthe tongue was the bolus swallowed, although often aug-mented by additional material from a powerful Stage IImovement in the preceding one or two cycles. Immedi-ately before swallowing (HTT), the bolus changed shapeby broadening at its leading edge. It then suddenlystarted to cross the hypopharynx. The leading edge en-tered the esophagus rapidly, the trailing edge some timelater.
38 K.M. Hiiemae and J.B. Palmer: Food Transport and Bolus Formation
The Soft Palate and the Posterior Oral Seal
Figures 4 and 5 show serial tracings from VFG records.The shape of the soft palate as shown is as accurate as themethod allows. The apparent absence of continuous softpalate-tongue contact during processing was so unex-pected [13] that particular care was taken during theslow-motion/stop-frame analysis to document the posi-tion and shape of the soft palate. The behavioral records,in which soft palate–tongue contact was annotated cycleby cycle, show a regular pattern of soft palate elevation.Toward the end of IP, or just after, the soft palate ‘‘lifts’’away from the pharyngeal surface of the tongue regard-less of any preceding tongue–soft palate contact. In longprocessing periods, a rhythmic pattern of tongue–softpalate contact as the teeth approached occlusion, with aloss of contact toward the end of IP, was sometimes seen.In other records, a long series of cycles with no tongue–soft palate contact whatsoever was followed by the oddcycle where contact occurred during IP. There was indi-vidual variation in the shape of the elevated soft palate:in some individuals it became ‘‘triangular’’ (Figs. 4, 5);in others, the palate simply moved away from the tongue,i.e., the space between the pharyngeal surface of thetongue and the soft palate widened.
Although the VFG records showed no consistentpattern of tongue soft–palate contact, contact did occur intwo specific sets of circumstances. First, when the jawwas opened widely and the tongue domed posteriorly,the uvula appeared to be in contact with the posteriorsurface of the tongue, which may be a function of thegeometry of the system, given a relaxed soft palate andvery wide jaw opening. Intermittently, soft palate–food–pharyngeal surface of the tongue contact occurred atearly stages in Stage II transport, sometimes close tomaximum gape. In other sequences, such contact oc-curred just before HTT.
Discussion
We consider the most immediately significant findingsfrom this study to be those related to bolus formation andswallowing when normal human subjects feed on con-ventional semisolid or solid foods. They can be summa-rized as follows.
1. Stage II transport of material through the fauces is,normatively, a prolonged and gradual process. Smallaliquots of swallowable material are passed throughthe fauces by the squeeze-back mechanism originallydescribed by Ardran and Kemp [1].
2. Stage II transport occurs in synchrony with continuedfood processing, i.e., aggregation of swallowable foodoccurs while inadequately chewed food continues tobe processed.
3. The accumulating bolus on the pharyngeal surface ofthe tongue is maintained on that surface without anydiscernible anatomical constraint. There is no controlof the forming bolus by the soft palate. To the best ofour knowledge, this is a new finding.
4. The behavior of the soft palate during complete feed-ing sequences has never been reported. The observa-tion that tongue–soft palate contact is both intermit-tent and irregular raises questions as to its role and tothe concept of the posterior oral seal as defined byDantas et al. [13].
5. The forming bolus rides on the pharyngeal tonguesurface such that its position, relative to standard skel-etal landmarks, such as the LBM, changes cycle tocycle, although there is a gradual overall progressiontoward the valleculae. To the best of our knowledge,this has never been fully described, although it hasbeen reported by Palmer et al. [8]. This finding raisesquestions as to the validity of using skeletal land-marks such as LBM in descriptions of swallowingwhen the bolus is derived from normal-sized bites ofstandard solid foods.
The methods used here were not novel. They were usedby Ardran and Kemp in their pioneering research in the1950s [1]. Since then, given the substantial radiation ex-posure needed for cinefluorography, its use has been re-stricted on ethical grounds to necessary clinical diagnos-tic testing. The development of VFG technology and thedramatic improvements in image intensifier, videocam-era, and recorder performance has made it possible toconduct studies such as this with minimal risk to normalsubjects. Nevertheless, stringent safety controls properlyconstrain experimental design. We cannot replicateArdran and Kemp’s study of 250 subjects, many withrepeated recording sessions, but we have confirmedmany of their findings. Their observations on the mecha-nisms involved in taking food from a spoon describe thepresent results.
During the feeding sequence, the consistency ofingested food is changed by the action of the teeth and/ortongue acting against the hard palate and its mixing withsaliva. Ideally, perhaps, all the test foods used in thisstudy should have been thoroughly mixed with bariumpowder, as was done for a study in rats [14], rather thandusted externally or given a thin coat of Esophotrast. Inpractice, the barium was rapidly distributed throughoutthe developing food mass, rendering it clearly visible. Insome trials, subjects signaled that they had fully com-pleted the feeding sequence and had swallowed all theingested material. In fact, in those cases where recordingwas not promptly stopped, the tapes showed a thin filmof residual material, by inference a mixture of very fineparticles, saliva, and some barium, on the pharyngeal
K.M. Hiiemae and J.B. Palmer: Food Transport and Bolus Formation 39
surface of the tongue and often some in the valleculae.The subjects were unaware of this material.
It has been axiomatic that bolus formation anddeglutition in man differs, even from the higher primates,given the shorter human face and longer vertically ori-ented oropharynx, so as to avoid the risk of aspiration.Although recent work [8,9,12,15–17] has shown thattriturated material enters the pharynx before the boluscrosses the hypopharynx, the results of the present studyconfirm that the concept of a mouth sealed from theoropharynx until a bolus is ready for swallowing cannotbe sustained for solid foods. Palmer et al. [8,9], andDengel et al. [12] reported sequence data in which foodaccumulation distal to the fauces occurs. Dua et al. [16]found material in the oropharynx before the swallow wasinitiated in 60% of liquid and 76% of solid food swal-lows by using videoendoscopic and videofluoroscopictechniques. They reported mean time intervals betweenmaterial in the valleculae and initiation of swallow (de-termined by submental EMG and adduction of the vocalfolds) of 3.2 ± 0.5 (SE) sec for liquids and 2.1 ± 0.3 secfor solids. These values are consistent with the datashown in Figure 3.
It follows that new explanations for the biome-chanics and physics of solid food bolus formation haveto be found. Prinz and Lucas [18] argued that the initia-tion of swallowing is not dependent on separate thresh-olds for particle size and particle lubrication, as proposedby Hutchings and Lillford [19], but rather occurs when abatch of food particles is cohesive enough, despite itsweight, to form a finite bolus. They hypothesized that thecombination of particle size reduction with the additionof saliva as a binding agent, increasing viscous cohesion,creates a bolus that could cross the oropharynx withoutthe risk of small particle aspiration. Their model was,however, developed before the present study and wastested with carrots and brazil nuts. No aspiration of anypart of the boli forming on the pharyngeal surface of thetongue occurred in any of our subjects, which suggeststhat the mechanical properties of the bolus may be piv-otal to oropharyngeal accumulation. If we assume thatthe Prinz and Lucas model is applicable to our findings,it then follows that triturated food has reached somecritical threshold when Stage II transport begins and thatthe physical properties of the initial aliquots of food par-ticles and saliva are maintained for long periods (sec-onds) while additional aliquots are added to the material.
The observations that the soft palate had no pre-dictable relationship to the upper pharyngeal surface ofthe tongue and that no contact whatsoever could be dis-cerned in some sequences were wholly unexpected.Equally unexpected was the finding that, regardless ofwhether previous contact had occurred as the teeth ap-proached occlusion, a distinct upward movement of the
soft palate away from the surface of the tongue as thejaws opened was normative. The palatal shape changeranged from slight to distinct ‘‘hooking.’’ These findingssuggest that, like the tongue and the jaws, the soft palateexhibits some rhythmic behavior.
During Stage II, the bolus on the pharyngeal sur-face of the tongue rides upward and downward (towardthe esophagus) with each cycle. This rhythmic change inbolus position raises questions as to the utility of skeletalreference points such as the LBM or the valleculae asmarkers of bolus position in normal feeding on semisolidor solid foods. Such landmarks clearly serve as measuresof bolus progression in the liquid swallow, with the teethtogether or near occlusion. The apparent (VFG) align-ment of the LBM relative to the deeper tissues (hyoid,larynx, etc.) is a function of the mandibular angle, i.e.,the angle between the posterior border of the ramus andthe lower border of the mandiblesensu strictu.Thisangle is normally 120° ± 5 but can range from just above90° in some individuals to of the order of 135° in others,most of whom could have normal occlusion and a normalfacies [20]. Because the hyoid is in continuous motionduring feeding and little, if any, attention has been paidto the dynamic relationships between the VFG images ofthe hyoid and the LBM, we suggest that the use of theLBM as a guide to bolus position be restricted to that partof the masticatory cycle when the teeth are in or veryclose to occlusion. In the case of the pharyngeal boliformed during feeding on solid foods, the cycling of thebolus on the tongue surface makes this criterion of bolusprogression even less useful. The situation concerningthe floor of the valleculae is somewhat different. Oncethe leading edge of the bolus has reached the valleculae,it remains in that position until HTT, although the trail-ing (mostly oral) edge may continue to cycle.
Classically, the oral cavity and the oropharynxhave been considered to be separate physiological spacesseparated by a gate, i.e., the posterior oral seal. In com-mand swallows, liquid boli are formed in the oral cavityand are very rapidly propelled through the fauces, acrossthe oropharynx, and into the esophagus. This appears notnecessarily to be the case for spontaneous liquid swal-lows [16] or normally ingested and processed solid andsemisolid foods. Instead, the most posterior oral cavityand the oropharynx (oropharyngeal surface of thetongue) form a single functional space into which tritu-rated food is passed by the squeeze-back mechanism butwithin which the forming bolus is passively carried onthe cycling tongue surface.
We have presented data that argue for a revisionof the current paradigm for the human swallow as occursin normal feeding. Figure 7 compares the current swal-lowing model with our proposed process model. Theprocess model was first developed from experimental
40 K.M. Hiiemae and J.B. Palmer: Food Transport and Bolus Formation
data on nonhuman mammals [11,21] and reflects a physi-ological, not a clinical or diagnostic, approach to feedingmechanisms. Importantly, our data and those presentedby Linden et al. [17] and Dua et al. [16] show that themechanisms of bolus formation and deglutition in manmore closely resemble those of other mammals than hasbeen recognized. These resemblances have been fullydiscussed in a major review [22]. We argue that thetraditional model, largely based on studies using the liq-uid command swallow, masks the variations in the tem-poral sequence of events and the effects of food consis-tency on the site of bolus formation.
The oral propulsive phase is not the start of aseamless sequence of coordinated behaviors moving abolus directly from the oral cavity to the esophagus. In-stead, aliquots of triturated material are propelledthrough the fauces onto the pharyngeal surface of thetongue. We consider Stage II transport (movement ofmaterial through the fauces) to be a more accurate de-scriptor of the process because it describes the movementof material from the oral cavity to the oropharynx withno connotation as to the timing of subsequent events.After Stage II transport, the liquid bolus travels directlyacross the oropharynx and hypopharynx and into theesophagus. Boli derived from semisolid or solid foodsaccumulate on the oropharyngeal surface of the tongue,whereas Stage II transport of material from the oral cav-ity continues for periods that may be as long as 10+ sec.At some pivotal time, the trigger for which is not yetknown [22], the oropharyngeal bolus is propelled acrossthe hypopharynx through the opened upper esophageal
sphincter into the esophagus. We argue that the tradi-tional pharyngeal phase of swallowing has to be equatedwith HTT, which is the same for all boli (Fig. 3).
We recognize that the present results and conclu-sions not only argue for a revision of the classic swal-lowing paradigm but also question the current models forthe neuronal control of swallowing. In advancing theprocess model for the description of events in completefeeding sequences, we are suggesting that the commandswallow, however valuable for diagnostic purposes, doesnot adequately represent typical behavior in normal sub-jects.
To summarize, bolus formation and deglutition ofliquids involves a rapid sequence of events: a volume offluid is propelled from the oral cavity, crosses the fauces,passes down and across the pharyngeal surface of thetongue, and enters the hypopharynx and then the esopha-gus. In contrast, triturated food is accumulated on thepharyngeal surface of the tongue after passage throughthe fauces, with a variable period in which additionalaliquots of food are moved distally. After a variable pe-riod of elapsed time, the pharyngeal bolus is swallowed.A revision of the current paradigm/descriptive terminol-ogy for the normal human swallow is proposed.
Acknowledgments.We are grateful to Xuezhuen Wu for superb tech-nical assistance. Drs. Gail Dengel and JoAnne Robbins kindly providedthe SIP software. We wish to acknowledge the contributions made byour subjects in the furtherance of this study. This work was supportedby an award No. RO1-DC 02123 from the National Institute on Deaf-ness and other Communication Disorders, National Institutes of Health.
Fig. 7. The process and swallowing modelsfor solids and liquids. Although thedifferences in sequence duration areindicated by the different lengths of the bars,it is important to note that the actual timespent in processing and Stage II transport ishighly variable depending on initial foodconsistency (Fig. 2). Bolus accumulation inthe oropharynx occurs throughout the periodlabeled Stage II transport; this process doesnot involve propulsion as traditionallydescribed in the classic swallowing model.Oral propulsion is associated with theinitiation of hypopharyngeal transit time.
K.M. Hiiemae and J.B. Palmer: Food Transport and Bolus Formation 41
References
1. Ardran GM, Kemp FH: A radiographic study of movements ofthe tongue in swallowing.Dent Pract 5:252–261, 1955
2. Hiiemae KM, Heath MR, Heath G, Kazazoglu E, Murray J,Sapper D, Hamblett K: Natural bites, food consistency and feed-ing behavior in man.Arch Oral Biol 41:175–189, 1996
3. Kazazoglu E, Heath MR, Ferman AM, Davis GR: Recordingmandibular movement: technical and clinical limitations of the‘Sirognathograph’.J Orofacial Pain 8:165–177, 1994
4. Lucas PW, Ow RKK, Chew CL, Keng SB: Relationship be-tween jaw movement and food breakdown in human mastica-tion. J Dent Res 65:400–404, 1986
5. El Malik A: The part played by the tongue in mastication anddeglutition.J Anat 89:250–255, 1955
6. Tomura Y, Ide Y, Kamijo Y: Studies on the morphologicalchanges of tongue movement during mastication by X-ray TVcinematography. In: Kawamura Y, Dubner R (eds.):Oral-FacialSensory and Motor Functions.Tokyo: Quintessence, 1981, pp45–52
7. Imai A, Tanaka M, Tatsuta M, Kawazoe T: Ultrasonographicimages of tongue movement during mastication.J Osaka DentUniv 29:61–69, 1995
9. Palmer JB, Hiiemae KM, Liu J: Tongue-jaw linkages in humanfeeding: a preliminary videofluorographic study.Arch Oral Biol42:429–441, 1997
10. Hiiemae KM, Hayenga S, Reese A: Patterns of jaw and tonguemovement in the macaque: a cinefluorographic study.Arch OralBiol 40:229–246, 1995
11. Hiiemae KM, Crompton AW: Mastication, food transport andswallowing. In: Hildebrand M, Bramble DM, Liem KF, Wake
DB (eds.): Functional Vertebrate Morphology.Cambridge:Belknap Press of Harvard University Press, 1985, pp 262–290
12. Dengel G, Robbins J, Rosenbeck JC: Image processing in swal-lowing and speech research.Dysphagia 6:30–39, 1991
13. Dantas RO, Dodds WJ, Massey BT, Shaker R, Cook IJ: Mano-metric characteristics of the glossopalatal sphincter.Dig Dis Sci35:161–166, 1990
14. Hiiemae KM: Masticatory function in mammals.J Dent Res46:883–893, 1967
15. Palmer JB: Bolus aggregation in the oropharynx does not de-pend on gravity.Arch Phys Med Rehabil 79:691–696, 1998
16. Dua KS, Ren J, Bardarn E, Xie P, Shaker R: Coordination ofdeglutitive glottal function and pharyngeal bolus transit duringnormal eating.Gastroenterology 112:73–83, 1997
17. Linden P, Tippett MA, Johnston J, Siebens A, French J: Bolusposition at swallow onset in normal adults: preliminary obser-vations.Dysphagia 4:146–150, 1989
18. Prinz JF, Lucas PW: Mastication and swallowing: an optimisa-tion model.Proc R Soc Lond B 264:1715–1721, 1997
19. Hutchings JB, Lillford PJ: The perception of food texture—thephilosophy of the breakdown path.J Texture Stud 19:103–115,1988
20. Broadbent BH: A new X-ray technique and its application toorthodontic practice.Angle Orthodont 1:45–66, 1931
21. Hiiemae KM: Process and mechanism: mechanoreceptors in themouth as the primary modulators of rhythmic behavior in feed-ing? In: Verrillo R (ed.):Sensory Research: Multimodal Per-spectives.Mahwah: Lawrence Erlbaum & Associates, 1993, pp263–283
22. Thexton AJ, Crompton AW: The control of swallowing. In:Linden RA (ed.):The Scientific Basis of Eating.Basel: Karger,1998, pp 168–222
42 K.M. Hiiemae and J.B. Palmer: Food Transport and Bolus Formation
