determinants of rhythm and rate in suckling

THE JOURNAL OF EXPERIMENTAL ZOOLOGY 278:1–8 (1997)

© 1997 WILEY-LISS, INC.

JEZ 765

Determinants of Rhythm and Rate in SucklingR.Z. GERMAN,1* A.W. CROMPTON,2 D.W. HERTWECK,1 ANDA.J. THEXTON2

1Department of Biological Sciences, University of Cincinnati, Cincinnati,Ohio 45221

2Museum of Comparative Zoology, Harvard University, Cambridge,Massachusetts

3Division of Physiology, United Medical and Dental Schools (St. Thomas’Campus), London SE1 7EH, United Kingdom

ABSTRACT Suckling was studied in infant miniature pigs to determine (a) the necessary stimu-lus for eliciting rhythmic behavior and (b) whether the rhythm of the feeding movements could beentrained with a rhythmic pulsed delivery of milk. The animals fed on an automated milk deliv-ery system, which supplied pulses of milk either at fixed, predetermined rates or on demand. Therhythm of the suckling response was quantified from the teat pressure changes produced by theanimal, which were highly correlated with jaw movement. Suckling frequency was measured asthe dominant frequency in the teat pressure wave, determined by fast Fourier transform. Wheneach animal was allowed to determine its own rate of milk delivery, the preferred frequency ofsuckling was approximately 3.8 Hz. When animals attempted to suckle on the teat but milk wasnot delivered, suckling was erratic and arrhythmic. The first aliquot of milk delivered to the ani-mal elicited rhythmic suckling at approximately 4.6 Hz, which was maintained when milk wasdelivered at a range of fixed rates (0.2–0.56 Hz) an order of magnitude below the preferred suck-ling frequency. When milk was delivered at a fixed rate (2.0–5.6 Hz) close to the animals’ pre-ferred rhythm, suckling proceeded at a lower frequency (3.9 Hz) than when the milk was deliveredat the much lower rate. However, variation in the delivery rate (2.0–5.6 Hz) did not cause a sig-nificant difference in the suckling frequency. These findings provided evidence against entrain-ment. The higher suckling frequency elicited by the slower delivery rate was suggestive of a negativefeedback loop; in the infant/sow relationship, such a mechanism could favor a particular volumedelivery per unit time. J. Exp. Zool. 278:1–8, 1997. © 1997 Wiley-Liss, Inc.

Feeding in both adults and infants is a rhyth-mic activity (Dellow and Lund, ’71; Lambert etal., ’86; Goldberg and Chandler, ’90). Peripheralsensory input may also modify the nature of on-going rhythmic behavior in adult feeding (Lundet al., ’81; Thexton and Crompton, ’89; Thextonand McGarrick, ’94). Less, however, is knownabout what is necessary in the infant to elicitrhythmic suckling.

Because suckling can be elicited by naturalstimulation of the mouth and lips at a time whenthe descending pathways from the cerebral hemi-spheres are unmyelinated or incompletely myeli-nated and, in the human, even when the cerebralhemispheres are absent (Marshall-Hall, 1833;Gamper, ’26), peripheral sensory input is likelyto be of major importance in activating those brainstem mechanisms that generate the rhythmic ac-tivity of suckling. It is, however, not knownwhether the intrinsic mechanisms of the infantdetermine the oral rhythm or whether it is the

periodic delivery of milk that determines or influ-ences the rhythm.

This study attempted to define, in the infantpig, firstly, the peripheral stimulus necessary toinduce suckling and, secondly, whether the suck-ling rhythm could be changed by altering, over awide range, the rate at which pulses of milk weredelivered to the oral cavity.

MATERIALS AND METHODSThree infant pigs were obtained from Charles

River Laboratories (NH); two were 2 days post-natal and one was 7 days postnatal. The animalswere not weaned and were easily trained to sucklefrom a soft veterinary nipple (NASCO, FortAtkinson, WI) delivering a commercial infant pigformula (Soweena/Litterlife; Merrick Foods). Data

*Correspondence to: R.Z. German, Department of Biological Sci-ences, ML-006, University of Cincinnati, Cincinnati, OH 45221-0006.

Received 15 March 1996; Revision accepted 3 December 1996

2 R.Z. GERMAN ET AL.

were collected on suckling over a period of 2 weeksin each case and included at least 14 feeding se-quences per individual. A sequence is defined asa recording of one animal’s feeding time from itsinitial approach to the nipple through to satia-tion. Each sequence lasted between 30 and 120sec, i.e., each sequence contained between about100 and 500 suckling cycles.

Infant pigs were fed in a Plexiglas box with thenipple at one end. Unrestrained animals were al-lowed to enter the box and feed freely from thenipple. Milk was delivered with an automatedpulsed delivery system, designed and constructedby one of the authors.

The automated pulsed delivery system was asystem of pumps controlled by hybrid analogue/digital circuitry (custom designed at Guy’s Hospi-tal, London, UK). The system was used to delivera predetermined volume of milk at either a pre-set fixed rate or on demand. In the latter case,every time the pressure level in the teat reacheda preset (+ve or –ve) level, a single delivery ofmilk was triggered (Fig. 1). The automated deliv-ery system employed four diaphragm pumps (LuftInc., MA), operating in parallel to deliver milk.They functioned serially in time in a preset orderto allow adequate time for each pump to fill whenthe pumping system delivered at more than 4 Hz.

The controller triggered the four pumps in a re-peating sequence. The control system consistedof three units, each serving one of three differentfunctions: master period generator and counter,teat pressure threshold detector, and a logic unitto determine which function had priority. The mas-ter period generator generated TTL pulses every10 msec. When a preset number was detected by

the counter, an output pulse was generated. Un-der routine conditions, this pulse passed throughthe logic unit to a sequencing unit that guidedeach pulse to the next pump in a preset order.The pulse also reset the master period generatorto zero to repeat the process. Input to the pres-sure threshold detector unit was a signal repre-senting pressure in the teat, generated by theinfant pig. An output pulse was generated if theincoming signal passed a preset (positive or nega-tive) pressure level. If the logic unit was appro-priately set, a sequence of pressure changes inthe teat generated a pumping rhythm that over-rode any rhythm generated by the period genera-tor and counters.

The four pumps delivered milk into a commonmanifold, which was connected to the teat withsemirigid tubing. A strain gauge was placed onan intervening short section of soft, extensible tub-ing so that a measure of distension (reflecting in-ternal pressure) could be recorded. The straingauge signal permitted the output of the fourpumps to be equalized accurately and acted as asignal marker for the delivery of fluid through theteat. The standard veterinary teat was modifiedby the addition of a central piece of soft tubing(sealed to the inside of the teat), which connectedto the semirigid tubing and carried milk to theteat orifice. The remainder of the teat was in con-tinuity with a side arm but was otherwise com-pletely sealed. This enabled the pressure in thenipple to be recorded by a transducer housed in atube connected to the side arm but at a safe dis-tance (10 cm) from the overly enthusiastic feed-ing activities of the animals. The transducer wasa Millar (Houston, TX) microtip pressure (SPR-249) used with a TCB 500 control box. This ar-rangement had the slight disadvantage that it wasdifficult to use with a negative pressure thresh-old to trigger milk delivery because the teat ori-fice into the mouth was part of the pumped systemand was therefore sealed from the rest of the teatcontaining the pressure transducer. There wasconsequently no direct connection between themouth and the pressure transducer, and the onlyway that a negative pressure could be transmit-ted to the inside of the teat was by suctionexpansion of the entire teat. However, the alter-nating compressive forces of the jaw and tonguecould easily compress the sealed teat to generate apositive pressure to trigger milk delivery (Fig. 1).

In its preset delivery mode, the controller oper-ated by triggering the four pumps in a regularrepeating sequence. In its demand mode of op-

Fig. 1. Schematic diagram of the automated food deliv-ery system. Hybrid analogue/digital circuitry controls fourpumps which deliver a preset aliquot of milk through thenipple.

DETERMINANTS OF RHYTHM AND RATE IN SUCKLING 3

eration, the input to the controller was a signalrepresenting pressure in the teat, generated bythe pig. A single output pulse to one of the pumpswas generated each time the incoming teat pres-sure signal passed a preset level. The delay be-tween pressure level detection and delivery of milkwas 80 msec. Four replicates of feeding in demandmode were made for each individual, each lastingat least 30 sec.

In experiments carried out in the preset deliv-ery mode, a standard three-stage pattern of milkdelivery was used. In stage one, the animal ap-proached the nipple and attempted to feed, butfor the first 2–5 sec, milk was deliberately notdelivered. In stage two, milk was delivered at alow rate, termed “long interval delivery” (between0.20 and 0.56 Hz, i.e., 1.8–5 sec between deliv-eries). This rate was roughly one-tenth of thenormal feeding frequency of the animals when ac-tivating the milk delivery system in its demandmode. Finally, in stage three, milk was deliveredat a rate closer to that of normal feeding, termed“short interval delivery” (2–5.6 Hz, i.e., 180–500msec between deliveries). This rate was ten timesthe long interval rate. Feeding continued at thisdelivery rate until the animal was satiated. Fivelong interval delivery rates were used (0.20, 0.25,0.30, 0.40, and 0.56 Hz), and five short intervaldelivery rates were used (2.0, 2.5, 3.0, 4.0, and5.6 Hz). Three replicates or trials were carriedout for the 0.25/2.5 Hz and for the 0.4/4.0 Hz de-livery rates (Table 1). At the beginning of eachfeeding session, the volume of milk per pump ac-tuation was calculated by running the systemwithout a pig present and collecting the milk de-livered from the teat in a graduated cylinder; thepumps were then adjusted to deliver 0.47 ml. Thethree-stage pattern of milk delivery can be seenin Figure 2.

Filming and recordingThe movements of suckling were recorded by

cineradiography at 100 frames per second using

a Siemens Tridoros 150G3 cineradiographic ap-paratus. The animals stood still for most of thetime while suckling, and the oral movements wererecorded cineradiographically in the lateral planeduring these times. When filming at 100 framesper second (fps), maximum gape occurred in asingle, easily identified frame of film so that acount of the number of frames from one instanceof maximum gape to the next indicated cyclelength. The output of the teat pressure transducerwas recorded together with the pressure signal ofthe pulsed milk delivery and a frame-by-framesynchronization signal from the camera on a Bell& Howell (Pasadena, CA) FM tape recorder. Thetape recordings were then digitized at 250 Hz(more than an order of magnitude greater thanthe base frequency of the animal’s behavior) andanalyzed with the aid of the LabView SoftwareProgram (National Instruments, Baltimore, MD).

Measurement of responseThe fluctuation of pressure within the teat,

generated by the suckling animal, was used asa measure of the rhythmic response. For somesequences, both jaw movement (determined ra-diographically) and teat pressure data were col-lected as measures of the rhythmic response.The experimental design included analysis of tworegimes: demand feeding, where the animal con-trolled the rate of milk delivery, and presetdelivery, where the rate of milk delivery was ex-perimentally set. Set rate delivery was dividedinto three stages of no milk delivery, long inter-val delivery, and short interval delivery.

Fourier analysis was used to dissect the teatpressure waveforms into their component frequen-cies in order to indicate the relative significance,or power, of those components of the total wave-form that were of biological interest, i.e., 0–10 Hz.Because the signal was not stationary, windowsof either 250 (1 sec duration, used for quantita-tive analysis) or 1,000 time points (4 sec dura-tion, used for graphical display) were subject to

TABLE 1. Repeats of delivery rates

Preset long interval Preset short interval Replicates per animal Total replicatesdelivery (Hz) delivery (Hz) (at each rate) (at all rates for all animals)

5.6 0.56 3 274.0 0.40 1 93.0 0.30 1 92.5 0.25 3 272.0 0.20 1 9

All three animals were subject to milk delivery at all of the above rates (Hz).


Fourier analysis. We used the SERIES module ofSYSTAT (Wilkenson, ’91) to extract the dominantfrequencies in each window of data. The windowwas moved in 1 sec increments, for quantitativeanalysis (half-second increments for graphical dis-play), through the entire time series, calculatinga new measurement of suckling frequency for eachtime segment of the record. The individual val-ues for each time window were used graphically,as plots of frequency (vertical axis) vs. time in se-quence (horizontal axis), to show the extent of anychange with time. The power, or the relative im-portance or contribution, of a particular frequencywas represented by the relative darkness plottedfor that frequency at that point in time. Thus, adark line at the 4 Hz level for the first 5 sec oftime indicated that, for that time, the dominantfrequency in nipple pressure generated by the ani-mal was 4 Hz (Fig. 3). The dominant frequencyfrom Fourier analyses of 1 sec windows was themeasure used in the subsequent statistical analy-ses for each time segment analyzed. The finer timeanalysis (1 sec windows) was used for quantita-tive analyses as it would be more accurate. This,however, did not produce an easily interpreted fig-ure; therefore, the 4 sec windows were used forgraphical display.

The frequency estimate for each stage was av-eraged over the successive time windows for eachfeeding sequence for each animal (Kirby, ’93). Dif-ferences in frequency of suckling between trials,between different delivery rates, between individu-als, and between feeding regimes were tested us-ing a three-factor multiway analysis of variance(ANOVA; MGLH module of SYSTAT; Wilkenson,

Fig. 2. Effects of varying milk delivery on sucklingrhythm. Upper trace is intraoral pressure generated by theanimal, measured by the pressure transducer in the teat, andthis reflects jaw movement. Lower trace is the pressure gen-erated by the pumps and represents the time course of milkdelivery. A: When no milk is delivered, there is neither rhyth-

mic response in jaw movement nor rhythmic change in nipplepressure. B: The first aliquot of milk elicits rhythmic move-ment, which continues despite the long interval delivery (lowrate delivery). C and D: Jaw movement slows down whenmilk is delivered at a higher interval (higher rate) delivery.

’91). This method permitted us to determinewhether our factors of interest, i.e., differencesbetween rates and regimes, were significant, giventhe confounding variation that exists among in-dividuals (Shaffer and Laude, ’85). Because we col-lected usable data from animals at most twice aday, it was not possible to test for within-dayvariation within individuals. Specific significantdifferences among factor levels were determinedusing Tukey HSD post-hoc tests with a Bonferronicorrection (Neter et al., ’90).

Some analyses were carried out on subsectionsof the data collected during the short interval milkdeliveries. Because suckling rate tended to changenonlinearly during some of these suckling se-quences, analysis of covariance (ANCOVA) withtime-in-sequence as a covariate was inappropri-ate. Consequently, some analyses were carried outon small sections of record where the rhythmicresponse (teat pressure) had a constant periodic-ity when measured over that period of time, i.e.,when the waveform components were statisticallystationary. We refer to these portions of an entiresequence as subsequences.

RESULTSGeneral behavior during various

feeding regimesSeveral aspects of behavior were consistent for

all experimental trials, including demand feeding,set rate long interval delivery, and set rate shortinterval delivery. When placed in the feeding box,animals freely approached the nipple and took itinto the mouth. A sequence of suckling cycles usu-


ally lasted 30–120 sec, providing adequate milkwas delivered. After this, animals appeared sati-ated and usually appeared to play with the nipple,although little feeding occurred. There was noobvious difference in gross behavior among thedifferent feeding situations. During a suckling se-

quence, animals occasionally broke off suckling for200–800 msec, without relinquishing hold of thenipple. Such interruptions defined subsections,which varied irregularly from 2 to 23 sec dura-tion, within the overall suckling sequence. Swal-lows, defined by the presence of milk in thepiriform recesses on cineradiographs, did not oc-cur during these breaks. In general, swallows oc-curred every two to three cycles. Cycles containingswallows were not statistically different in lengthfrom cycles without swallows, as determined fromframe counts in cineradiographic films.

Visual inspection of the cineradiographic recordsindicated that a dorsal/ventral movement of thejaw and tongue was used for suckling, similar tothat previously described (German et al., ’92). Itwas also evident from the cineradiographs thatjaw closure was associated with compression ofthe teat by the tongue, and this, in turn, was as-sociated with an increase in teat pressure. Whenthe successive times at which jaw gape reachedits minimum during suckling were compared withthe times at which the simultaneously recordedteat pressure reached a maximum, there was astrong correlation (R2 > 0.95). The pressure withinthe teat was consequently used as a measure ofrhythmic oral movement generated by the animal.

Suckling frequency during demand feedingAt the beginning of each session of demand feed-

ing, the delivery threshold (the level to which thenipple pressure had to rise or fall before trigger-ing the delivery of a pulse of milk) was set so asto ensure that each animal could obtain milk regu-larly. If the threshold was set for a negative pres-sure, the animal was rarely able to develop asufficiently negative pressure in the mouth forthat pressure to be transmitted to the transducersealed in the teat and so be successful in obtain-ing milk. However, if the threshold was set at asuitable positive pressure level, the alternatingcompressive forces on the teat, generated by tongueand jaw elevation, were able to compress it so thata positive pressure was readily generated inside thesealed teat and all animals could reach thresholdand feed to satiation. Suckling frequencies were inthe range of 3.5–4.4 Hz (Table 2), corresponding tocycle lengths in the range 227–286 msec. Each ani-mal had a preferred frequency which was signifi-cantly different from the other animals (P < 0.001).

Response to changing stimuliAll animals had a consistent response to the pre-

set, three-stage milk delivery design of no milk,

Fig. 3. Dominant frequency in teat pressure records, basedon Fourier analysis plotted as a function of time. Each plotshows the animal response to three delivery regimes: no milk,long interval delivery, short interval delivery. Each plot is adifferent animal. The power of each possible frequency (2–10Hz) is represented as values on a gray scale, where very darkpoints represent the dominant frequencies at a given pointin time. Each diagram is a continuous record; each sequencediffered in length. A: Pig 1 produced a weak signal at about3 Hz when no milk was delivered; it produced a strong sig-nal at about 5 Hz during low rate delivery and a strong sig-nal at 4 Hz during high rate delivery. B: Pig 2 produced manystrong frequencies when no milk was delivered, indicating achaotic waveform. This animal formed several subsectionsduring high rate delivery, which started at about 5 Hz andthen dropped to about 4 Hz. C: Pig 3 produced no rhythmicteat pressure signal when milk was not delivered; it gener-ated a high frequency component of the signal during longinterval delivery, but that frequency dropped to 3.5–4 Hz dur-ing short interval delivery.


long interval delivery, short interval delivery(Fig. 2).

No delivery of milkWhen no milk was delivered, there was no overt

rhythmic response from any animal in more than15 trials per animal, where each trial lasted from2 to 30 sec. Furthermore, no consistent level ofpressure was generated (Fig. 2A). Behaviorally,the infants were excited, mouthed the teat vig-orously, and sometimes seemed to attack it.However, on one occasion a single drip of milkinadvertently left on the end of the teat wassufficient to elicit rhythmic activity lasting sev-eral seconds; this sequence was excluded fromanalysis. Rhythmicity was not evident in over 80%of the trials, but in the remainder a small rhyth-mic component of activity was detectable. The fre-quencies of the small rhythmic components were,however, significantly lower than in any otherfeeding stage, and the power of these frequencies(a measure of the significance of that frequencyto the data sequence) was significantly lower thanthe power of the main frequencies in the two sub-sequent stages with milk delivery (P < 0.001). Inthe sequences with some evidence of rhythmicity,the average frequencies of the rhythmic portionsfor the three animals were 2.3 (S.D. 0.6), 3.2 (S.D.1.1), and 2.8 (S.D. 1.1).

Long interval delivery of milkThe delivery of the first aliquot of milk greatly

reduced the animal’s display of excitement. It also

elicited obvious rhythmic jaw movement. Thechange in behavior usually occurred in less than1 sec after the first drop of milk entered the oralcavity, although in five sequences, a few secondspassed before rhythmic behavior began. Once therhythm was initiated, the patterns of jaw move-ment and of teat pressure were regular, eventhough the milk was delivered only every 2–5sec (Fig. 2B). Suckling sequences for all ratesof long interval delivery (0.20, 0.25, 0.30, 0.40,and 0.56 Hz) were characterized by rhythmicjaw movements.

Short interval delivery of milkWhen milk was delivered automatically at a

rate within a few Hz of the animal’s preferred fre-quency (determined by demand feeding on the au-tomated feed delivery), suckling was regular andrhythmic (Fig. 2C,D). Significant rhythmicity wasevident in every sequence.

Characteristic responses to the three deliverystages are graphically summarized in Figure 3A–C, where the darkness of the line indicates thesignificance of that frequency to the pressurewaveform. Although the sequences were of differ-ent duration, there is a consistent pattern acrossall three. In each case, no single clear frequencywas evident during the first period of time, whenno milk was delivered. During long interval de-livery, there was initially a relatively high fre-quency of response, but in two of the threeanimals, this frequency fell over the next few sec-onds. Finally, in the third stage, during short in-

TABLE 2. Suckling frequency for demand and preset rate delivery, standard deviations (S.D.) andnumber of seconds used in calculations

Pig 1 Pig 2 Pig3N N N

Frequency S.D. (sec) Frequency S.D (sec) Frequency S.D. (sec)

Demand feeding 3.5 1.0 103 4.4 1.2 161 3.5 .7 139Preset long intervaldelivery (Hz)

5.6 4.3 .9 24 4.6 .3 16 4.9 .4 154.0 4.6 .4 21 4.8 .3 15 4.4 .9 203.0 3.8 .4 20 4.9 .3 23 5.0 .6 192.5 4.9 .2 33 4.7 .5 50 4.3 .3 462.0 3.5 .4 32 5.0 .4 22 4.6 .5 28

Mean response 4.22 4.80 4.64Preset short intervaldelivery (Hz)

0.56 3.6 .1 27 4.7 .7 24 3.4 .1 70.40 4.0 .3 82 4.2 .5 110 3.5 .5 470.30 3.9 .3 43 4.5 .5 52 3.3 .2 470.25 4.1 .2 91 4.1 .2 173 3.7 .4 1300.20 3.9 .2 52 4.1 .4 42 3.3 .3 35

Mean response 3.90 4.32 3.44


terval delivery, each animal suckled at a slowerfrequency. In one animal (Fig. 3B), the subsec-tions were clearly discernable, each one startingat a high frequency of suckling, and then droppedto a lower frequency.

The results from the full model ANOVA indi-cated that there was a highly significant differ-ence between the second and third stages of milkdelivery. Animal response to long interval deliv-ery was a suckling frequency that was signifi-cantly higher than suckling at short intervaldelivery (P < 0.0001). This effect accounted for40% of the total variation in the data set. In con-trast, the significant difference between individu-als (P = 0.009) accounted for only 19% of the totalvariation. The responses to different delivery ratesnested in delivery stages (i.e., five rates nestedwithin long interval delivery and five rates nestedwithin short interval delivery) were not signifi-cant (P = 0.857), even when differences due tostage and to individual were accounted for. Infantsuckling rate was consequently not significantlydifferent when milk was delivered at rates be-tween 0.2 and 0.56 Hz, i.e., the long interval de-livery stage. Nor was there any difference inindividual response to the changes in rates withinthe short interval delivery stage, i.e., at rates from2.0 to 5.6 Hz.

Changes in suckling rate over timeIn almost all situations (demand feeding, long

interval delivery, and short interval delivery),suckling slowed over the course of the sequenceand within any subsection. This pattern was par-ticularly evident in individuals exhibiting clearsubsections (Fig. 3B). These subsections were aseries of ten to 20 cycles of suckling, followed bya brief absence of activity, lasting less than 1 sec.Usually, the first subsection was significantlyfaster than the following. The first cycle of thesecond subsection was slower than the first cycleof the first subsection but always faster than thelast few cycles of the first subsection.

DISCUSSIONExtensive work documents the significance of

peripheral stimulation in eliciting rhythmic oralbehavior (Bremer, ’23; Lund and Dellow, ’71, ’73;Nakamura et al., ’76; Lund, ’76; Dubner et al.,’78; Thexton, ’76), although such studies were car-ried out on acute experimental animals and maynot always have produced the same behavior thatwould have occurred in the normally feeding ani-mal. Nevertheless, in some cases (Bremer, ’23;

Thexton and McGarrick, ’87) the activities, elic-ited by fluid stimulation of the mouth of decer-ebrate animals, appear to be appropriate to thestimulus; i.e., they resemble licking, lapping, orsuckling. Conversely, mechanical stimuli eliciteddifferent types of rhythmic activity. The evidenceof this paper also indicated that rhythmic suck-ling in the intact infant pig was not elicited sim-ply by mechanical contact with an artificial teat,although that teat had been used for all previousfeeding and the animals were fully habituated toit. The presence of only a trace of milk on the teatwas, however, sufficient to initiate several secondsof rhythmic activity, and the delivery of just 0.46ml of milk only once every 5 sec was an adequatestimulus for fully maintained rhythmic activity.

The greatest difference found in suckling fre-quency was between suckling in response to longinterval delivery and suckling in response to shortinterval delivery. During long interval delivery,where delivery rate was an order of magnitudeless than the preferred frequency (as determinedby demand feeding), suckling response was higherthan the preferred frequency (Table 2; Fig. 3).When the rate of pulsed delivery of milk was closeto the preferred suckling frequency, the actual fre-quency of the elicited suckling dropped signifi-cantly to its preferred level. This could mean thatthe intraoral stimulation produced by the deliv-ery of milk acted to decrease suckling frequency.Similarly, peripheral stimulation has been shownto be capable of lengthening the jaw cycle periodin other experimental situations (Lund et al., ’81).The effect of repeated delivery of milk might alsopartly explain the gradual slowdown in the suck-ling frequency that can occur with time during asuckling sequence (Fig. 3). However, in the lastsituation, the presence of milk is unlikely to bethe sole factor because a gradual slowing of anelicited rhythm can also occur in the acute ex-perimental animal when central pattern genera-tor activity is elicited by electrical stimulation ofdescending pathways (Dellow and Lund, ’71).

The finding of the inverse relationship betweendelivery rate and response suckling frequencymight also be interpreted as indicating that theinfant pig can respond in a negative feedbacksense to natural milk delivery by the sow. Thehypothesis is that the delivery rate of milk belowa certain threshold produces a motor response inthe infant that provides increased mechanicalstimulation to the teat. Such activity would elicita great milk ejection. Conversely, milk deliveryabove a particular level could reduce the oral mo-


tor activity and so reduce the mechanical stimu-lus for milk delivery.

When the animals controlled their own rate ofmilk delivery, each animal had a preferred suck-ling frequency. In this situation, each jaw closingmovement raised teat pressure and, providing thatthe pressure threshold level was reached, milkwas delivered within 80 msec and therefore withinthat jaw cycle. There was consequently a 1:1 re-lationship between milk delivery and jaw move-ment over most of such a feeding period. Thisfurther implies that the milk always entered themouth at the same point in the jaw cycle and thata constant phase relationship existed between jawmovement and milk delivery. In contrast to this,fixed rate delivery was often different from thesuckling frequency, which makes it extremelylikely that, over 100 or so suckling cycles, milkwas delivered at all phases of the jaw cycle. Nev-ertheless, the animals suckled readily and suc-cessfully under these conditions. This suggeststhat the activity of the pattern generator that op-erates in suckling is dependent upon the presenceof milk for its initiation but that the subsequentrate of operation is independent of the stimulusprovided by milk delivery. The infant pig is, how-ever, a precocious animal, and the inability to en-train the rhythm of suckling may have beenrelated to the relative neuronal maturity (Thextonet al., ’95) of that species. Therefore, this may im-ply that the mechanism, described by Kurasawaet al. (’88), for preventing disturbance to an es-tablished rhythm is also functional.

ACKNOWLEDGMENTSThis work was supported by NIH DE9967. We

thank Mr. R. Burton for invaluable help in theconstruction of the pump controller circuits, D.Hertweck for his computational assistance, andC. McCluskey and M. Maunz for their commentson the manuscript.

LITERATURE CITEDBremer, F. (1923) Physiologie nerveuse de la mastication chez

le chat et le lapin. Arch. Int. Physiol., 21:309–352.Dellow, P.G., and J.P. Lund (1971) Evidence for central

timing of rhythmical mastication. J. Physiol. (Lond.),215:1–13.

Dubner, R., B.J. Sessle, and A.T. Storey (1978) The NeuralBasis of Oral and Facial Function. Plenum Press, New York.

Gamper, E. (1926) The Human Mid-Brain (Film). British FilmInstitute, London.

German, R.Z., A.W. Crompton, L.C. Levitch, and A.J. Thexton(1992) The mechanism of suckling in two species of infantmammal: Miniature pigs and long-tailed macaques. J. Exp.Zool., 261:322–330.

Goldberg, L.J., and S.H. Chandler (1990) Central mechanismsof rhythmical trigeminal activity. In: Neurophysiology of theJaws and Teeth. A. Taylor, ed. Macmillan Press, London,pp. 268–289.

Kirby, K.N. (1993) Advanced Data Analysis with SYSTAT. VanNostrand Reinhold, New York.

Kurasawa, I., Y. Hirose, T. Sunada, and Y. Nakamura (1988)Phase-linked modulation of excitability of presynaptic ter-minals of low-threshold afferent fibers in the inferior al-veolar nerve during cortically induced fictive masticationof the guinea pig. Brain Res., 446:113–120.

Lambert, R.W., L.J. Goldberg, and S.H. Chandler (1986) Com-parison of mandibular movement trajectories and associ-ated patterns of oral muscle electromyographic activityduring spontaneous and apomorphine-induced rhythmic jawmovements in the guinea pig. J. Neurophysiol., 55:301–315.

Lund, J.P. (1976) Oral-fascial sensation in the control of mas-tication and voluntary movements of the jaw. In: Mastica-tion and Swallowing. B.J. Sessle and A.G. Hannam, eds.University of Toronto Press, Toronto, pp. 145–152.

Lund, J.P., and P.G. Dellow (1971) The influence of interac-tive stimuli on rhythmical masticatory movements in rab-bits. Arch. Oral. Biol., 16:215–223.

Lund, J.P., and P.G. Dellow (1973) Rhythmical masticatoryactivity of hypoglossal motoneurons responding to an oralstimulus. Exp. Neurol., 40:243–246.

Lund, J.P., and K.A. Olsson (1983) The importance of reflexesand their control during jaw movement. Trends Neurosci.,6:458–463.

Lund, J.P., S. Rossignol, and T. Murakami (1981) Interac-tions between the jaw-opening reflex and mastication. Can.J. Physiol. Pharmacol., 59:683–690.

Marshall-Hall, E. (1833) On the reflex function of the me-dulla oblongata and medulla spinalis. Phil. Trans.,26:635–665.

Nakamura, Y., Y. Kubo, S. Nozaki, and M. Takatori (1976)Cortically induced masticatory rhythm and its modificationby tonic peripheral inputs in immobilized cats. Bull. TokyoMed. Dental Univ., 23:101–107.

Neter, J., W. Wasserman, and M.H. Kutner (1990) AppliedLinear Statistical Models. Irwin, Burr Ridge, IL.

Shaffer, H.B., and G.V. Laude (1985) Aquatic prey capture inambystomatid salamanders: Patterns of variation in muscleactivity. J. Morphol., 183:273–284.

Thexton, A.J. (1976) To what extent is mastication pro-grammed and independent of peripheral feedback? In: Mas-tication. D.J. Anderson, and B. Mathews, eds. Wright &Son, Bristol, UK, pp. 213–220.

Thexton, A.J., and A.W. Crompton (1989) Effect of sensoryinput from the tongue on jaw movement in normal feedingin the opossum. J. Exp. Zool., 250:233–243.

Thexton, A.J., and J. McGarrick (1987) Effect of experimen-tally elicited rhythmic oral activity on the linguodigastricreflex in the lightly anesthetized rabbit. Exp. Neurol.,96:104–117.

Thexton, A.J., and J.D. McGarrick (1994) The electromyo-graphic activities of jaw and hyoid musculature in differ-ent ingestive behaviours in the cat. Arch. Oral Biol.,39:599–612.

Thexton, A.J., R.Z. German, and A.W. Crompton (1995) La-tency of the jaw opening reflex in the immature pig. Arch.Oral. Biol., 40:1133–1135.

Wilkenson, L. (1991) SYSTAT, The system for statistics.Evanston, IL.

determinants of rhythm and rate in suckling

Documents