Archives of Oral Biology 83 (2017) 161–168

Contents lists available at ScienceDirect

Archives of Oral Biology

journal homepage: www.elsevier.com/locate/archoralbio

Relationships between chewing rate, occlusion, cephalometric anatomy, MARK

muscle activity, and masticatory performance
⁎
Brent A. Fulksa, Kathryn X. Callaghanb, Claire D. Tewksburyb, Geoffrey E. Gerstnerb,
a
Department of Orthodontics and Pediatric Dentistry, School of Dentistry, University of Michigan, Ann Arbor, MI 48109-1078, USA
b
Department of Biologic and Materials Sciences, School of Dentistry, University of Michigan, Ann Arbor, MI 48109-1078, USA

A R T I C L E I N F O A B S T R A C T

Keywords: Objective: Mastication consists of rhythmic jaw openings and closings. Recent studies suggest that muscle me-
Mastication chanical properties determine the rhythmic rate; however, speed-accuracy tradeoffs may also play a role. This
Electromyography study evaluated how variation in chewing rate affected chewing performance, how masticatory muscle activity
Eating behaviour varied with chewing rate, and whether morphology and demographics contributed to performance.
Functional morphology
Design: Chewing performance and muscle activity were sampled in 23 healthy, fully-dentate adults, who chewed
Morphometrrics
a standardized test food to a metronome set at 0.5, 0.75, 1, 2 and 3 times their ‘natural’ chewing rates. Subjects
Occlusion
produced ten chews per trial, and five trials for each of the five rates. Surface electromyographic (EMG) activity
was sampled from masseter and temporalis muscles bilaterally. Demographic, occlusal, and cephalometric data
were also obtained.
Results: Chewing performance, defined by median particle size, was inversely related to chewing rate; however,
performance was not remarkably improved at rates slower than the natural chewing rate. Above the natural
chewing rate, variability in EMG bursts diminished, suggesting a reduction in muscle activity modulation at fast
rates. Occlusal contacts and most morphological features appeared to play a limited or no role in performance.
Conclusions: Results support the hypothesis that the ‘natural’ chewing rate is selected to be as fast as possible
while providing sufficient time to allow EMG modulation for improved performance. The interplay between
EMG modulation and individual variation in skeletal morphology is likely critical for optimal chewing perfor-
mance.

1. Introduction
Mastication is a mammalian chewing method for reducing food
rapidly and efficiently (Wall & Smith, 2001). It is controlled in part by
ponto-medullary pattern generating circuitry, which has been modeled
as consisting of a central timing network that generates the rhythm or
tempo at which jaw openings and closings occur, and interneuron cir-
cuits that control the sequencing of muscle activity patterns
(Nakamura & Katakura, 1995). Recent work suggests that additional
networks are involved in rhythmogenesis as well (Morquette et al.,
2012). It is also noteworthy that most animal studies have focused on
the trigeminal system (Westberg & Kolta, 2011); however, facial and
hypoglossal motor systems (Lund & Kolta, 2006) as well as numerous
suprabulbar sites play roles during mastication (Onozuka et al., 2002;
Quintero, Ichesco, Myers, Schutt, & Gerstner, 2013; Quintero, Ichesco,
Schutt et al., 2013).
Mastication must be efficient due, in part, to the fact that the
mammalian respiratory and digestive systems share a common pathway
in the oropharynx (Matsuo & Palmer, 2009). Also, because of their high
metabolic rates relative to ectotherms (Krosniunas & Gerstner, 2003),
mammals are limited in terms of how long they can cease breathing.
Since there is a brief interruption in respiration when swallowing oc-
curs (Matsuo & Palmer, 2009), food must be well-chewed and mixed
with saliva to minimize swallowing time and the concomitant brief
cessation in breathing. However, mammals have lost the ancestral
ability to replace teeth multiple times (Whitlock & Richman, 2013);

Abbreviations: AMI, occlusal contact area in maximum intercuspation (Fig. 4); ANB, angle formed between point A nasion and point B (Fig. 2); CoGn, distance between condylion and
gnathion (Fig. 2); CV, coefficient of variation; EMG, electromyography; FMA, angle formed by Frankfort horizontal and mandibular plane (Fig. 2); GoGn, distance between gonion and
gnathion (Fig. 2); X50, median particle size; N, particle size distribution (slope of the linear part of the Rosin-Rammler equation); SNA, angle formed between sella nasion and point A
(Fig. 2); SNB, angle formed between sella nasion and point B (Fig. 2); SN-GoGn, angle formed by line segment through sella and nasion and line segment through gonion and gnathion
(Fig. 2); TC, chewing cycle duration from onset to onset of two successive EMG bursts (Fig. 3); TB, EMG burst duration or time between onset and offset of an EMG burst (Fig. 3); TP, EMG
peak onset latency or time from EMG burst onset to time of peak activity (Fig. 3); UAFH/LAFH, ratio of upper anterior face height and lower anterior face height (Fig. 2)
⁎
Corresponding author.
E-mail address: geger@umich.edu (G.E. Gerstner).

http://dx.doi.org/10.1016/j.archoralbio.2017.07.020
Received 7 May 2017; Received in revised form 15 July 2017; Accepted 26 July 2017
0003-9969/ © 2017 Elsevier Ltd. All rights reserved.
B.A. Fulks et al. Archives of Oral Biology 83 (2017) 161–168

therefore, minimizing tooth wear is critical for their survival (Ungar,

2005). Indeed, excessive tooth wear has dire fitness consequences for
mammals including reduced fecundity (King et al., 2005; Wright, King,
Baden, & Jernvall, 2008) and increased mortality (Kojola, Helle,
Huhta, & Niva, 1998; Veiberg et al., 2007). Hence, mastication must
address the competing challenges of reducing food for safe deglutition
while minimizing tooth wear that occurs with food reduction (Ross,
Washington et al., 2009; Williams et al., 2011).
One unique feature of mastication is its relatively invariant rhythm
(Ekuni, Furuta, Takeuchi, Tomofuji, & Morita, 2012; Ross et al., 2007).
It was long believed that chewing rate was linked to jaw mass or lever
arm biomechanics (Druzinsky, 1993; Gerstner, Madhavan, & Braun,
2014; Ross, Reed et al., 2009), but recent work suggests this is not the
primary relationship (Carvalho & Gerstner, 2004; Gerstner,
Cooper, & Helvie, 2010; Ross et al., 2017; Ross, Reed et al., 2009;
Stover & Williams, 2011). Muscle mechanics and salivary flow rate have Fig. 1. Procedural steps for each subject.

been proposed as playing roles (Virot, Ma, Clanet, & Jung, 2017).
However, we posit that speed-accuracy trade-offs may also play a
role, based on the need to minimize tooth wear from food reduction and
to reduce food sufficiently for safe swallowing, as discussed above.
Specifically, modulation of bite force and jaw kinematics requires
feedback from oral proprioceptors and mechanoreceptors
(Lund & Kolta, 2006; Shimada et al., 2012; Svensson & Trulsson, 2011).
Neuronal conduction velocity limits exist (More et al., 2010), and
communication between peripheral and central trigeminal, facial and
hypoglossal cranial nerve systems is required for masticatory modula-
tion; therefore, chewing rate will be limited by time lags associated
with requisite feedback and feedforward. Additionally, evidence sug-
gests that speed-accuracy tradeoffs are governed by neurodynamical
limits within central motor systems (Beamish, Bhatti, MacKenzie, & Wu,
2006); therefore, if suprabulbar central systems are involved in masti-
catory modulation, there will be time delays associated with involve-
ment of these systems as well.
This study was undertaken to shed some light on these issues. We
hypothesize that, as for most human motor behaviors, speed-accuracy
tradeoffs exist (Beamish et al., 2006), i.e., slower chewing will allow
more accurate kinematics, thus improving the per-chew rate of food
reduction. But, because feeding competes for time with other behaviors,
there are practical limits on how slow chewing rate can be. Hence, we
hypothesize that chewing rate is just slow enough to allow feedback to
improve accuracy and performance. We further hypothesize that spe-
cific morphological features, e.g., short faces (Kiliaridis, Johansson,
Haraldson, Omar, & Carlsson, 1995; Proffit, Fields, & Nixon, 1983);
(Hartstone-Rose, Perry, & Morrow, 2012) and large occlusal contact
areas, will account for variation in chewing performance across sub-
jects.
Inspired by previous studies (Buschang, Throckmorton,
Travers, & Johnson, 1997; Plesh, Bishop, & McCall, 1987; Sanchez-
Ayala, Farias-Neto, Campanha, & Garcia, 2013; Throckmorton,
Buschang, Hayasaki, & Phelan, 2001), we manipulated chewing rate to
study how rate affects performance. We further examined how chewing
rate impacts muscle activity and how muscle activity impacts perfor-
mance. Finally, the study considered the roles of demographic, occlusal
and cephalometric variation in performance as well.
2. Materials and methods
2.1. Subjects
Twenty-three healthy, fully dentate adults served as subjects.
Procedures, in which all subjects were involved, are summarized in
Fig. 1 and described in more detail, below. Subjects’ rights were pro-
tected by the University of Michigan Medical IRB (IRB-MED), and
written informed consent was obtained from all subjects. Demographic
data and informed consent were obtained after subjects were screened
for the following inclusion criteria: (1) no chewing side preferences, (2)
no chewing difficulties, (3) no gum chewing habits, (4) no orthodontic
work within the previous year, (5) no temporomandibular disorders
(TMD), as defined by RDC-TMD criteria (Dworkin & LeResche, 1992),
(6) no musculoskeletal, gastrointestinal, nor neurological conditions,
(7) no medication use with known oral motor side effects, (8) no history
of eating disorders, (9) no recent radiation exposure. Intraoral exams
confirmed Angle’s Class I molar relationships and presence of full
dentitions, sans third molars.
2.2. Cephalometrics
Lateral cephalographs were taken from the right side. Landmarks on
the skull and mandible were digitized with proprietary software
(Dolphin Imaging, v.11.7, Dolphin Imaging and Management Solutions,
Chatsworth, CA), from which standard clinical and custom measure-
ments were made (Fig. 2). One orthodontically-trained investigator
(B.F.) took two readings of each landmark, which were averaged before
calculating variables.
2.3. Surface electromyography (EMG) signal recording and filtering
EMG activity of left and right superficial masseter and anterior
temporalis muscles was measured with surface electrodes in bipolar
configuration (Ag/AgCl, 1.8-m snap-on leads, MVAP Medical Supplies,
Newbury Park, CA). A ground electrode was placed over the left mas-
toid process. Data were digitized at 1 kHz (Octal Bioamp, PowerLab 8/
35, LabChart Pro v. 8.0.4, ADInstruments, Colorado Springs, CO), then
processed and filtered as per (Ives & Wigglesworth, 2003), viz., band-
pass filtered (20–500 Hz), notch-filtered (60 Hz), full-wave rectified,
and smoothed with a moving average window set to 5% of the sampling
rate.
2.4. Test food
Test food tablets (CutterSil Putty Plus, Universal Plus Hardener,
Heraeus Kulzer, South Bend, IN) were formed using a circular template
(Plexiglas, 4.76-mm thickness, 12-mm diameter), allowed to harden for
1 h, trimmed to remove flash, and weighed. Only tablets weighing
0.85 g + 0.05 g were used.
2.5. Experimental procedure
2.5.1. Subject chewing rates
Subjects were seated comfortably in a chair facing a table. Materials
for subjects to use during the experiment, e.g., Styrofoam cups and test
food tablets (see below), were placed conveniently on the table.
Subjects were first instructed to chew gum (Trident Original, Mondelez
International, Deerfield, IL). After chewing for at least 15 s, chews were

162
B.A. Fulks et al. Archives of Oral Biology 83 (2017) 161–168

Fig. 2. Top left panel shows landmarks used to construct variables

depicted in remaining panels. Key. Upper left: ANS, anterior nasal
spine; A, A point; B, B point; Co, condylion; Gn, gnathion; Go,
gonion; Me, menton; N, nasion; O, orbitale; P, porion; S, sella.
Upper right: SNA, SNB, ANB, angles formed between points spe-
cified in name; CoGn, GoGn, distances between points specified in
name. Lower left: SN-GoGn, angle formed between line segment
intersecting S and N and line segment intersecting Go and Gn.
Lower right: FMA, angle formed between line segment inter-
secting P and O and line segment intersecting Go and Me; UAFH,
distance from N to ANS; LAFH, distance from ANS to Me.

counted until the subject had chewed at least 30 consecutive times

without talking or swallowing. The uninterrupted number of chews and
the time (in ms) required to produce the chews were used to calculate a
mean ‘natural’ chewing rate, which was used to define five metronome
rates, viz., 0.5, 0.75, 1, 2, and 3 times ‘natural’ chewing rate, for each
subject.

2.5.2. Blocks, trials and tasks

Each subject completed five experimental blocks. Within each
block, subjects chewed test food tablets at each of the five rates to the
beat of a metronome; hence, five trials per block. Only one tablet was
chewed per trial. Chewing-rate order was randomized within blocks. No
instructions were given regarding chewing side. Subjects were told to
close to the beat of the metronome, to chew each tablet exactly ten
times (e.g., Fig. 3, top), and then to spit out all particles into a Styr-
ofoam cup (Fig. 4). A cup of water was supplied to use as necessary to
remove all particles between trials. Trials not conforming to all of the
above specifications were redone. By the end, each subject had chewed
up 25 tablets. The five trials representing a given chewing rate con-
stituted a task.

2.6. EMG analysis

EMG burst onsets and offsets occurred when EMG voltages went
10% above and below baseline voltages, respectively (Fig. 3). The time
Fig. 3. Top: Right temporalis EMG from one trial; note 10 EMG bursts. Bottom: Depiction
between two successive EMG burst onsets defined chewing cycle period of EMG variables, chewing cycle duration (TC), EMG burst duration (TB), and EMG peak
(TC); nine TC were calculated per muscle per trial. EMG burst duration onset latency (TP).
(TB), and peak amplitude onset latency (TP) were calculated for each
chewing cycle; ten TB and TP were calculated per muscle per trial.
2.7. Test food particle treatment and analysis
Within-trial variation in TC, TB, and TP was defined using the coefficient
of variation (CV) = standard deviation/mean for each muscle and trial.
Test food particles in each cup (Fig. 4) were washed, disinfected, left
Because we did not control chewing side, variables were averaged
to dry for twenty-four hours, and then separated through a series of
across left and right muscle pairs as well as across the five trials re-
seven sieves with mesh sizes of 5.6, 4.0, 2.8, 2.0, 0.85, 0.425, and
presenting a given chewing rate.

163
B.A. Fulks et al.
0.25 mm (U.S.A. Standard Test Sieves, Hogentogler & Co., Inc, Co-
lumbia, MD). Sieves were shaken for two minutes to aid particle se-
paration (Gilson Company, Inc, Lewis Center, OH). Contents in each
sieve were weighed to the nearest 0.01 g.
Chewing performance was measured using the Rosin-Rammler
equation, Qw = 100[1 − 2−(D/X
50
)N
] (Rosin & Rammler, 1933), where
Qw is the weight percentage of particles passing through a sieve size
smaller than D, X50 is median particle size (in mm), and N is the slope of
the linear part of the fitted equation (Olthoff, van der Bilt,
Bosman, & Kleizen, 1984). A small X50 indicates good performance, and
a small N indicates a broader particle size distribution (compare ex-
amples, Fig. 4). Calculations were made using an algorithm in MatLab
(Brezani & Zelenak, 2010).
2.8. Wax bites and occlusion analysis
Wax bite impressions were made by having subjects bite in max-
imum intercuspation (MI) on bite registration wax (Hygenic Yellow Bite
wax–Coltene/Whaledent). The impressions were placed on a con-
trasting black background and photographed from a distance of 1 m,
providing high-contrast areas representing tooth contacts (Fig. 4).
Millimeter rulers placed in-plane with the wax were used for calibra-
tion.
Eight-bit versions of wax images were used with a color threshold of
90 to digitize total premolar and molar bite contact areas (ImageJ
software, NIH, Bethesda, MD). Two researchers independently calcu-
lated contact areas, which were then averaged together for analysis.
Inter-rater reliability was also evaluated, which showed good inter-
rater agreement (Pearson’s product moment, r2 = 0.997; slope = 1.02;
intercept = −0.21).
2.9. Statistical procedures
General linear mixed models with repeated measures were used to
perform statistical tests (SPSS, v.24, Chicago, IL). Subject was treated as
a random effect, task as a repeated measure, and all continuous
164
Archives of Oral Biology 83 (2017) 161–168
Fig. 4. Top row: Chewed test food particles from
which chewing performance variables, median par-
ticle size (X50) and particle-size distribution (N),
were calculated. Left side, example of relatively poor
performance; right side, example of relatively good
performance. Bottom. Wax occlusal impression, from
which the area of tooth contact in maximum inter-
cuspation (AMI) was calculated. Black line distin-
guishes posterior contacts from anterior contacts;
only posterior contacts were used in calculations.
variables were entered as covariate fixed effects. A p < 0.05 defined
the level of significance. Pairwise comparisons from models reaching
statistical significance were made with a Bonferroni correction.
Multiple linear regression models, using the stepwise procedure, were
performed to identify variable subsets that best fitted variation in
performance. The resultant beta coefficients were evaluated for sig-
nificance. Tolerance values were also obtained to evaluate collinearity
among variables in the model. Some additional correlation analyses
were performed in Excel (MS Office 10, Ver 14), see Figs. 6 and 7.
3. Results
Task rate had a significant impact on X50 (F [4,20.3] = 14.6,
p < 0.001) and N (F [4,26.5] = 3.5, p < 0.02), Fig. 5. Pairwise
comparisons revealed that 0.5x and 0.75x task X50 were similar to each
other, but significantly smaller than those of the 2x and 3x tasks
(P < 0.007). No other pairwise comparisons were significant for either
the X50 or N analyses; however, Fig. 5 shows a trend for N to be greater
for 2x and 3x trials.
Table 1 shows the relationship of other variables to X50. (TC, TB, TP
were not included in this modelling due to strong collinearities with
task rate.) Of the cephalometric variables, significant effects were seen
for ANB, SNA, SNB and UAFH/LAFH. Results for effects on N are not
shown as no significant effects occurred.
Table 2 shows multiple linear regression analyses using X50 and N as
dependent variables. Three variables explained ∼32% of the variance
in X50, namely task, SNA, and UAFH/LAFH. Tolerance values showed
that each variable was independent of the others. Note that X50 was
positively correlated with task rate, i.e., larger X50 occurred with faster
chewing rates, corroborating findings in Fig. 5. X50 was also positively
correlated with UAFH/LAFH, meaning that smaller X50 (better perfor-
mance) was associated with subjects having lower UAFH/LAFH ratios.
However, X50 was negatively correlated SNA. This meant that smaller
X50 (better performance) was associated with larger SNA. Only task
explained variance in particle size distribution (N), with broader dis-
tribution (lower slope) being associated with slower task rates;
B.A. Fulks et al. Archives of Oral Biology 83 (2017) 161–168

Fig. 5. Histograms of median particle size (X50, left) and

particle size distribution, (N, right) for the 5 tasks. Horizontal
bars above plots indicate significance (p < 0.007). Error
bars = 1 SD.

however, only 5.2% of the variation was so explained.
Fig. 6A–C shows pairwise comparisons among the three EMG timing
variables. Non-linear correlations were used to demonstrate ‘best fit’
trends. Arrows in A–C indicate a transition from ‘tight’ bivariate re-
lationships for tasks faster than 1x (2x and 3x, collectively represented
by black circles), to looser relationships for 1x (gray circles) and slower
tasks (0.5x and 0.75x, collectively represented by open circles).
The ‘loosening’ of the bivariate relationships seen in Fig. 6A–C was
further confirmed by comparing chewing rate (TC) across all five rates
vs. the within-trial coefficient of variation (CV) for TC and TP (Fig. 6D,
E, respectively). The results showed that, whereas the CV for TC did not
systematically increase as chewing rate decreased (p = 0.41), the CV
for TP did (p < 0.00001). Note also that the CV for TC appears to
average between 0.10 and 0.15 (Fig. 6D), whereas the CV for TP is much
greater at all rates, especially so at slower rates (Fig. 6E). As Fig. 6F
shows, the CV for TP differed significantly across tasks (F(4,29.8)
= 22.03, p < 0.000001), with pairwise comparisons demonstrating
that the CV was significantly less for the 2x and 3x tasks compared with
the CV for 0.5x, 0.75x tasks (p < 0.001). Also, the 2x task CV was
significantly less than that for the 1x task (p < 0.001). No significant
differences existed between the CVs for 0.5x, 0.75x and 1x tasks
(p > 0.05).
Fig. 7 compares how X50 varied as TC, TB, and TP varied. Results
demonstrate a levelling out of X50 values as durations of each of these
variables increased. The results are less obvious for TC (Fig. 7A) than for
the other two variables, with a clear nadir in X50 occurring when TB was
between 400–600 ms (Fig. 7B) and when TP was between 200–400 ms
(Fig. 7C).
4. Discussion
Chewing rate significantly affected performance as defined by X50
(Fig. 5). The effect was not evident statistically for N, partly due to
substantial variation (error bars, Fig. 5, right). We expected perfor-
mance to improve at reduced chewing speeds, based on previous stu-
dies (Buschang et al., 1997; Plesh et al., 1987; Sanchez-Ayala et al.,
2013; Throckmorton et al., 2001), and on speed-accuracy tradeoffs
underlying human motor performance (Fitts, 1954), and hypothesized
neuromotor limits (Beamish et al., 2006; More et al., 2010).
Our results (Fig. 6) corroborate previous findings that masseter EMG
bursts and inter-burst intervals decrease with increased chewing rate
(Plesh et al., 1987). In addition, our study demonstrates that intra-EMG
burst variables (i.e., TB and TP) appear relatively ‘tightly’ coupled to
chewing at fast rates and relatively loosely coupled to chewing at slow

Fig. 6. Analyses involving EMG temporal variables and X50. Shown are masseter data; similar data were obtained for the temporalis. Text convention in lower right of A–E is: abscissa vs.
ordinate, regression equation, variance (R2). In A–E, black circles = pooled 2x and 3x tasks, gray circles = 1x tasks, open circles = pooled 0.5x and 0.75x tasks. Abscissa in A–E = time
in ms. Ordinates in A–C = time in ms; D–F = coefficient of variation (CV) for TC (D), or TP (E,F). The trend-line method used in A–E was the method that accounted for the most variance,
linear, logarithm, or polynomial; these regressions are provided mainly to demonstrate trends. F. Error bars = 1 SD; horizontal lines represent statistically significant pair-wise differences
(p < 0.001).

165
B.A. Fulks et al. Archives of Oral Biology 83 (2017) 161–168

Fig. 7. Scatterplots and correlations involving EMG temporal variables and X50. Shown are masseter data; similar data were obtained for the temporalis. Text convention in lower right of
each plot is: abscissa vs. ordinate, regression equation, variance (R2). Black circles = pooled 2x and 3x tasks; gray circles = 1x tasks; open circles = pooled 0.5x and 0.75x tasks. Abscissa
in all plots = time in ms. Ordinates are X50 in mm. The trend-line method used was the method that accounted for the most variance, linear, logarithm, or polynomial; these regressions
are provided mainly to demonstrate trends (see text).

Table 1 or central processing delays are likely to be the rate-limiting features

Relationship between independent variables and X50. that determine the chewing rates at which EMG burst modification can
occur.
Variable N Meana SD df F p-value
Chewing rates slower than 1x were associated with continued re-
Task 115 See Fig. 5 See Fig. 5 4, 18.8 14.72 0.000013 duction in X50 (Fig. 7A), which corroborates findings in previous stu-
Gender 14F; 9 M na na 1, 4.5 0.86 0.40 dies, e.g., (Buschang et al., 1997; Plesh et al., 1987; Sanchez-Ayala
Age 23 23.4 3.5 1, 4.5 0.00011 0.99 et al., 2013; Throckmorton et al., 2001). However, when X50 was
Height 23 171.5 11.6 1, 4.5 0.50 0.52
Weight 23 67.1 12.5 1, 4.5 0.020 0.89
evaluated in association with intra-burst EMG variables, there was a
AMIb 22 75.6 29.5 1, 4.5 0.11 0.75 clearer picture of diminishing returns on performance at slower rates,
FMAc 23 22.6 6.2 1, 4.5 3.17 0.14 i.e., the non-linear best-fit correlations in Fig. 7B,C suggest little change
ANBc 23 1.3 2.1 1, 4.5 8.06 0.041 in X50 at the slowest rates (0.5x and 0.75x, collectively represented by
SNAc 23 80.6 3.0 1, 4.5 8.18 0.040
open circles). This supports our hypothesis that the natural chewing
SNBc 23 79.4 3.1 1, 4.5 8.15 0.040
SN-GoGnc 23 28.9 6.6 1, 4.5 0.33 0.60 rate is as fast as possible with minimal loss in performance. Moreover,
CoGnc 23 125.9 9.2 1, 4.5 1.16 0.34 the increased variation observed for TB and TP at slower rates
GoGnc 23 86.9 5.6 1, 4.5 0.33 0.59 (Fig. 6A–C, E,F) suggests that the ability to modify EMG activity on a
UAFH/LAFHc 23 81.3 7.5 1, 4.5 9.78 0.030 chew-by-chew basis is likely to be important in improving performance.
a Bite force and muscle fatigue probably impact performance; how-
Units: AMI, in mm2; FMA, ANB, SNA, SNB, SN-GoGn in degrees; GoGn, CoGn, in mm;
UAFA/LAFH, a ratio. ever, we could not directly evaluate these. Higher bite forces are related
b
Key:See text and Fig. 4 for variable description. to shorter vertical face dimensions (Kiliaridis et al., 1995; Proffit et al.,
c
Key: See Fig. 2 for variable descriptions. 1983), due partly to biomechanics of the skeletal-muscle system
(Hartstone-Rose et al., 2012). The mechanical advantage of a muscle is
rates (Fig. 6). Additionally, variation in chewing rhythm (TC) appears to the ratio of its moment arm to the moment arm of the load. Increasing
be independent of chewing rate (Fig. 6D), whereas TP increased in the gonial angle lengthens the mandibular load arm but not the muscle
variability as chewing rate decreased (Fig. 6E). Presumably, TC, which moment arm. As the gonial angle increases, the mechanical advantage
represents the masticatory rhythm, is controlled by a central rhythm of masticatory muscles decreases. We believed that this variation in
generator, whereas other neuronal populations produce EMG burst mechanical advantage would be best captured in FMA and SN-GoGn;
activity patterns, e.g., TB and TP (Nakamura & Katakura, 1995). Hence, however, these variables were not significantly related to X50 or N.
the results in Fig. 6 suggest growing independence in the coupling Rather, four other cephalometric variables contributed to variation in
between the rhythm-generating circuits producing TC and those pro- X50, viz., ANB, SNA, SNB, and UAFH/LAFH (Table 1). Because ANB,
ducing TB and TP as chewing rate decreased. We hypothesize that, be- SNA and SNB are obviously inter-related (Fig. 2), the stepwise multiple
ginning near the ‘natural’ chewing rate, sufficient time exists for mod- linear regression procedure (Table 2), was used to capture the variable
ulatory input of EMG activity to occur, which allows for the observed subset that best explained variation in X50. In this case, better perfor-
increase in EMG burst variation vs. chewing rate variation. We hy- mance (lower X50) was associated with subjects with either large SNA
pothesize that neuronal conduction velocities (More et al., 2010) and/ angles or low UAFH/LAFH ratios. Given the fact that none of the lower

Table 2
Stepwise multiple linear regression results.

Dependent variable Independent Variable Adjusted r2 Betab p-value Zero-order correlation Partial Correlation Tolerance

X50 Constant
Task 0.17 0.42 0.000005 0.42 0.47 1.0
SNAa 0.3 −0.33 0.000324 −0.37 −0.38 0.96
UAFH/LAFH 0.324 0.18 0.04592 0.25 0.21 0.96

N
Constant
Task 0.052 0.25 0.017 0.25 0.25 1.0

a
See Fig. 2 for description of variables.
b
Standardized coefficient reported.

166
B.A. Fulks et al.
face variables, viz., SN-GoGN, GoGn, CoGn, FMA, was significantly
associated with variation in X50, variation in upper face morphology
may somehow have impacted performance. Specific ways in which
UAFH/LAFH and SNA can vary in association with upper face include
the following. First, given that subjects with relatively small UAFH/
LAFH ratios had smaller X50, it is possible that subjects with relatively
small upper anterior face heights had relatively better performance.
Likewise, subjects with large SNA had smaller X50. Inspection of Fig. 2
will demonstrate that moving N towards A along line segment N-A or
moving N along line segment S-N towards S are ways of increasing
angle SNA without changing lower face or jaw sizes. However, further
work will be required to identify the functional advantages afforded by
the observed results. We would hypothesize that UAFH/LAFH and ANS
are surrogates for morphological features related to the orientation and
length of fibers in the superficial masseter and anterior temporalis
muscles, a hypothesis supported by other findings relating anterior
masseter muscle contractile thickness to anterior face height and bite
force (Bakke et al., 1992).
Finally, occlusal contacts did not play a role in performance
(Table 1), which corroborates some previous results (Fontijn-Tekamp,
van der Bilt, Abbink, & Bosman, 2004; Wilding, 1993) but not all
(Owens, Buschang, Throckmorton, Palmer, & English, 2002). Occlusal
area, as we measured it, does not capture three-dimensional functional
relationships, including loading, stress and strain regimes, which are
likely to be more important than static, two-dimensional measurements
(Ross, Iriarte-Diaz, & Nunn, 2012). A three-dimensional occlusal as-
sessment would provide a better test of occlusion-performance re-
lationships.
There were several study weaknesses that should be mentioned.
First, visual observation revealed varying chewing strategies across
subjects at rates slower than 1x. For instance, some appeared to reduce
jaw opening speed, but to close relatively fast to the metronome beat.
Others seemed to reduce both jaw open and closing rates. If speed-ac-
curacy tradeoffs play a role in performance, they would likely play the
most significant role during closing. Therefore, it would be useful to
monitor jaw movements in a future study and to instruct subjects to
perform a specific jaw movement profile so that EMG and jaw move-
ment behaviour would be more similar among subjects.
Second, we did not control for chewing side, which necessitated
pooling of bilateral muscle data. Future studies should control or at last
account for chewing side, so that insights differentiating working from
balancing side muscle activity would be possible.
Third, our subjects varied little demographically. Future studies
should broaden subject demographics so that developmental and aging
components of performance can be assessed.
Fourth, we recognize that the rhythmic jaw movements we eval-
uated are self-conscious and performed to the auditory beat of a me-
tronome. As such, the neuromotor mechanisms subserving them are
likely to be at least partially different from those involved with spon-
taneous or ‘natural’ mastication. This ‘unnatural’ quality should be
taken into consideration when evaluating the data. It is also likely that
the artificial setup, including the forced changes in chewing tempo,
greatly influenced tongue and facial muscle activity patterns in ways
that would not be observed under routine conditions.
Fifth, EMG data may have been impacted by at least two confounds,
variation in inter-electrode distance and movement artifact. We care-
fully controlled for these by filtering signals to remove movement ar-
tifact and standardizing inter-electrode distances between muscles and
subjects according to ‘best practices’ (Stepp, 2012). However, it is likely
that many muscle contraction parameters varied systematically with
chewing rate changes, and that these could have caused systematic
changes in inter-electrode distances and in movement artifact. Never-
theless, it is important to recognize that, even with changes in inter-
electrode distance, activity from superficial fibers still dominate the
signal, especially when EMG is recorded with small electrodes and with
small inter-electrode distances (Stepp, 2012), both of which we used in
167
Archives of Oral Biology 83 (2017) 161–168
this study. Moreover, the variables selected for analysis, viz., TC, TB and
TP, are probably less influenced by these confounds than, say, EMG
amplitude (Stepp, 2012). It is also important to recognize that data
were averaged across working and balancing muscle pairs, and that our
results were similar for temporalis and masseter muscles. It is less likely
that these confounds could have had consistent effects across muscle
pairs, across muscles and across subjects. Therefore, we believe the
EMG data likely reflect mainly activity of the target muscles.
Finally, muscle fatigue could have been a potential confound.
However, subjects had 30-s recovery periods between trials, and we
varied the order of trials in blocks. Thus, muscle fatigue probably did
not confound reported results.
5. Conclusions
Chewing rate significantly impacted performance, with slower
chewing rates being associated with smaller particle sizes and a broader
distribution or spread of particle sizes. Rates faster than ‘natural’
chewing rate were considerably worse in performance, whereas rates
slower than ‘natural’ chewing rate were only slightly better in terms of
performance than the ‘natural’ rate. Muscle activity was impacted by
chewing rate, with evidence suggesting that within-trial variation in
EMG intra-burst variables increased at rates near to and slower than
‘natural’ chewing rate. Results support the hypothesis that the ‘natural’
chewing rate is selected to be as fast as possible while providing suf-
ficient time to allow EMG modulation for improved performance. The
interplay between EMG modulation and individual variation in skeletal
morphology is likely critical for optimal chewing performance.
Acknowledgements
This work was supported by Delta Dental Foundation of Michigan
[grant number 210790], a University of Michigan LeGro Research
Fellowship, and a University of Michigan Rackham Graduate Student
Research Grant. The authors wish to thank Dr. Shyamala Nagaraj and
Josh Errickson at the University of Michigan Center for Statistical
Consultation and Research (CSCAR) for statistical assistance. The au-
thors have no conflict of interests to report and received no consulting
fees, nor were any of them involved in financial arrangements that
would impact the study.
References
Bakke, M., Tuxen, A., Vilmann, P., Jensen, B. R., Vilmann, A., & Toft, M. (1992).
Ultrasound image of human masseter muscle related to bite force, electromyography,
facial morphology, and occlusal factors. Scandinavian Journal of Dental Research,
100(3), 164–171.
Beamish, D., Bhatti, S. A., MacKenzie, I. S., & Wu, J. (2006). Fifty years later: A neuro-
dynamic explanation of Fitts' law. Journal of The Royal Society Interface, 3(10),
649–654.
Brezani, I., & Zelenak, F. (2010). Improving the effectivity of work with Rosin-Rammler
diagram by using MATLAB GUI tool. Acta Montanistica Slovaca, 15(2), 152–157.
Buschang, P. H., Throckmorton, G. S., Travers, K. H., & Johnson, G. (1997). The effects of
bolus size and chewing rate on masticatory performance with artificial test foods.
Journal of Oral Rehabilitation, 24(7), 522–526.
Carvalho, T. C., & Gerstner, G. E. (2004). Licking rate adaptations to increased man-
dibular weight in the adult rat. Physiology and Behavior, 82(2–3), 331–337.
Druzinsky, R. E. (1993). The time allometry of mammalian chewing movements: Chewing
frequency scales with body mass in mammals. Journal of Theoretical Biology, 160(4),
427–440.
Dworkin, S. F., & LeResche, L. (1992). Research diagnostic criteria for tempor-
omandibular disorders: Review, criteria, examinations and specifications, critique.
Journal of Craniomandibular Disorders, 6(4), 301–355.
Ekuni, D., Furuta, M., Takeuchi, N., Tomofuji, T., & Morita, M. (2012). Self-reports of
eating quickly are related to a decreased number of chews until first swallow, total
number of chews, and total duration of chewing in young people. Archives of Oral
Biology, 57(7), 981–986.
Fitts, P. M. (1954). The information capacity of the human motor system in controlling
the amplitude of movement. Journal of Experimental Psychology: Human
Perception & Performance, 47(6), 381–391.
Fontijn-Tekamp, F. A., van der Bilt, A., Abbink, J. H., & Bosman, F. (2004). Swallowing
threshold and masticatory performance in dentate adults. Physiology and Behavior,
B.A. Fulks et al.
83(3), 431–436.
Gerstner, G. E., Cooper, M., & Helvie, P. (2010). Chewing rates among domestic dog
breeds. Journal of Experimental Biology, 213, 2266–2272.
Gerstner, G. E., Madhavan, S., & Braun, T. M. (2014). Relationships between masticatory
rhythmicity, body mass and cephalometrically-determined aesthetic and functional
variables during development in humans. Archives of Oral Biology, 59(7), 711–721.
Hartstone-Rose, A., Perry, J. M. G., & Morrow, C. J. (2012). Bite force estimation and the
fiber architecture of felid masticatory muscles. Anatomical Record-Advances in
Integrative Anatomy and Evolutionary Biology, 295(8), 1336–1351.
Ives, J. C., & Wigglesworth, J. K. (2003). Sampling rate effects on surface EMG timing and
amplitude measures. Clinical Biomechanics, 18(6), 543–552.
Kiliaridis, S., Johansson, A., Haraldson, T., Omar, R., & Carlsson, G. E. (1995).
Craniofacial morphology, occlusal traits, and bite force in persons with advanced
occlusal tooth wear. American Journal of Orthodontics & Dentofacial Orthopedics,
107(3), 286–292.
King, S. J., Arrigo-Nelson, S. J., Pochron, S. T., Semprebon, G. M., Godfrey, L. R., Wright,
P. C., et al. (2005). Dental senescence in a long-lived primate links infant survival to
rainfall. Proceedings of the National Academy of Sciences of the United States of America,
102(46), 16579–16583.
Kojola, I., Helle, T., Huhta, E., & Niva, A. (1998). Foraging conditions, tooth wear and
herbivore body reserves: A study of female reindeer. Oecologia, 117, 26–30.
Krosniunas, E., & Gerstner, G. (2003). A model of vertebrate resting metabolic rate:
Balancing energetics and O2 transport in system design. Respiratory Physiology and
Neurobiology, 134, 93–113.
Lund, J. P., & Kolta, A. (2006). Generation of the central masticatory pattern and its
modification by sensory feedback. Dysphagia, 21(3), 167–174.
Matsuo, K., & Palmer, J. B. (2009). Coordination of mastication, swallowing and
breathing. Japanese Dental Science Review, 45(1), 31–40.
More, H. L., Hutchinson, J. R., Collins, D. F., Weber, D. J., Aung, S. K., & Donelan, J. M.
(2010). Scaling of sensorimotor control in terrestrial mammals. Proceedings of the
Royal Society of London B: Biological Sciences, 277(1700), 3563–3568.
Morquette, P., Lavoie, R., Fhima, M. D., Lamoureux, X., Verdier, D., & Kolta, A. (2012).
Generation of the masticatory central pattern and its modulation by sensory feed-
back. Progress in Neurobiology, 96(3), 340–355.
Nakamura, Y., & Katakura, N. (1995). Generation of masticatory rhythm in the brainstem.
Neuroscience Research, 23, 1–19.
Olthoff, L. W., van der Bilt, A., Bosman, F., & Kleizen, H. H. (1984). Distribution of
particle sizes in food comminuted by human mastication. Archives of Oral Biology,
29(11), 899–903.
Onozuka, M., Fujita, M., Watanabe, K., Hirano, Y., Niwa, M., Nishiyama, K., et al. (2002).
Mapping brain region activity during chewing: A functional magnetic resonance
imaging study. Journal of Dental Research, 81(11), 743–746.
Owens, S., Buschang, P. H., Throckmorton, G. S., Palmer, L., & English, J. (2002).
Masticatory performance and areas of occlusal contact and near contact in subjects
with normal occlusion and malocclusion. American Journal of
Orthodontics & Dentofacial Orthopedics, 121(6), 602–609.
Plesh, O., Bishop, B., & McCall, W. (1987). Mandibular movements and jaw muscles'
activity while voluntarily chewing at different rates. Experimental Neurology, 98(2),
285–300.
Proffit, W. R., Fields, H. W., & Nixon, W. L. (1983). Occlusal forces in normal- and long-
face adults. Journal of Dental Research, 62(5), 566–570.
Quintero, A., Ichesco, E., Myers, C., Schutt, R., & Gerstner, G. E. (2013). Brain activity and
human unilateral chewing: An fMRI study. Journal of Dental Research, 92(2),
136–142.
Quintero, A., Ichesco, E., Schutt, R., Myers, C., Peltier, S., & Gerstner, G. E. (2013).
Functional connectivity of human chewing: An fcMRI study. Journal of Dental
Research, 92(3), 272–278.
168
Archives of Oral Biology 83 (2017) 161–168
Rosin, P., & Rammler, E. (1933). Gesetzmassigkeiten in der Kornzusammensetzung des
Zementes. Zement, 31, 427–433.
Ross, C. F., Eckhardt, A., Herrel, A., Hylander, W. L., Metzger, K. A., Schaerlaeken, V.,
et al. (2007). Modulation of intra-oral processing in mammals and lepidosaurs.
Integrative and Comparative Biology, 47, 118–136.
Ross, C. F., Iriarte-Diaz, J., & Nunn, C. L. (2012). Innovative approaches to the re-
lationship between diet and mandibular morphology in primates. International
Journal of Primatology, 33(3), 632–660.
Ross, C. F., Iriarte-Diaz, J., Platts, E., Walsh, T., Heins, L., Gerstner, G. E., et al. (2017).
Scaling of rotational inertia in primate mandibles. Journal of Human Evolution, 106,
119–132.
Ross, C. F., Reed, D. A., Washington, R. L., Eckhardt, A., Anapol, F., & Shahnoor, N.
(2009). Scaling of chew cycle duration in primates. American Journal of Physical
Anthropology, 138, 30–44.
Ross, C. F., Washington, R. L., Eckhardt, A., Reed, D. A., Vogel, E. R., Dominy, N. J., et al.
(2009). Ecological consequences of scaling of chew cycle duration and daily feeding
time in primates. Journal of Human Evolution, 56(6), 570–585.
Sanchez-Ayala, A., Farias-Neto, A., Campanha, N. H., & Garcia, R. C. M. R. (2013).
Relationship between chewing rate and masticatory performance. Journal of
Craniomandibular Practice, 31(2), 118–122.
Shimada, A., Yamabe, Y., Torisu, T., Baad-Hansen, L., Murata, H., & Svensson, P. (2012).
Measurement of dynamic bite force during mastication. Journal of Oral Rehabilitation,
39(5), 349–356.
Stepp, C. E. (2012). Surface electromyography for speech and swallowing systems:
Measurement, analysis, and interpretation. Journal of Speech, Language, and Hearing
Research, 55, 1232–1246.
Stover, K. K., & Williams, S. H. (2011). Intraspecific scaling of chewing cycle duration in
three species of domestic ungulates. Journal of Experimental Biology, 214, 104–112.
Svensson, K. G., & Trulsson, M. (2011). Impaired force control during food holding and
biting in subjects with tooth- or implant-supported fixed prostheses. Journal of
Clinical Periodontology, 38(12), 1137–1146.
Throckmorton, G. S., Buschang, B. H., Hayasaki, H., & Phelan, T. (2001). The effects of
chewing rates on mandibular kinematics. Journal of Oral Rehabilitation, 28(4),
328–334.
Ungar, P. S. (2005). Reproductive fitness and tooth wear: Milking as much as possible out
of dental topographic analysis. Proceedings of the National Academy of Sciences,
102(46), 16533–16534.
Veiberg, V., Mysterud, A., Gaillard, J. M., Delorme, D., Van Laere, G., & Klein, F. (2007).
Bigger teeth for longer life? Longevity and molar height in two roe deer populations.
Biology Letters, 3(3), 268–270.
Virot, E., Ma, G., Clanet, C., & Jung, S. (2017). Physics of chewing in terrestrial mammals.
Scientific Reports, 7, 43967.
Wall, C. E., & Smith, K. K. (2001). Ingestion in mammals. Encyclopedia of Life
Sciences, 1–6.
Westberg, K. G., & Kolta, A. (2011). The trigeminal circuits responsible for chewing.
International Review of Neurobiology, 97, 77–98.
Whitlock, J. A., & Richman, J. M. (2013). Biology of tooth replacement in amniotes.
International Journal of Oral Science, 5(2), 66–70.
Wilding, R. J. (1993). The association between chewing efficiency and occlusal contact
area in man. Archives of Oral Biology, 38(7), 589–596.
Williams, S. H., Vinyard, C. J., Wall, C. E., Doherty, A. H., Crompton, A. W., & Hylander,
W. L. (2011). A preliminary analysis of correlated evolution in mammalian chewing
motor patterns. Integrative and Comparative Biology, 51(2), 247–259.
Wright, P., King, S. J., Baden, A., & Jernvall, J. (2008). Aging in wild female lemurs:
Sustained fertility with increased infant mortality. Primate Reproductive Aging: Cross-
Taxon Perspectives, 36, 17–28.