3D enamel profilometry reveals faster growth but similar stress severity in Neanderthal versus Homo sapiens teeth

Abstract

Early life stress disrupts growth and creates horizontal grooves on the tooth surface in humans and other mammals, yet there is no consensus for their quantitative analysis. Linear defects are considered to be nonspecific stress indicators, but evidence suggests that intermittent, severe stressors create deeper defects than chronic, low-level stressors. However, species-specific growth patterns also influence defect morphology, with faster-growing teeth having shallower defects at the population level. Here we describe a method to measure the depth of linear enamel defects and normal growth increments (i.e., perikymata) from high-resolution 3D topographies using confocal profilometry and apply it to a diverse sample of Homo neanderthalensis and H. sapiens anterior teeth. Debate surrounds whether Neanderthals exhibited modern human-like growth patterns in their teeth and other systems, with some researchers suggesting that they experienced more severe childhood stress. Our results suggest that Neanderthals have shallower features than H. sapiens from the Upper Paleolithic, Neolithic, and medieval eras, mirroring the faster growth rates in Neanderthal anterior teeth. However, when defect depth is scaled by perikymata depth to assess their severity, Neolithic humans have less severe defects, while Neanderthals and the other H. sapiens groups show evidence of more severe early life growth disruptions.

Full text



Scientic Reports | (2021) 11:522 | 
www.nature.com/scientificreports
3D enamel prolometry reveals
faster growth but similar
stress severity in Neanderthal
versus Homo sapiens teeth
Kate McGrath1,2*, Laura Sophia Limmer3, Annabelle‑Louise Lockey3,
Debbie Guatelli‑Steinberg4,5, Donald J. Reid2, Carsten Witzel6, Emmy Bocaege5,
Shannon C. McFarlin2,7 & Sireen El Zaatari3
Early life stress disrupts growth and creates horizontal grooves on the tooth surface in humans
and other mammals, yet there is no consensus for their quantitative analysis. Linear defects are
considered to be nonspecic stress indicators, but evidence suggests that intermittent, severe
stressors create deeper defects than chronic, low‑level stressors. However, species‑specic growth
patterns also inuence defect morphology, with faster‑growing teeth having shallower defects at
the population level. Here we describe a method to measure the depth of linear enamel defects and
normal growth increments (i.e., perikymata) from high‑resolution 3D topographies using confocal
prolometry and apply it to a diverse sample of Homo neanderthalensis and H. sapiens anterior teeth.
Debate surrounds whether Neanderthals exhibited modern human‑like growth patterns in their teeth
and other systems, with some researchers suggesting that they experienced more severe childhood
stress. Our results suggest that Neanderthals have shallower features than H. sapiens from the Upper
Paleolithic, Neolithic, and medieval eras, mirroring the faster growth rates in Neanderthal anterior
teeth. However, when defect depth is scaled by perikymata depth to assess their severity, Neolithic
humans have less severe defects, while Neanderthals and the other H. sapiens groups show evidence
of more severe early life growth disruptions.
e incremental nature of dental development allows researchers to precisely assess the tempo and duration of
tooth growth. Aspects of dental ontogeny, including molar eruption patterns, are roughly correlated with the
pace of successive life history events like weaning, sexual maturation, and longevity in primates1, but much more
stands to be learned, particularly among closely related taxa2,3. An enduring debate surrounds when the modern
human-like life history ‘package’ appeared in our lineage, marked by a relatively longer and slower developmental
period compared to other apes4. While it is now accepted that species such as Homo erectus exhibited a shorter
developmental window than our own species, active controversy still exists in regards to the growth patterns of
our most recently extinct relative, the Neanderthal5,6.
e most reliable way to glean growth information from teeth is to physically or virtually section them, or
image already broken tooth surfaces, thus gaining access to their internal microstructure7–9. ese methods are
inaccessible for most studies due their destructive nature and/or prohibitive cost, and further restricted by their
inherent sample size limitations. Instead, many researchers analyze the near-weekly growth increments called
perikymata that outcrop on the outer tooth surface (Fig.1). ere have been many advances in the microscopic
OPEN
             
             
            
          
           
              
USA. 
            
  
 *
Content courtesy of Springer Nature, terms of use apply. Rights reserved


Vol:.(1234567890)
Scientic Reports | (2021) 11:522 | 
www.nature.com/scientificreports/
analysis of perikymata in recent years, allowing for detailed analyses of normal growth patterns as well as growth
disruptions caused by early life stress10–13.
Stressors like illness, injury, or malnutrition disrupt enamel formation, creating enamel defects on the surface
of teeth, the most common of which is linear enamel hypoplasia (LEH)14. Hypoplastic defects usually appear
as lines or grooves on the enamel surface, following the course of normal perikymata around the tooth crown,
but they can also have pits or larger areas of missing enamel15. e qualitative analysis of LEH is a mainstay of
bioarchaeology and paleobiology, with high LEH prevalence interpreted as a signal of poor living conditions
among past populations14. A major limitation is that qualitative defect identication makes it dicult to com-
pare results among studies as defects appear on a continuum. Two methods for quantifying LEH defect depth
have been proposed16,17, but there is yet to be a methodological consensus for data collection and analysis, nor
a standard approach for the identication of defects on the basis of their quantitative morphology. Most micro-
scopic analyses of LEH focus on perikymata spacing within defects, but more research is needed to understand
and quantify ‘normal’ perikymata variation in dierent crown regions and tooth types, both within and among
species, in order for deviations from normality to be used as an indicator of stress events.
In addition to the replicability of defect identication, measuring LEH defect depth also sheds light on growth
and thus life history patterns. At the population level, defect depth has been shown to reect species-specic
enamel growth (i.e., extension) rates in great apes18. Faster-growing mountain gorillas have signicantly shallower
defects, relating to their shallower growth increments below the enamel surface, and mirroring their faster life
history patterns overall19. Previous work hypothesized that the relationship between enamel growth rates and
defect depth exists among hominins as well20, with dierences in depth, and thus perceptibility via qualitative
observation, inuencing estimates of LEH prevalence. However, no defect depths are published for Homo sapiens,
Figure1. Canine epoxy replicas included in this study. Le Moustier 1 lower right canine (LRC), Les Rois 2B
LRC, Çatalhöyük 1938.1 LRC, and Saxon burial at Hildesheim Befund 1701 lower le canine. Digital elevation
models generated within SensoSCAN soware (S Neox, https ://www.senso far.com/metro logy/produ cts/sneox
/) are superimposed on the replicas, with blue areas showing surface depressions. Defects are marked with
arrows and perikymata are visible as small grooves on the surface. e scale is 3mm and corresponds to the full
replicas.
Content courtesy of Springer Nature, terms of use apply. Rights reserved


Vol.:(0123456789)
Scientic Reports | (2021) 11:522 | 
www.nature.com/scientificreports/
and only one study currently exists on fossil hominins comparing Australopithecus africanus and H. naledi21.
erefore, very little is known about how defect depth might relate to enamel growth and life history patterns
in our own species and extinct relatives.
Defect depth also provides information on the severity of growth disruptions themselves, and is best under-
stood by analyzing the cellular activity around the time of the disruption. Evidence from experimental research
on sheep incisors suggests that the intensity of stressors inuences defect dimensions, with higher doses of
growth-disrupting uoride creating broader and deeper defects22. Further evidence from analyses of deer teeth
suggest that more severe defects occur in animals exposed to high, intermittent doses of uoride, particularly
in cases where enamel secretion was abruptly halted altogether, with no resumption of cellular activity23, form-
ing the equivalent of plane-form hypoplastic defects in primates24. In contrast, histological studies of modern
human teeth suggest that the classic LEH defect type, also called furrow-form defects, reect less severe disrup-
tions to late-stage enamel secretion than plane form defects24. However, it is oen dicult to discern plane- vs.
furrow-form defects without histology, the former of which represents a more severe growth disruption from the
cellular perspective but of a shorter duration, while the latter oen represents a longer-forming but less severe
growth disruption24–26.
Researchers have long hypothesized that defect depth provides information about the severity of the stressor
that disrupted growth in hominoids16,21,27, just as location and width of the defects tell us something about their
timing and duration, respectively. Evidence from great ape studies supports this, with several wild-captured indi-
viduals exhibiting particularly deep defects that might reect major stress events17. One known gorilla exhibits
a defect depth of 276µm (compared to the species median of 41µm), and according to associated veterinary
records, this defect likely formed around the time that she was captured from the wild to live in a zoo17. At the
population level, mountain gorillas that developed their teeth during a period of intense human encroachment
have defects that are almost twice as deep as those that lived under increased protection17. Further, anged
orangutans, which exhibit higher levels of the stress hormone cortisol throughout development, have defects
that are more than twice as deep as unanged or developmentally arrested males with either the same or lower
cortisol levels28. e duration of the disruption is also likely to aect classic furrow-form defect dimensions,
as is usually approximated by counting the number of perikymata involved in the defect, with each additional
perikyma carving more deeply into the enamel wall20,21. At the cellular level, the number of aected enamel-
secreting cells, as well as the relative timing of the growth disruption in relation to the life of the cells, aects
the depth of the defect at the outer enamel surface24,25. Taken together, LEH defect depth is likely to reect the
interaction of multiple factors, including the intensity of the stressor, the duration of the growth disruption, and
the developmental timing of the disruption, in addition to interspecic growth variation. However, a growing
body of research supports the link between greater stress, as can be dened in terms of intensity and/or duration,
and deeper LEH defects, particularly when an eort is made to account for enamel growth variation.
Here, we describe a method to create high-resolution 3D models of the enamel surface using confocal pro-
lometry, thus allowing for the identication and precise measurement of perikymata and LEH defect depths, and
apply it to a sample of 17 Neanderthal and 18 H. sapiens anterior teeth. Perikymata depth is yet to be quantied
in any hominin species despite its potential for providing baseline information about normal growth patterns
of the underlying tissue. We compare perikymata and defect depths to published enamel growth information,
like enamel extension rates, which vary among tooth types and taxa. Previous work suggests that Neanderthals
have enamel extension rates that are an average of 33% higher than European H. sapiens in their anterior teeth
(28–35%)29,30, providing an opportunity to assess defect depth and its relationship to enamel growth patterns
in these groups. We scale defect depth by perikymata depth as a means to control for the ‘normal’ interspecic
and inter-tooth variation in surface morphology, as this allows for direct comparisons of defect severity, or the
relative proportion of enamel that is missing within defects between Neanderthals and H. sapiens, as well as
among the H. sapiens samples. In this way, we are able to test new hypotheses about the severity (i.e., intensity
and/or duration) of stress experiences in past populations on the basis of LEH defect depth while accounting
for species dierences in enamel growth.
Specically, we aim to:
1. Test whether lines or grooves identied as LEH defects by visual inspection of the enamel surface truly rep-
resent localized reductions in enamel thickness by comparing their depth to perikymata depth, and assess
the relationship between the two variables;
2. Test whether perikymata and defect depths mirror established growth dierences between tooth types
and taxa, i.e., whether faster-growing, thinner-enameled incisors have shallower defects than canines, and
whether the anterior teeth of Neanderthals have shallower depths compared to Upper Paleolithic, Neolithic,
and medieval H. sapiens;
3. Use ratios of defect to perikymata depth (i.e., LEH severity ratios) to compare relative stress severity between
taxa and among H. sapiens samples. We compare the magnitude of the reductions in enamel thickness (i.e.,
LEH defect depth) while controlling for species-specic variation in perikymata depth, allowing for infer-
ences about the intensity and/or duration of stress episodes to be made.
Results
We compared defect depths to perikymata depths to ensure that defects identied by eye represent LEH, or local-
ized reductions in enamel thickness, and indeed there is no overlap in their distributions within teeth (N = 280
perikymata and 71 defects) (Table1; TableS1). ere is a signicant positive correlation between perikymata
and defect depths in the combined sample (R2 = 0.39; p < 0.001), but a large proportion of variation in defect
depth remains unexplained by perikymata depth (Fig.S1). Within individual dentitions, incisors have shallower
Content courtesy of Springer Nature, terms of use apply. Rights reserved


Vol:.(1234567890)
Scientic Reports | (2021) 11:522 | 
www.nature.com/scientificreports/
perikymata (F(1,265) = 143.7, p < 0.001) and defect depths (F(1,52) = 21.5, p < 0.001) compared to canines, includ-
ing instances of matched defects across the same dentitions (TableS2).
e Neanderthal sample has signicantly shallower perikymata than the H. sapiens sample (F(1,12) = 27.8,
p < 0.001) (Fig.2). e H. sapiens sample has perikymata that are an average of 2.33 times (range 1.69–3.65)
deeper than the Neanderthal sample in incisors, and 3.24 times (range 3.05–3.65) deeper in canines. Perikymata
depths range from shallowest in the La Chaise Neanderthal upper canines and incisors to the deepest in the
Çatalhöyük lower canines and incisors.
e Neanderthal sample also has signicantly shallower LEH defects than H. sapiens (F(1,16) = 30.0, p < 0.001)
(Fig.2). Defect depths follow a similar pattern with La Chaise Neanderthal upper incisors having the shallowest
defects and H. sapiens lower canines from the Saxon burials at Hildesheim and Çatalhöyük having the deepest
defects. However, there are no signicant dierences in LEH severity ratios between Neanderthals and H. sapiens,
as calculated by dividing each defect by the median perikymata depth for each tooth (F(1,16) = 0.30, p = 0.598).
When further subdivided by sample, the Neolithic Çatalhöyük specimens exhibit signicantly lower severity
ratios compared to the other H. sapiens samples from the medieval and Upper Paleolithic periods, and compared
to the Neanderthals (F(3,14) = 15.7, p < 0.001; posthoc comparisons—Neolithic vs. medieval p = 0.004, Neolithic
vs. Upper Paleolithic and Neanderthals p < 0.001) (Table1, Fig.3).
To test the effect of leveling algorithms that have been used in other studies to remove the effect of
curvature15,20, and ultimately move toward a methodological consensus for measuring defect depth, plane
and sphere-type form removal algorithms (within SensoSCAN) were applied to two digital elevation mod-
els (DEMs) from Le Moustier canines (Fig.S2). Leveled values underestimate the depth of defects, being on
average 24.3% shallower than raw depths (range = 16.1–27.9%; N = 4), and with a mean absolute dierence of
9.9µm (range = 6.9–11.1µm; N = 4). When the same algorithms are applied to perikymata, they have the oppo-
site eect, making perikymata appear articially deeper. e mean dierence in perikymata depth is 36.1%
(range = 14.3–64.2%; N = 4), with a mean absolute dierence of 1.0µm (range = 0.40–1.63µm; N = 4).
Discussion
A long-standing debate surrounds the appearance of modern human-like development and life history. Evidence
from studies of brain, skeletal, and dental growth is equivocal, with some researchers arguing that Neanderthals
exhibited modern human-like growth patterns, while others suggest that they had faster growth rates, particularly
during early development5,6,29–31. Our data support the hypothesis that Neanderthals had faster-growing anterior
teeth29, as evidenced by their shallower perikymata and LEH defects. e extent to which faster anterior tooth
growth rates relate to other aspects of life history is yet to be fully understood, but recent evidence suggests that
the tempo and mode of brain ontogeny, and possibly skeletal growth, diers between the two species6. Among
extant apes, faster-growing species like mountain gorillas have shallower LEH defects, faster enamel growth
(extension) rates, accelerated brain and somatic growth, and faster life history schedules17–19,32,33. However,
more comparative data from extant and fossil hominoids are needed in order to further assess the relationships
among these variables in hominins.
e application of novel methods such as ours unlocks new information about growth and developmental
stress in past populations, improving our understanding of dental development and life history evolution. We
provide a way to reliably identify LEH defects on the basis of their being deeper than nearby ‘normal’ growth
increments, facilitating cross-study comparisons. When analyzing samples with known dierences in their
enamel growth patterns, our results show that it is important to take that variation into account when interpreting
defect depths. ere is a positive relationship between perikymata and defect depth, with deeper defects occur-
ring in teeth with deeper perikymata (Fig.S1). Defect and perikymata depths also follow the predicted pattern
based on documented dierences in enamel growth rates between dierent tooth positions, including when
comparing matched defects across the same dentitions (TableS2). Shallower features occur in teeth with faster
average enamel extension rates and thinner enamel18, such as in incisors vs. canines29,34. is inter-tooth dier-
ence is understood to be a consequence of growth-related variation in underlying enamel geometry, namely the
angles that striae of Retzius make as they meet the outer enamel surface18. Faster enamel extension rates are also
associated with less tightly packed perikymata on the tooth surface35, as exhibited by Neanderthals compared to
modern humans, particularly in the cervical crown36. Neanderthals also have lower periodicities, or the number
Table 1. Perikymata and LEH defect depths by taxon.
Tax on Tooth type N perikymata (pk) Median and range pk depth
(µm) N defects Median and range defect
depth (µm) LEH severity ratio (defect/
pk depth)
H. neanderthalensis
(Sites: Le Moustier, La Chaise,
La Chaise Suard, Biache-Saint-
Vaast, Kulna, Monsempron,
Rochelot)
LC
LI1
LI2
UC
UI1
UI2
10
10
10
50
20
30
1.57 (1.17–3.09)
0.83 (0.71–1.18)
0.84 (0.46–1.78)
1.04 (0.42–3.05)
1.28 (0.49–3.05)
0.81 (0.53–1.43)
5
3
2
13
5
8
26.4 (19.6–46.2)
15.2 (9.8–27.5)
19.0 (10.7–27.3)
24.1 (13.3–45.7)
13.4 (9.5–22.7)
17.2 (10.1–31.3)
16.8 (12.5–29.4)
18.3 (11.8–33.1)
22.6 (12.7–32.5)
27.1 (9.9–49.2)
13.4 (8.7–15.6)
20.7 (13.8–35.6)
H. sapiens
(Sites: Les Rois, Saint Germain-
la-Rivière, Çatalhöyük, Saxon
burials at Hildesheim)
LC
LI1
LI2
UC
UI1
UI2
40
30
30
20
20
10
4.32 (1.28–11.29)
2.42 (0.96–7.95)
2.52 (0.83–8.17)
4.18 (1.68–8.47)
1.79 (1.14–3.09)
2.04 (1.55–3.42)
11
8
6
3
6
1
47.5 (32.8–101.7)
38.4 (20.6–48.4)
36.7 (20.9–50.5)
41.7 (32.7–79.1)
28.4 (18.7–52.4)
29.6
16.1 (5.7–40.0)
19.7 (4.2–46.5)
19.6 (5.6–52.3)
7.8 (6.1–28.3)
16.3 (9.4–30.5)
14.5
Content courtesy of Springer Nature, terms of use apply. Rights reserved


Vol.:(0123456789)
Scientic Reports | (2021) 11:522 | 
www.nature.com/scientificreports/
of days represented by striae of Retzius, providing further evidence of faster growth where perikymata are more
widely spaced compared to H. sapiens29. e Neanderthal sample in this study has two to three times shallower
perikymata in their incisors and canines compared to the H. sapiens sample, following the same pattern (but at
a higher magnitude) as the enamel extension rate dierences documented in previous studies29,30.
When compared with published defect depths from nonhuman great apes and hominins, the H. sapiens
sample has the highest median depths (47.5µm in mandibular canines), perhaps as a result of their slower
average extension rates, larger striae of Retzius angles, and thicker enamel compared to apes18. e Neanderthal
specimens, in contrast, have a median defect depth (30.3µm) that is more similar to earlier hominins Australo-
pithecus africanus (26.0µm) and H. naledi (26.9µm) as reported by Skinner21, as well as nonhuman apes (23.6µm
in mountain gorillas). However, the A. africanus and H. naledi depths21 were collected from leveled DEMs and
therefore likely underestimate the true depth, based on our comparisons of leveled vs. raw data (Fig.S2). In terms
of dierences among the H. sapiens samples, relatively little is known about enamel growth variation within our
species. A histological analysis of contemporary European and South Africans37, and the variation in the timing
of dental eruption38, suggest that considerable variation can be expected.
Figure2. Perikymata and linear hypoplastic defect depths by tooth type. Neanderthals have signicantly
shallower perikymata (n = 280; 150 H. sapiens and 130 Neanderthal) and defects (n = 71; 35 H. sapiens and 36
Neanderthal) than H. sapiens. Inter- and intra-individual comparisons suggest that faster-growing and thinner-
enameled teeth (i.e., Neanderthals compared to H. sapiens; incisors compared to canines) have shallower
perikymata and defects at the population level. LLI1/ULI1 = lower/upper central incisor; LLI2/ULI2 = lower/
upper lateral incisor; LC = lower/upper canine. Figure generated in RStudio (version 1.3.959)50.
Content courtesy of Springer Nature, terms of use apply. Rights reserved


Vol:.(1234567890)
Scientic Reports | (2021) 11:522 | 
www.nature.com/scientificreports/
If enamel growth variation explains only a moderate proportion of population-level defect depth, we hypoth-
esize that severity—the intensity and/or duration of the stressor—helps to explain the remaining variation21,39. We
propose that using a ratio of defect to perikymata depth provides a way to tease apart the inuence of growth vari-
ation (whether tooth- or species-specic) vs. stress severity in defect formation. e quantity of missing enamel
tissue within each defect is measured in relation to the local perikymata depth, allowing for comparisons of LEH
severity ratios among samples with diering growth patterns. e interplay of severity and duration can further
shape the appearance of enamel defects, particularly those of the classic furrow-form type. Guatelli-Steinberg
etal.39 found that Neanderthal defects represent shorter growth disruptions compared to those in Inuit forag-
ers, and defects of a shorter duration are likely to be shallower than those of a longer duration21. Our method
is in line with the existing theoretical framework surrounding LEH defect formation in which the intensity of
the stressor (i.e., the number of aected ameloblasts), the duration of the stress episode, and the timing of the
stressor in relation to cellular developmental stage inuence defect depth15,21,23–26.
In addition to the life history debate, researchers have long supposed that Neanderthals experienced more
severe early life stress compared to Upper Paleolithic humans in Europe40. Ogilvie etal.40 analyzed a large sample
of 669 Neanderthal teeth and suggested that they had high LEH prevalence compared to recent human popula-
tions, which they interpreted as a sign of lower foraging eectiveness and higher food stress from the time of
weaning through adolescence. Our results show that while Neanderthals have defects that are absolutely shallow,
their LEH severity ratios are not signicantly dierent from those in the H. sapiens sample as a whole.
When the H. sapiens sample is divided by time period, the Neolithic specimens from Çatalhöyük stand out
as having signicantly less severe LEH (i.e., lower severity ratios) compared to Upper Paleolithic and medieval
H. sapiens as well as Neanderthals (Table1, Fig.3). Çatalhöyük, a well-documented early farming settlement in
south-central Anatolia (Turkey; 7100–5150cal BCE), shows little evidence of early life stress beyond the mere
presence of LEH defects, which are mostly short in duration41. Skeletal evidence suggests that the ontogenetic
patterns at Çatalhöyük match those of well-nourished contemporary populations41. e Çatalhöyük specimens
included in this study have some of the deepest defects in the sample, but they also have the deepest perikymata,
leading Çatalhöyük to have signicantly lower LEH severity ratios than the other groups. We hypothesize that
deeper perikymata are at least partially explained by slower enamel growth rates, which might also explain their
deeper defects in the absence of more severe stress. Expanded analyses incorporating dental histology could test
whether enamel growth rates are indeed slower in Neolithic vs. foraging or more recent agricultural populations,
tting with bioarchaeological evidence of reduced energetic investment in maintenance and growth during the
transition to agriculture, coinciding with dramatic reductions in stature in many parts of the world42.
e reduction in LEH severity ratios is not merely a consequence of adaptive changes through time as the
medieval Saxon burials sample has higher ratios than the Neolithic sample. e Saxon burials are derived from
a time period of active warfare and increasing population density (c. 700–1030 CE) in modern day Hildesheim,
Germany. e Upper Paleolithic individuals have LEH severity ratios that are the highest in the study, but they
Figure3. Severity ratios by sample. Ratios are calculated by dividing defect depths by perikymata depths
from the same teeth (N = 59), or median perikymata depth for the species and tooth type where perikymata
were not well-preserved enough to reliably measure (N = 12). e Neolithic sample from Çatalhöyük (Neo)
has signicantly lower severity ratios compared to the Medieval (Med), Upper Paleolithic (UP), and Middle
Paleolithic (MP) samples, as noted with asterisks. Figure generated in RStudio (version 1.3.959)50.
Content courtesy of Springer Nature, terms of use apply. Rights reserved


Vol.:(0123456789)
Scientic Reports | (2021) 11:522 | 
www.nature.com/scientificreports/
are not signicantly dierent from the Neanderthals nor the medieval Saxons. While they relied on a lot of the
same food resources as Neanderthals, evidence suggests that Upper Paleolithic humans had a more varied and
exible diet, which could have led to lower mortality and higher fertility43. Snodgrass and Leonard43 hypothesized
that Neanderthals would have experienced seasonal periods of “intense energy stress,” and indeed, Smith etal.44
found a pronounced week-long internal growth disruption in a Neanderthal tooth that is consistent with illness
and associated weight loss during the coldest part of winter. However, little concrete evidence currently exists to
help explain the severe LEH in Upper Paleolithic humans.
We suggest that defect depth is the most biologically appropriate signal to measure when attempting to iden-
tify and characterize LEH morphology as defects are dened as reductions in enamel thickness. e dierence in
scale between defect depth vs. changes in perikymata spacing associated with defects makes the former easier to
identify, characterize, and verify through repeated and independent measurements. In contrast to spacing-only
analyses, which require near-perfect surface preservation, defect depth can be analyzed in any samples with
preserved enamel (i.e., present and not covered in calculus or plant matter). Minor variation in perikymata spac-
ing, or accentuated perikymata, could reect minor growth disruptions, corresponding to the more numerous
accentuated lines oen visible in thin section. If so, detailed perikymata analyses using the method described
here provide a way to quantify those defects that fall in the grey area between clear growth disruptions and ‘nor-
mal’ growth increments. Additionally, a better understanding of perikymata packing patterns in both 2- and 3D
would contribute to species attributions and assessments of hybridization, as in the case of Les Rois mandible
B (included in this study), which displays a mixed morphology of Neanderthal and H. sapiens-like traits45. By
analyzing changes in defect and perikymata morphology through time, it might be possible to determine at what
point ‘contemporary’ developmental patterns appeared, if such unique patterns exist. Our preliminary data sug-
gest that recent H. sapiens have evolved highly variable perikymata morphology with a wide range of perikymata
depths, mirroring the high variation found in other developmental variables like dental eruption times38.
At the level of the individual defect, there is evidence to suggest that outliers, or particularly deep defects,
reect more severe stress episodes, i.e., intense in terms of the cellular reaction, physiological stressor, and/or
longer-lasting in the case of furrow-form defects17,21–23. Only once more data are gathered from extant samples
with associated records will models be able to accommodate dierential stress experiences as an explanatory vari-
able, and even then, enamel defects will continue to reect nonspecic stressors in terms of etiology. However, by
analyzing defect depth in relation to perikymata depth, information about the magnitude of growth disruptions
among populations, as operationalized using the LEH severity ratio, can be gleaned in the absence of associated
data. Further, if the relationship between defect depth and enamel extension rates continues to be supported
in more taxa, population-level perikymata and defect depth may be used to model aspects of enamel growth in
samples that cannot be sectioned or virtually imaged via synchrotron or other methods. Indeed, defects appear on
a continuum, and there exists an active debate about to what extent LEH defects reect ‘normal’ growth processes
as opposed to pathologies46. is method will ultimately contribute to discussions around interindividual suscep-
tibility to defect formation, and provide more information about the early life stress experiences of individuals.
LEH analyses are a standard part of bioarchaeological and paleoanthropological analyses, and it is critical that
they strive to take morphological and growth variation into consideration and form interpretations accordingly.
Materials and methods
Sample. e sample includes 35 high resolution epoxy replicas created from permanent anterior teeth (man-
dibular and maxillary incisors and canines) from Homo neanderthalensis (N = 17 teeth; sites: Le Moustier, La
Chaise, La Chaise Suard, Biache-Saint-Vaast, Kulna, Monsempron, and Rochelot) and Homo sapiens (N = 18
teeth; sites: Les Rois, Saint Germain-la-Rivière, Çatalhöyük, Saxon burials at Hildesheim) (TableS1). Anterior
teeth were selected because relatively more of their crowns are made up of imbricational striae, where periky-
mata and defects are visible on the surface. Specimens were selected for analysis based on their surface preserva-
tion, i.e., whether they had visible perikymata and little to moderate wear, calculus, and other debris obscuring
the outer enamel surface. Only the best preserved antimere, right or le, was selected for analysis to avoid
repeated measurements of the same defects10,12. Perikymata and LEH defects were measured within the mid-
crown dened here as the middle 3/5ths of crown height, avoiding the cuspal and cervical regions most aected
by wear and calculus, and reducing the potential inuence of changes in underlying geometry on surface feature
depth15. Defect depth was measured as the maximum dierence between the occlusal shoulder and the deepest
point within the groove (Fig.4). Perikymata depth was measured as the maximum dierence between the level
of a perikyma groove and a perikyma ridge (Fig.5).
Impressions were collected from original teeth using Coltène’s President Jet Regular and/or Light Body
dental impression material (Coltène-Whaledent). High-resolution positive replicas were made using Loctite
Hysol E-60NC and/or EPO-TEK 301 epoxy. No coatings or treatments were applied to the replicas for imaging.
At least two teeth with partially overlapping developmental periods were analyzed per specimen, except in the
case of isolated teeth, to conrm systemic defect manifestation46.
Equipment and settings. Individual defects were rst identied via surface visualization (by eye and/or
a hand lens) and then imaged using the Sensofar S Neox confocal prolometer. e replica was positioned on
the microscope stage using a tilting stage that eases orientation of the defect perpendicular to the light source
(Fig.4A). Once positioned, a quick scan was performed to assess whether the features of interest (i.e., the occlusal
shoulders of the target defect or the region with retained perikymata) were level (Fig.4B). e Z-plane step size
for each scan was set to 1µm with a Z-range encompassing the full height of the defect and the occlusal shoul-
ders, ranging between 75 and 250µm. e curvature of the crown negatively aects the quality of individual
scans and the ability of the researcher to track which defect or perikyma is being measured in later stages of
Content courtesy of Springer Nature, terms of use apply. Rights reserved


Vol:.(1234567890)
Scientic Reports | (2021) 11:522 | 
www.nature.com/scientificreports/
analysis. erefore, the entire crown, or only crown regions of interest in cases where there is incomplete surface
preservation, was imaged in strips of crown height of less than 10mm each, with 10 percent overlap between
image frames (Fig.4C). e eld of view for each scan can be adjusted based on the shape of the tooth. However,
a width of at least 500µm and a length that fully encompasses both the region before and aer the defect was
used here (e.g., 656 × 3241µm in Fig.4B,C).
e ultimate resolution of the Sensofar microscopes provides the appropriate vertical resolution (0.64µm/
pixel; 0.31µm optical resolution using the 20x lens) to capture the morphology of defects and perikymata as
higher resolution images would be too large to analyze on most systems, while lower resolution images would
not allow especially perikymata to be reliably measured. e type of light used to probe the surface should be
selected based on the material properties of the epoxy replica, but here, white light (as opposed to blue, green, or
red) produces the best quality scans. Within the proprietary soware SensoSCAN, the confocal fusion algorithm
Figure4. A step-by-step guide to the method. (A) Le Moustier 1 LRC epoxy replica with the cusp toward the
right of the image (scale = 3mm). An LEH defect is marked by an arrow. (B) An overview image within the
SensoSCAN soware (S Neox, https ://www.senso far.com/metro logy/produ cts/sneox /) with a bounding box
marking the crown region to be scanned. (C) e resulting digital elevation model (DEM), showing the defect
as a reduction in enamel thickness. (D) e DEM is read into ImageJ soware47,48 where a transect is drawn
from one occlusal shoulder to the other to extract 2D coordinates for analysis. (E) Once plotted, the FindPeaks
plug-in49 is used to measure maximum defect depth from the occlusal side of the defect (in µm).
Content courtesy of Springer Nature, terms of use apply. Rights reserved


Vol.:(0123456789)
Scientic Reports | (2021) 11:522 | 
www.nature.com/scientificreports/
(under expert options) as opposed to continuous provides the highest percentage of measured points (98–99.9%).
We provide imaging settings for the Sensofar S Neox prolometer, but other noncontact prolometry systems
are also available and provide a similar level of vertical resolution.
Image processing and measurement protocol. e resulting x, y, and z coordinates were saved in .dat
le format and read into the Fiji distribution of ImageJ47 using the XYZ2DEM plug-in48. is plugin imports the
coordinates and interpolates a digital elevation model or DEM using a Delaunay triangulation (Fig.4D). Scale
information (i.e., pixel length, scan area) is available within the SensoSCAN soware and can be entered when
loading the DEM, and by using the ‘set scale’ function in the ‘analyze’ menu. Once opened, the DEM appears
in grey-scale format (Fig.4E), although shading or shadows can be applied to the DEM to increase contrast,
if desired. As the orientation of the DEMs are ipped when they read into ImageJ, we used the transform tab
of the menu to ip them vertically before analysis. e line tool was used to draw a transect orthogonal to the
feature(s) of interest, across the occlusal shoulders of defects (Fig.4D) or across several perikymata (Fig.4B).
Due to curvature in both the mesiodistal and cuspo-cervical directions, each transect was drawn across the por-
tion of the DEM that was orthogonal with the microscope during scanning, i.e., level on either side of the defect
or perikyma being measured, which is usually in the midline of the DEM.
Figure5. Perikymata measurements in Les Rois 2B lower right canine. (A) Digital elevation model (DEM) of
midcrown perikymata. (B) e same DEM loaded into ImageJ soware47 using XYZ2DEM plug-in48, with a
transect drawn across a level region of normal (i.e., without clear defects) perikymata. (C) Extracted 2D prole
using the FindPeaks plug-in49, allowing for the measurement of maximum perikymata depth in four separate
furrows (in µm).
Content courtesy of Springer Nature, terms of use apply. Rights reserved


Vol:.(1234567890)
Scientic Reports | (2021) 11:522 | 
www.nature.com/scientificreports/
Once the transect was drawn and a 2D prole plotted, the FindPeaks plug-in was used to locate the Z-coor-
dinate of the peak associated with the occlusal edge of the defect or perikyma ridge, as well as the coordinate for
the deepest point within the defect oor or perikyma groove (Figs.4E, 5C; package can be found within BAR, a
collection of Broadly Applicable Routines)49. Depending on the length of the transect and the depth of features
being targeted, the threshold for the FindPeaks parameters might need modication: for LEH defects, the default
settings are usually sucient (i.e., defects are substantially deeper than any other nearby features like perikymata,
and are therefore automatically identied by the soware), but particularly when measuring shallow features,
the minimum peak amplitude might need to be reduced (e.g., to a value above median local perikymata depth
to measure shallow defects, or to a value below median perikymata depth to measure perikymata). e longer
the transect drawn within the DEM, the more likely that the default amplitude will need adjustment to identify
the grooves. We measured maximum defect depth three times to calculate an average depth for each defect
using three separate transects. To measure perikymata depth, the process was the same, except a transect was
drawn across just across a few perikymata that occupy a relatively at area of the crown without obvious defects
(Fig.5B). Given the smaller height of perikymata, and their relatively greater potential for wear over the lifespan,
it was particularly important to measure perikymata in unworn to lightly worn teeth. In this study, 10 separate
perikymata were measured within dierent areas of the midcrown in each well-preserved tooth, as evidenced
by Tomes’ pit processes within the perikyma grooves46.
An alternative approach to collecting defect depths has been proposed16,21 in which algorithms are used to
eliminate the eects of tooth curvature before extracting 2D measurements from 3D scans. In order to assess the
eect of leveling, we used plane and sphere ‘form-removal’ leveling algorithms within SensoSCAN soware and
compare defect and perikymata depths obtained in raw vs. leveled DEMs. In this study it was not necessary to
lter noise before analysis; 2D transects are used to measure the features of interest, making it feasible to avoid
the typical aberrations (e.g., bubbles, large scratches, spikes, etc.) when drawing transects. However, if spikes
are visible in the DEM and/or resulting 2D proles, the ‘reduce noise’ function can be used to eliminate outliers
before collecting depth data see17 for specications.
Replicability. To assess the replicability of the imaging and analysis process, we measured the depth of
matched defects as well as perikymata depth within the same regions of the midcrown in the same replicas using
two models of confocal microscopes hosted at two institutions: the Sensofar S Neox (Université de Bordeaux)
and PLu Neox (Universität Tübingen). e mean dierence in LEH defect depth between the two microscopes is
2.3% (range = 0.8–3.5%; N = 5), with a mean absolute dierence of 1.1µm (range = 0.3–1.8µm; N = 5). e mean
dierence in perikymata depth in the same replicas between the two microscopes is 1.7% (range = 1.1–2.4%;
N = 3), with a mean absolute dierence of 0.3µm (range = 0.02–0.77; N = 3).
Statistical analyses. Assumptions of normality and homogeneity of variance were visually assessed using
residual diagnostic plots. Perikymata and defect depth values were natural log-transformed prior to analyses
because their distributions are right-skewed. We conducted a pairwise correlation analysis using the Pearson
method to assess the relationship between perikymata and defect depth for all teeth in which perikymata were
well-preserved (n = 29) using the median depths for each tooth. We used separate linear mixed models to test
whether there are dierences in perikymata and LEH defect depths between the two taxa and between incisors
and canines, as well as LEH severity ratios between taxa and among temporal groups (i.e., Neanderthals, Upper
Paleolithic, Neolithic, and Medieval). For 12 out of 71 defects, median perikymata depth for the species and
tooth type, rather than that exact tooth, was used as perikymata were not well-preserved enough to reliably
measure. We included specimen ID as a random eect in all models as multiple defects and perikymata were
measured per specimen, and some of the defects are ‘matched’ meaning that they represent the same systemic
growth disruption, but on dierent teeth. Analyses were performed in RStudio (Version 1.3.959)50 using package
nlme for the mixed models.
Data availability
All data analyzed in this study are included in the article and its Supplementary Information les.
Received: 28 July 2020; Accepted: 16 December 2020
References
1. Smith, B. H., Crummett, T. L. & Brandt, K. L. Ages of eruption of primate teeth: a compendium for aging individuals and compar-
ing life histories. Am. J. Phys. Anthropol. 37, 177–231 (1994).
2. Kelley, J., Schwartz, G. T. & Smith, T. M. Age at rst molar emergence in Pan troglodytes verus and variation in the timing of molar
emergence among free-living chimpanzees. J. Hum. Evol. 145, 102823 (2020).
3. Smith, T. M. et al. First molar eruption, weaning, and life history in living wild chimpanzees. Proc. Natl. Acad. Sci. U. S. A. 110,
2787–2791 (2013).
4. Dean, M. C. Growing up slowly 160,000 years ago. Proc. Natl. Acad. Sci. U. S. A. 104, 6093–6094 (2007).
5. Zollikofer, C. P. & de León, M. S. e evolution of hominin ontogenies. Semin. Cell. Dev. Biol. 21, 441–452 (2010).
6. Hublin, J. J., Neubauer, S. & Gunz, P. Brain ontogeny and life history in Pleistocene hominins. Philos. Trans. R. Soc. B 370, 20140062.
https ://doi.org/10.1098/rstb.2014.0062 (2015).
7. Smith, T. M. Incremental dental development: methods and applications in hominoid evolutionary studies. J. Hum. Evol. 54,
205–224 (2008).
8. Le Cabec, A., Tang, N. & Taoreau, P. Accessing developmental information of fossil hominin teeth using new synchrotron
microtomography-based visualization techniques of dental surfaces and interfaces. PLoS ONE https ://doi.org/10.1371/journ
al.pone.01230 19 (2015).
Content courtesy of Springer Nature, terms of use apply. Rights reserved


Vol.:(0123456789)
Scientic Reports | (2021) 11:522 | 
www.nature.com/scientificreports/
9. Bromage, T. G., Perez-Ochoa, A. & Boyde, A. Portable confocal microscope reveals fossil hominid microstructure. Microscop.
Anal. 19, 5–7 (2005).
10. Guatelli-Steinberg, D. et al. Chronology of linear enamel hypoplasia formation in the Krapina Neanderthals. PaleoAnthropol. 2014,
431–445 (2014).
11. Hassett, B. R. Evaluating sources of variation in the identication of linear defects of enamel: a new quantied method. J. Archaeol.
Sci. 39, 560–565 (2012).
12. Bocaege, E. & Hillson, S. Disturbances and noise: dening furrow-form enamel hypoplasia. Am. J. Phys. Anthropol. 161, 744–751
(2016).
13. Temple, D. H., Nakatsukasa, M. & McGroarty, J. N. Reconstructing patterns of systemic stress in a Jomon period subadult using
incremental microstructures of enamel. J. Archaeol. Sci. 39, 1634–1641 (2012).
14. Go odman, A. H. & Rose, J. C. Assessment of systemic physiological perturbations from dental enamel hypoplasias and associated
histological structures. Yearb. Phys. Anthropol. 33, 59–110 (1990).
15. Hillson, S. & Bond, S. e relationship of enamel hypoplasia to the pattern of tooth crown growth: a discussion. Am. J. Phys.
Anthropol. 104, 89–103 (1997).
16. Skinner, M. F. & Skinner, M. M. Orangutans, enamel defects, and developmental health: a comparison of Borneo and Sumatra.
Am. J. Phys. Anthropol. 96, 301–314 (2017).
17. McGrath, K. et al. Quantifying linear enamel hypoplasia in Virunga Mountain gorillas and other great apes. Am. J. Phys. Anthropol.
166, 337–352 (2018).
18. McGrath, K. et al. Faster growth corresponds with shallower linear hypoplastic defects in great ape canines. J. Hum. Evol. 137,
102691. https ://doi.org/10.1016/j.jhevo l.2019.10269 1 (2019).
19. Stoinski, T. S., Perdue, B., Breuer, T. & Ho, M. P. Variability in the developmental life history of the genus Gorilla. Am. J. Phys.
Anthropol. 152, 165–172 (2013).
20. Guatelli-Steinberg, D. Analysis and signicance of linear enamel hypoplasia in Plio–Pleistocene hominins. Am. J. Phys. Anthropol.
123, 199–215 (2004).
21. Skinner, M. F. Developmental stress in South African hominins: Comparison of recurrent enamel hypoplasias in Australopithecus
africanus and Homo naledi. S. Afr. J. Sci. 115, 1–10 (2019).
22. Suckling, G. & urley, D. C. Histological, macroscopic and microhardness observations of uoride-induced changes in the enamel
organ and enamel of sheep incisor teeth. Arch. Oral Biol. 29, 165–255 (1984).
23. Kierdorf, H. & Kierdorf, U. Disturbances of the secretory stage of amelogenesis in uorosed deer teeth: a scanning electron-
microscopic study. Cell Tissue Res. 289, 125–135 (1997).
24. Witzel, C., Kierdorf, U., Schultz, M. & Kierdorf, H. Insights from the inside: Histological analysis of abnormal enamel microstruc-
ture associated with hypoplastic enamel defects in human teeth. Am. J. Phys. Anthropol. 136, 400–414 (2008).
25. Witzel, C. et al. Reconstructing impairment of secretory ameloblast function in porcine teeth by analysis of morphological altera-
tions in dental enamel. J. Anat. 209, 93–110 (2006).
26. Kierdorf, H., Witzel, C., Upex, B., Dobney, K. & Kierdorf, U. Enamel hypoplasia in molars of sheep and goats, and its relationship
to the pattern of tooth crown growth. J. Anat. 220, 484–495 (2012).
27. Moggi-Cecchi, J., Pacciani, E. & Pinto-Cisternas, J. Enamel hypoplasia and age at weaning in 19th-Century Florence, Italy. Am.
J. Phys. Anthropol. 93, 299–306 (1994).
28. Kralick, A. E. & McGrath, K. Orangutan canine linear enamel hypoplasia defects assessed in association with anging status. Am.
J. Phys. Anthropol. 171(S69), 149 (2020).
29. Smith, T. M. et al. Dental evidence for ontogenetic dierences between modern humans and Neanderthals. Proc. Natl. Acad. Sci.
U. S. A. 107, 20923–20928 (2010).
30. Smith, T. M., Toussaint, M., Reid, D. J., Olejniczak, A. J. & Hublin, J.-J. Rapid dental development in a middle Paleolithic Belgian
Neanderthal. Proc. Natl. Acad. Sci. U. S. A. 104, 20220–20225 (2007).
31. Rozzi, F. V. & De Castro, J. M. Surprisingly rapid growth in Neanderthals. Nature 428(6986), 936–939 (2004).
32. McFarlin, S. C. et al. Early brain growth cessation in wild virunga mountain gorillas (Gorilla beringei beringei). Am. J. Primatol.
75, 450–463 (2013).
33. Galbany, J. et al. Body growth and life history in wild mountain gorillas (Gorilla beringei beringei) from Volcanoes National Park,
Rwanda. Am. J. Phys. Anthropol. 163, 570–590 (2017).
34. Smith, T. M., Olejniczak, A. J., Reh, S., Reid, D. J. & Hublin, J.-J. Brief communication: enamel thickness trends in the dental arcade
of humans and chimpanzees. Am. J. Phys. Anthropol. 136, 237–241 (2008).
35. Guatelli-Steinberg, D., Floyd, B. A., Dean, M. C. & Reid, D. J. Enamel extension rate patterns in modern human teeth: two
approaches designed to establish an integrated comparative context for fossil primates. J. Hum. Evol. 63, 475–486 (2012).
36. Guatelli-Steinberg, D., Reid, D. J. & Bishop, T. A. Did the lateral enamel of Neandertal anterior teeth grow dierently from that of
modern humans?. J. Hum. Evol. 52, 72–84 (2007).
37. Reid, D. J. & Dean, M. C. Variation in modern human enamel formation times. J. Hum. Evol. 50, 329–346 (2006).
38. Liversidge, H. M. Worldwide variation in human dental development. In Growth and Development in the Genus Homo (eds
ompson, J. L. et al.) 73–113 (Cambridge University Press, Cambridge, 2003).
39. Guatelli-Steinberg, D., Larsen, C. S. & Hutchinson, D. L. Prevalence and the duration of linear enamel hypoplasia: a comparative
study of Neandertals and Inuit foragers. J. Hum. Evol. 47, 65–84 (2004).
40. Ogilvie, M. D., Curran, B. K. & Trinkaus, E. Prevalence and pattern of dental enamel hypoplasia among the Neandertals. Am. J.
Phys. Anthropol. 79, 25–41 (1989).
41. Larsen, C. S. et al. Bioarchaeology of Neolithic Çatalhöyük reveals fundamental transitions in health, mobility, and lifestyle in
early farmers. Proc. Natl. Acad. Sci. U. S. A. 116, 12615–12623 (2019).
42. Wells, J. C. & Stock, J. T. Life History Transitions at the origins of agriculture: A model for understanding how niche construction
impacts human growth, demography and health. Front. Endocrinol. 11, 325 (2020).
43. Snodgrass, J. & Leonard, W. R. Neandertal energetics revisited: insights into population dynamics and life history evolution.
PaleoAnthropol. 2009, 220–237 (2009).
44. Smith, T. M. et al. Wintertime stress, nursing, and lead exposure in Neanderthal children. Sci. Adv. https ://doi.org/10.1126/sciad
v.aau94 83 (2018).
45. Ramirez Rozzi, F. V. et al. Cutmarked human remains bearing Neandertal features and modern human remains associated with
the Aurignacian at Les Rois. J. Anthropol. Sci. 87, 153–185 (2009).
46. Hillson, S. Tooth Development in Human Evolution and Bioarchaeology (Cambridge University Press, Cambridge, 2014).
47. Schindelin, J. et al. Fiji: an open-source platform for biological-image analysis. Nat. Methods 9, 676 (2012).
48. Schlüter, M. XYZ2DEM Importer Plugin at https ://image j.nih.gov/ij/plugi ns/xyz2d em-impor ter.html (2012).
49. Ferreira, T. FindPeaks Plugin at https ://image j.net/Find_Peaks (2014).
50. R Core Team. R: A Language and Environment for Statistical Computing. R Foundation for Statistical Computing. Vienna, Austria.
At https ://www.R-proje ct.org (2019).
Content courtesy of Springer Nature, terms of use apply. Rights reserved


Vol:.(1234567890)
Scientic Reports | (2021) 11:522 | 
www.nature.com/scientificreports/
Acknowledgments
We thank Yann Heuzé, Alain Queelec, Éric Pubert, William Rendu, Margaret Stanton, Horst and Uwe Kierdorf
and the University of Hildesheim, Christopher Knüsel, Ottmar Kullmer and the Senckenberg Museum, the
Domarchäologie Hildesheim and the Çatalhöyük Research Project for their guidance, helpful discussions, and
access to dental replicas for this project.
Author contributions
K.M. conceived of the methodological protocol with contributions from D.G.-S., D.J.R., E.B., S.E.Z., S.C.M.. K.M.,
D.G.-S., E.B., S.E.Z. designed the study. S.E.Z., E.B., C.W. provided replicas for analysis. K.M., L.S.L., A.L.-L. col-
lected confocal images. K.M. conducted analyses and wrote the manuscript with contributions from all authors.
Funding
is project has received funding from the European Union’s Horizon 2020 research and innovation program
under the Marie Sklodowska-Curie Grant Agreement No 798117, as well as support from the Deutsche Forschun-
gsgemeinscha (DFG) Grant No 353106138, Ministerium für Wissenscha, Forschung und Kunst Baden-Würt-
temberg, and the British Academy for EB’s Fellowship at the University of Kent (EB).
Competing interests
e authors declare no competing interests.
Additional information
Supplementary Information e online version contains supplementary material available at https ://doi.
org/10.1038/s4159 8-020-80148 -w.
Correspondence and requests for materials should be addressed to K.M.
Reprints and permissions information is available at www.nature.com/reprints.
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and
institutional aliations.
Open Access is article is licensed under a Creative Commons Attribution 4.0 International
License, which permits use, sharing, adaptation, distribution and reproduction in any medium or
format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the
Creative Commons licence, and indicate if changes were made. e images or other third party material in this
article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the
material. If material is not included in the article’s Creative Commons licence and your intended use is not
permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from
the copyright holder. To view a copy of this licence, visit http://creat iveco mmons .org/licen ses/by/4.0/.
© e Author(s) 2021
Content courtesy of Springer Nature, terms of use apply. Rights reserved

1.
2.
3.
4.
5.
6.
Terms and Conditions
Springer Nature journal content, brought to you courtesy of Springer Nature Customer Service Center GmbH (“Springer Nature”).
Springer Nature supports a reasonable amount of sharing of research papers by authors, subscribers and authorised users (“Users”), for small-
scale personal, non-commercial use provided that all copyright, trade and service marks and other proprietary notices are maintained. By
accessing, sharing, receiving or otherwise using the Springer Nature journal content you agree to these terms of use (“Terms”). For these
purposes, Springer Nature considers academic use (by researchers and students) to be non-commercial.
These Terms are supplementary and will apply in addition to any applicable website terms and conditions, a relevant site licence or a personal
subscription. These Terms will prevail over any conflict or ambiguity with regards to the relevant terms, a site licence or a personal subscription
(to the extent of the conflict or ambiguity only). For Creative Commons-licensed articles, the terms of the Creative Commons license used will
apply.
We collect and use personal data to provide access to the Springer Nature journal content. We may also use these personal data internally within
ResearchGate and Springer Nature and as agreed share it, in an anonymised way, for purposes of tracking, analysis and reporting. We will not
otherwise disclose your personal data outside the ResearchGate or the Springer Nature group of companies unless we have your permission as
detailed in the Privacy Policy.
While Users may use the Springer Nature journal content for small scale, personal non-commercial use, it is important to note that Users may
not:
use such content for the purpose of providing other users with access on a regular or large scale basis or as a means to circumvent access
control;
use such content where to do so would be considered a criminal or statutory offence in any jurisdiction, or gives rise to civil liability, or is
otherwise unlawful;
falsely or misleadingly imply or suggest endorsement, approval , sponsorship, or association unless explicitly agreed to by Springer Nature in
writing;
use bots or other automated methods to access the content or redirect messages
override any security feature or exclusionary protocol; or
share the content in order to create substitute for Springer Nature products or services or a systematic database of Springer Nature journal
content.
In line with the restriction against commercial use, Springer Nature does not permit the creation of a product or service that creates revenue,
royalties, rent or income from our content or its inclusion as part of a paid for service or for other commercial gain. Springer Nature journal
content cannot be used for inter-library loans and librarians may not upload Springer Nature journal content on a large scale into their, or any
other, institutional repository.
These terms of use are reviewed regularly and may be amended at any time. Springer Nature is not obligated to publish any information or
content on this website and may remove it or features or functionality at our sole discretion, at any time with or without notice. Springer Nature
may revoke this licence to you at any time and remove access to any copies of the Springer Nature journal content which have been saved.
To the fullest extent permitted by law, Springer Nature makes no warranties, representations or guarantees to Users, either express or implied
with respect to the Springer nature journal content and all parties disclaim and waive any implied warranties or warranties imposed by law,
including merchantability or fitness for any particular purpose.
Please note that these rights do not automatically extend to content, data or other material published by Springer Nature that may be licensed
from third parties.
If you would like to use or distribute our Springer Nature journal content to a wider audience or on a regular basis or in any other manner not
expressly permitted by these Terms, please contact Springer Nature at
onlineservice@springernature.com