a repository copy of Morphological parametric mapping of 21 skin sites throughout the body using

Background: Changes in body posture cause changes in morphological properties at different skin sites. Although previous studies have reported the thickness of the skin, the details of the postures are not generally given. This paper presents the effect of a change in posture on parameters such as thickness and surface roughness in 21 load-bearing and non-load-bearing sites. Materials and methods: A total of 12 volunteers (8 males and 4 females) were selected in an age group of 18 – 35 years and of Fitzpatrick skin type I-III. Images were captured using a clinically-approved VivoSight ® optical coherence tomography system and analysed using an algorithm provided by Michelson Diagnostics. Results: Overextension (extending joints to full capacity) resulted in changes to thickness, roughness and undulation of the skin around the body. Discussion and conclusion: The load-bearing regions have thicker skin compared to non-load-bearing sites. This is the irst time that undulation topography of the stratum corneum – stratum lucidum and the dermal – epidermal junction layers have been measured and reported using statistical values such as Ra. The data presented could help to deine new skin layer models and to determine the variability of the skin around the body and between participants.


Introduction
The skin is a complex system made up of different layers, namely the epidermis (E), dermis (D) and hypodermis. The epidermis is itself made up of various layers including the stratum corneum (SC) and stratum lucidum (SL). Although abundant research exists on the chemistry, composition and function of the skin, little is known about morphological parameters such as thickness, roughness and undulation at different anatomical locations. The thickness of the layers varies with age (Li et al., 2006;Tsugita et al., 2013;Vashi et al., 2016), ethnicity (Vashi et al., 2016), gender (Lee and Hwang, 2002), anatomical location (Tsugita et al., 2013;Lee and Hwang, 2002;Barker, 1951) and presence of conditions such as atopic dermatitis and psoriasis (Byers et al., 2017).
Two previous studies (Maiti et al., 2016;Beaudette et al., 2017) found that morphological parameters of the skin layers changed when the forearm was stretched, meaning that variation in skin layer measurements depend on body segment position. Therefore, comparing studies and individuals is challenging due to variations in measurement protocols, intra-individual body site and inter-individual hormonal differences. Inter-individual hormonal variations involve differences in fatty acid, lipids and cholesterols that inluence skin properties among individuals (Norlen et al., 1999). The aims of the project are therefore: 1) to determine differences in skin measurements at two extreme postures of the body (relaxed and overextension); and 2) quantify morphological parameters such as SC or E thickness (T SC /T E ), roughness of the skin surface (R SS ) and SC-SL/D-E (R SCLJ /R DEJ ) junction layers. The data presented in this paper could help in proposing new skin layer models, determining variability of the skin in the body, understanding skin differences between individuals and testing the eficacy of pharmaceutical products on the skin.

Materials and methods
Skin sites of 12 volunteers were captured using OCT to obtain the thickness of the epidermis (E) and stratum corneum (SC); skin surface roughness (R SS ); and undulation topographies of the dermal-epidermal (R DEJ ) and stratum corneum-stratum lucidum (R SCLJ ) junction layers at two different postures: relaxed and in overextension.

Volunteers and the OCT system
Twelve volunteers aged 18-35 years old, with Fitzpatrick skin type I-III and with no previous history of skin ailments, were recruited. The effect of change in posture on skin parameters was investigated for same three out of the total 12 volunteers. The study was approved by the Ethics committee of the University of Shefield Medical School. Images of 28 skin sites around the body were captured using a clinicallyapproved VivoSight® OCT system (Michelson Diagnostics, Kent, UK). The VivoSight® is a Fourier domain OCT system with a 20 kHz swept source laser at 1300 nm centre wavelength, 7.5 μm lateral and 5 μm axial resolution. It captures 20 frames per second with an image size of 1342 � 460 pixels. The OCT image volume obtained from each skin site was 6 � 6 � 2 mm 3 (width x length x depth).

Measurement protocol
Informed consent was irst obtained from the volunteers. The skin sites were cleaned with Medipal alcohol wipes (Pal, Leicestershire) to remove sebum and then acclimatised for a period of 15 min at 19-21 � C room temperature and 30-40% humidity. A non-permanent tattoo skin marker pen (Shenzhen Badamu Keji Ltd, China) was used to mark a point at every site where measurements were taken (Table 1). Overextension was deined as the maximum angle of extension a joint could sustain without causing pain, and was taken as the lowest angle measured across all volunteers.
The skin sites of all 12 volunteers were then measured at the overextended position. Skin sites of three of the 12 volunteers were also captured in the relaxed state to compare with the same volunteers in overextended. The angles obtained during both states (�5 � ) are reported in Table 1. There was no contact between the OCT probe and the skin sites during either the relaxed or overextended state.
Facial sites were also captured using OCT: the frontalis (above eyebrow), superior orbital (below eyebrow), external meatus (ear canal), post auricular (behind ear), malar region (cheek) and suborbicularis oculi fat (eye bag). Overextension was achieved using a force applied to the skin (for example, below the eyebrow) by the investigator as shown in Fig. 1. The force applied was based on lexibility of the sites and therefore this method could not be repeated accurately. Hence, the measurements of facial sites were analysed and represented in the range of two values.
A total of 50 images were captured per site at each position for each participant, of which six good quality images were analysed using the algorithm to obtain statistical differences. The maximum thickness measured was limited by the depth resolution of the laser. Hence, for load-bearing sites such as the heels, ingertips and feet, only SC thickness and SC-SL junction layer undulations are presented. Skin surface roughness and undulation topography of the junction layer deines irregularities in terms of smoothness, peaks and troughs. These parameters were calculated by averaging the height of peaks and troughs over the sample length, and are given as mean � SEM.
Statistical analysis for the normally-distributed data was carried out using paired t-tests and 2-way ANOVA tests in GraphPad Prism 7 (GraphPad Software, San Diego, USA) with signiicance as p < 0.05. Pearson correlation was performed between thickness, roughness and undulation and skin type. Only data with strong relationships (r > 0.6 and p < 0.05) are presented in the results section. All the p values between morphological parameters are reported in brackets.

Results
Mean weight and height of the volunteers were 63.3�15 kg and 176.1�12 cm (Table 2).

Differences in skin morphology between load-bearing and non-loadbearing sites
The SC was thicker in load-bearing sites such as the ingertip (297 μm), heel (311 μm) and foot (601 μm) compared to non-loadbearing sites (Fig. 2).
The T E , R SS and R DEJ of all non-load-bearing sites varied from 80-126 μm, 2-5 μm and 2-10 μm respectively. The T SC, R SS and R SC-SL for   There was no signiicant relationship between skin type and any morphological parameter except for SC thickness at the heel (r ¼ À0.86, p ¼ 0.0004) and D-E junction layer undulation of the back of the knee (r ¼ 0.78, p ¼ 0.003), Fig. 3.
No signiicant difference (p > 0.99) was seen in any morphological parameters between volar forearm skin and other sites. This supports the methodology of using the forearm instead of the cheek for the design and development of shaving devices, sensors etc.

Effect of posture loading: difference between relaxed and overextended states
Posture loading was applied to the sites below the neck (non-facial regions) to investigate morphological changes in the skin between the relaxed and overextended states (Fig. 4). Overextension caused a signiicant reduction of SC thickness (Fig. 4a) in sites such as the foot (450 μm-347 μm; p ¼ 0.02); and E thickness (Fig. 4b) in sites such as the back of the knee (115 μm-86 μm; p ¼ 0.04), neck (127 μm-102 μm; p ¼ 0.03) and hand (130 μm-107 μm; p ¼ 0.02). The only site that increased in thickness with overextension was the SC of the heels, from 267 μm to 316 μm (p ¼ 0.02); the underarm and volar forearm did not show a signiicant change.
The hand and the back of the knee showed a signiicant decrease (p < 0.01) in D-E Junction layer undulation with overextension (Fig. 4f).
The thickness of load-bearing sites (mean 315 μm) such as the foot, ingertip and heel was signiicantly (p < 0.01) higher than the E thickness of non load-bearing sites (Fig. 4b) in both types of postures.   4. Changes in morphological parameters in relaxed (n ¼ 3) and overextended (n ¼ 12) states. Thickness of stratum corneum (a) and epidermis (b); average skin surface roughness of stratum corneum (c) and epidermis (d); average undulation of stratum corneum-stratum lucidum junction layer (e) and dermal-epidermal junction layer (f). * represents a signiicant difference (p < 0.05) between the postures. SC: stratum corneum; SL: stratum lucidum; D: dermis; E: epidermis. However, the SC is part of the epidermis so thickness should not be compared. The mean roughness of all load-bearing sites (2.9 μm; Fig. 4c) was signiicantly (p < 0.01) lower than non load-bearing sites (4.1 μm; Fig. 4d).

Effect of manual loading
Manual loading was applied to the facial regions (Fig. 5a, b and c). There was a lack of repeatability between two consecutive measurements on any facial sites during manual loading. The E thickness of the above-eyebrow site increased with overextension from 110.5 to 132.3 μm. However, E thickness in the below-eyebrow site changed with overextension from 112.4 to 89.0 μm. Opposite trends to thickness were seen in roughness and junction layer undulation for the above-eyebrow site (R SS : 2.7 to 2.4 μm; R DEJ junction: 2.8 to 2.6 μm).
There was no difference between males and females in terms of skin surface roughness ( Fig. 6c and 6d) or junction undulation ( Fig. 6e and  6f); except for undulation of the D-E junction in the chest, which was signiicantly (p ¼ 0.02) smaller in males (2.8 � 0.3 μm) than females (4.4 � 0.7 μm). There was no signiicant difference between male and female skin surface roughness of either load-bearing (3.0 � 0.7 μm) or non-load bearing (3.3 � 0.5 μm) sites.

Discussion
Overextension generally decreased skin thickness, due to muscle activation and bulging that initiated stretching and contraction of the skin. In addition, the orientation of Langer's lines (Langer, 1861) affects the epidermal layer thickness: a higher decrease was seen when Langer's lines were perpendicular to the direction of skin extension such as back of knee, shoulder blade (Ni Annaidh et al., 2012) i.e., the skin will better resist deformation when the force is aligned with ibre direction. In the current paper, not all posture movements were in the direction of Langer's lines. Hence, the thickness of the skin decreases with overextension as seen in most of the load bearing and non-load bearing sites except heels, underarm and volar forearm. In the current study, there were no signiicant differences in the underarm or volar forearm. The SC thickness in the heel increased with overextension, possibly because of calluses on the heels. Although similar observations were reported in earlier studies (Maiti et al., 2016;Beaudette et al., 2017), comparison is dificult as only one participant was studied in one study and the sites reported were forearm and lower back.
The forces applied in manual loading did vary with elasticity of the skin, and the lack of repeatability resulted in inaccuracy of the loading protocol. For example, on application of force on an eye bag (Fig. 7a) the thickness did not change signiicantly. However, the roughness of the skin surface and E-DE junction layers did change signiicantly. In posture-type loading (Fig. 7b), overextension decreased thickness and roughness parameters. Efforts to standardise the applied force on the facial sites will be developed in the future to attain repeatable measurements.
The higher SC thickness in female ingertips, feet and heels might be due to the greater number of corneocytes (Machado et al., 2010) or higher pH content (Danby et al., 2016). Although the difference was not signiicant, epidermal layers in the female lower back and chest were thicker than in males; this may be due to the larger hip size and presence of mammary glands.
The SEM for skin roughness in the heels of the female participants was very high, possibly because of the low numbers in this study or lifestyle differences between the male and female cohorts.
Junction layer undulation could not be compared to previous studies due to lack of data in the literature. One study found a higher undulation in the lower forearm (Maiti et al., 2016), although this might have been due to the small number of volunteers used in the two studies.
The limitation of the maximum depth that the VivoSight® could measure meant that SC thickness of load-bearing sites was compared to epidermal thickness of non-load-bearing sites. Future studies could be conducted with higher depth (lower resolution) OCT to compare epidermal thickness of these sites.
Skin data was calculated based on the assumption that the participant maintained their posture during imaging. However, muscle bulging and activation is also inluenced by muscle strength. Future studies could quantify muscle strength, skin top surface stretching using DIC while measuring thickness and roughness similar to earlier reported studies (Maiti et al., 2016;Panchal et al., 2019). A 3D skin model built through DIC surface strain and OCT sub surface will be useful for clinicians for diagnosis of skin tumour in future. In addition, the data collected in the current study is limited to Fitzpatrick skin type I-III and age group 18-35 years. Future studies will be conducted on a broader range of skin types/ethnicities and age groups with visco-elastic variation among the volunteers.

Conclusions
The study demonstrated the in vivo variation of skin properties such Fig. 6. Comparison of morphological parameters. Thickness of stratum corneum (a) and epidermis (b); skin surface roughness for load-bearing (c) and non load-bearing (d) sites; undulations at the stratum corneum-stratum lucidum (e) and dermal-epidermal junction layer (f) in the overextended state. * represents signiicance difference (p < 0.05) between males and females. SC: stratum corneum; SL: stratum lucidum; D: dermis; E: epidermis. as thickness and roughness with posture, gender and site. Change in posture led to signiicant differences in the thickness and roughness of non-facial sites. However, these changes cannot easily be compared to previous literature due to lack of detail in the measurement protocol. The stratum corneum thickness, roughness and undulation in loadbearing sites are different from non load-bearing sites. This is the irst time that stratum corneum-stratum lucidum roughness and epidermaldermal junction layer undulation topography have been measured and reported using statistical values such as Ra for 21 sites. These morphological parameters lay a foundation for future work such as modelling the biomechanics of the skin, investigating the structure of unaffected skin compared to affected skin for eczema and skin cancer treatment, and studying drug infusion therapies.

Declaration of competing interest
The authors declare that they have no conlict of interest. Fig. 7. Difference between facial (deformation with low repeatability) and non-facial sites (deformation with high repeatability) undergoing manual and posture loading (represented by arrow) in a) eye bag and b) wrist sites. 'X' denote the measurement point.