Enamel surface topography analysis for diet discrimination. A methodology to enhance and select discriminative parameters. (5th January 2018)

• Record Type: Journal Article
• Authors: Francisco, Arthur; Blondel, Cécile; Brunetière, Noël; Ramdarshan, Anusha; Merceron, Gildas
• Is Part Of: Surface topography. Volume 6:Number 1(2018:Jan./Mar.)
• Publication Date: 2018-01-05
• Subjects: surface statistics -- dental microwear -- surface texture -- surface anisotropy; Surfaces (Physics) -- Periodicals; Surfaces (Physics) -- Measurement -- Periodicals; 530.417
• DOI: 10.1088/2051-672X/aa9dd3
• Languages: English
• ISSNs: 2051-672X

Abstract

Abstract: Tooth wear and, more specifically, dental microwear texture is a dietary proxy that has been used for years in vertebrate paleoecology and ecology. DMTA, dental microwear texture analysis, relies on a few parameters related to the surface complexity, anisotropy and heterogeneity of the enamel facets at the micrometric scale. Working with few but physically meaningful parameters helps in comparing published results and in defining levels for classification purposes. Other dental microwear approaches are based on ISO parameters and coupled with statistical tests to find the more relevant ones. The present study roughly utilizes most of the aforementioned parameters in their more or less modified form. But more than parameters, we here propose a new approach: instead of a single parameter characterizing the whole surface, we sample the surface and thus generate 9 derived parameters in order to broaden the parameter set. The identification of the most discriminative parameters is performed with an automated procedure which is an extended and refined version of the workflows encountered in some studies. The procedure in its initial form includes the most common tools, like the ANOVA and the correlation analysis, along with the required mathematical tests. The discrimination results show that a simplified form of the procedure is able to more efficiently identify the desired number of discriminative parameters. Also highlighted are some trends like the relevance of working with both height and spatial parameters, as well as the potential benefits of dimensionless surfaces. On a set of 45 surfaces issued from 45 specimens of three modern ruminants with differences in feeding preferences (grazing, leaf-browsing and fruit-eating), it is clearly shown that the level of wear discrimination is improved with the new methodology compared to the other ones.