Statistical Models for Landmark Coordinate Data - An Invariant Approach to Statistical Analysis of Shapes - page 90

mity in dentition, requirements of the palate to separate the oral and
nasal orifices, and attachment sites for soft tissues, may account for
the relatively reduced variances for these two landmarks. The inter-
pretations that we offer as potential reasons for local differences in
variation come from prior knowledge of the anatomy and development
of the primate skull, and should be further investigated. Prior knowl-
edge from any relevant discipline can and should be used to interpret
local differences in variation.

3.8.3 Neurocranial morphology in craniosynostosis

Eight landmarks from the external surface of the neurocranium
(calotte or skullcap) were collected from computed tomography (CT)
scans of 16 infants and children diagnosed with premature closure of
the sagittal suture. These children range in age from six weeks to
seven years of age. The mean form matrix, mean form, and variance-
covariance matrix for these data are given in Table 3.4a- c.

Table 3.4a. Mean form of the calotte of children with sagittal synostosis
( mm ). (Coordinates corresponding to the mean form matrix).
Landmark (number)

X Y Z

Right asterion (1) 41.509 42.599 24.623

Left asterion (2) 43.008 -40.134 26.122

Bregma (3) -38.021 -1.320 -40.743

Lambda (4) 68.148 -0.770 -27.897

Right frontal boss (5) -68.448 27.304 12.088

Left frontal boss (6) -67.125 -28.598 13.242

Right parietal boss (7) 9.464 50.580 -3.954

Left parietal boss (8) 11.464 -49.661 -3.481

Variances are very large in this small sample. One reason for this lies
in the extremely large age range represented by the children that
make up this sample. Given that more than two thirds of adult head
size is attained by two and one half years of age, this sample includes
an extreme range of sizes, from specimens very close to the smallest
human postnatal size possible to those approaching 75% of their even-
tual adult head size. Additionally, the frontal and parietal bosses
(landmarks 5, 6, 7, 8) represent “fuzzy landmarks” (see Chapter 2 ), the
measurement error of which has been shown to be relatively larger
than that of more traditional landmarks (Valeri et al., 1998).