Lance Waller
Rollins School of Public Health
Department of Biostatistics
Emory University
A Spatial Analysis of Epidermal Nerve Fibers
Epidermal nerve fibers represent thin nerves extending from the dermis
to the epidermis and transmit sensations of heat and pain to the central
nervous system. Patients suffering from neuropathy (e.g., as a result
of diabetes) exhibit quantifiable loss in the number of ENFs present in
skin biopsies. Stained images of the biopsies suggest increased
clustering in the spatial pattern of ENFs preceding the loss in number,
and suggest the possibility of early diagnosis if such changes in
pattern can be quantified. We review the basic neurology of ENFs, and
methods for quantifying clustering in spatial point processes, then
apply these ideas to data from the University of Minnesota involving
repeated skin biopsies from samples of non-diseased patients and those
suffering from small fiber sensory neuropathy, often a precursor to
neuropathy. Methods include kernel-based estimation of the
pair-correlation function, and Monte Carlo assessments of significance.
This is joint work with Traci Leong (Emory), William Kennedy
(Minnesota), Gwen Wendelschafer-Crabb (Minnesota) and David Walk
(Minnesota).
Back to Colloquium Series