Faculty Bibliography

Approximate entropy (ApEn) is a recently developed formula to quantify the amount of regularity in data. We examine the potential applicability of ApEn to clinical endocrinology to quantify pulsatility in hormone secretion data. We evaluate the role of ApEn as a complementary statistic to widely employed pulse-detection algorithms, represented herein by ULTRA, via the analysis of two different classes of models that generate episodic data. We conclude that ApEn is able to discern subtle system changes and to provide insights separate from those given by ULTRA. ApEn evaluates subordinate as well as peak behavior and often provides a direct measure of feedback between subsystems. ApEn generally can distinguish systems given 180 data points and an intra-assay coefficient of variation of 8%. This suggests ApEn as applicable to clinical hormone secretion data within the foreseeable future. Additionally, the models analyzed and extant clinical data are both consistent with episodic, not periodic, normative physiology

Estrogen treatment induces synaptic plasticity accompanied by damaged structures and aggregates of peroxidase in astrocytes in the hypothalamic arcuate nucleus of the rat. Synaptic plasticity also occurs within the arcuate nucleus after physiologic surges of estrogen. Although the function of estrogen-induced peroxidase is unclear at present, in other systems peroxidase can generate free radicals by catalyzing the oxidation of some molecules, including estrogen. Because free radicals underlie remodeling in a number of tissues, estrogen-induced free radicals could mediate synaptic remodeling within the arcuate nucleus. Although they contain estrogen-inducible peroxidase, astrocytes do not contain estrogen receptors as measured by conventional techniques, suggesting that estrogen-inducible peroxidase arises from some novel mechanism. Estrogen could induce peroxidase within receptor-deficient astrocytes by binding to receptors in neurons and stimulating the release of some factor that interacts with astrocytes. Alternatively, estrogen could act directly on astrocytes in the absence of estrogen receptors. Although astrocytes in the hypothalamus of the rat do not contain classical nuclear estrogen receptors, they do bind fluorescein-conjugated estradiol in extranuclear sites. The distribution of fluorescein-conjugated estradiol binding within the hypothalamus overlaps that of peroxidase-rich astrocytes, and double labeling reveals many cells with the stellate morphology of astrocytes, containing both peroxidase and fluorescein-conjugated estradiol binding. However, because peroxidase and fluorescein-conjugated estradiol always occupy different compartments of the cell, the fluorescein-conjugated estradiol is not binding to peroxidase

Despite the well known role of the light-dark cycle in the entrainment of circadian rhythms, very little is known about the neurochemical events that mediate the effects of light on the mammalian circadian clock. Recent anatomical and pharmacological data support the hypothesis that acetylcholine may be involved in relaying light-dark information from the retina to, or within, the circadian clock of rodents. If acetylcholine is required for this response, it should be possible to block the phase-shifting effects of a light pulse by blocking cholinergic neurotransmission. To test this possibility, hamsters free-running in constant darkness received an intraventricular injection of the anticholinergic drug, mecamylamine (450 micrograms), 10 min before being exposed to a 5-min pulse of light known to induce sub-maximal phase shifts in the circadian rhythm of wheel-running behavior. Compared to vehicle-injected control animals, mecamylamine treatment blocked or reduced both the phase-advancing and phase-delaying effects of light. These results support the hypothesis that acetylcholine is involved in mediating the phase-shifting effects of light on the mammalian circadian clock