Capturing dependencies in images in an unsupervised manner is important for many image-processing applications and for understanding the structure of natural image signals. Data generative linear models such as principal component analysis (PCA) and independent component analysis (ICA) have shown to capture low level features such as oriented edges in images. However, those models capture only linear dependency and therefore its modeling capability is limited. We propose a new method for capturing nonlinear dependencies in images of natural scenes. This method is an extension of the linear ICA method and builds on a hierarchical representation. The model makes use of lower level linear ICA representation and a subsequent mixture of Laplacian distribution for learning the nonlinear dependencies in an image. The model parameters are learned via the expectation maximization (EM) algorithm and it can accurately capture variance correlation and other high order structures in a simple and consistent manner. We visualize the learned variance correlation structure and demonstrate applications to automatic image segmentation and image denoising. (c) 2006 Elsevier B.V. All rights reserved.

Onychomycosis is a common and difficult-to-treat fungal infection of the nail unit that gradually leads to dystrophic changes of the nail plate and nail bed. If untreated, infection progresses and may lead to discomfort, reduced quality of life, and risk of complications in patients with comorbid conditions (eg, diabetes, human immunodeficiency virus, peripheral vascular disease). Onychomycosis treatments are designed to eradicate causative pathogens (most commonly Trichophyton rubrum and Trichophyton mentagrophytes), restore healthy nails, and prevent recurrence or spread of infection. Given the deep-seated nature of most cases of onychomycosis, an effective antifungal agent needs to achieve and maintain sufficient drug concentrations throughout the nail unit for the duration of healthy nail in-growth. Oral antifungal drugs are the most effective available therapy and are generally well tolerated, but may be limited by safety concerns and the potential for drug-drug interactions (DDIs). Thus, treating physicians and pharmacists must be cognizant of a patient’s current medications; indeed, it may not be feasible to treat onychomycosis in patients with diabetes, heart disease, or depression because of the risk for DDIs. Current topical therapy is not associated with risk of DDIs. Tavaborole and efinaconazole, two recently approved topical agents, have demonstrated good nail penetration and high negative culture rates in clinical trials of patients with onychomycosis. This article provides the treating physician and pharmacist with information on the safety and effectiveness of current oral (allylamine, azole) and topical (ciclopirox, efinaconazole, tavaborole) treatment to aid in making informed treatment decisions based on the unique characteristics (medication history, comorbidities, nature of onychomycosis) of each patient.

Somatosensory thalamus and cortex in rodents contain topological representations of the facial whisker pad. The thalamic representation of a single whisker ("barreloid") is presumed to project exclusively to the cortical representation ("barrel") of the same whisker; however, it was not known when this correspondence is established during early development, nor how precise the thalamocortical projection is at birth, before formation of barrels and barreloids. To answer these questions, we retrogradely labeled thalamocortical projection neurons in fixed brain slices from 0-8 d old (P0-P8) mice, by placing paired deposits of two fluorescent dyes in adjacent barrels or (before barrel formation) in adjacent loci in upper cortical layers. At all ages studied, a negligible fraction of the retrogradely labeled cells was double labeled, implying that branches of single thalamocortical axons never extended within layer IV over an area wider than a single barrel. In P0 preparations, 70% of paired dye deposits placed 75-200 microns apart resulted in statistically significant segregation of labeled cell clusters in the thalamus. Quantitative analysis indicated that on P0 about 70% of thalamocortical axons were within 1.3 presumptive barrel diameters from their topologically precise target. In P4-P8 preparations, the great majority of thalamic cells retrogradely labeled from a single barrel were found in a single barreloid, implying a 1:1 projection of barreloids to barrels. The postnatal increase in topological precision was reproduced by a computer simulation, which assumed that many aberrant axons corrected their initial targeting error by extending terminal arborizations asymmetrically, towards the center of their appropriate barrel.

The thalamocortical projection to rodent somatosensory ("barrell") cortex is highly ordered in both the radial and the tangential dimensions. During a brief period of postnatal development, thalamocortical axons establish two tiers of terminations, in the deep layers and in layer IV, and form whisker-specific clusters within layer IV; however, little is known about the cues that guide them to their appropriate radial and tangential positions. To gain insight into potential mechanisms underlying this process, we studied the development of thalamocortical termination patterns in mouse barrel cortex at high spatial resolution. Developing thalamocortical axons were labeled in fixed slices with the lipophilic carbocyanine dye Dil and imaged with a laser scanning confocal microscope. On the day of birth (postnatal day 0, P0) axons coursed through layers VI and V, with little or no branching. By P2 the lower tier of terminations, at the border of layers VI and V, was clearly identifiable. Below this tier axons coursed obliquely or tangentially, forming a dense meshwork of intersecting fibers, but with no apparent branching. By P4 the upper tier of terminations, in layer IV, was clearly recognizable, and consisted of periodic, dense clusters of terminal arborizations. In marked contrast to the oblique and apparently disorderly course followed by axons in layer VI and lower layer V, axons in upper layer V heading toward the upper tier were organized in loose bundles running radially, suggesting that axons destined to terminate in a particular layer IV barrel had already reached their appropriate tangential coordinates within the lower tier. Thus, the pattern of thalamocortical terminations in layer IV seems to be projected from the deep tier of terminations, and does not develop from an initially profuse arborization pattern through pruning of inappropriate branches.

The objective of this study was to determine the effects of tibial rotation on in situ strain in the peripatellar retinaculum and patellofemoral contact pressures and areas. Patellofemoral joint biomechanics demonstrate a strong correlation with the etiology of patellofemoral disorders, such as chondromalacia, and are significantly influenced by tibial rotation. Six human cadaveric knees were used along with a patellofemoral joint testing jig that permits physiological loading of the knee extensor muscles. Patellofemoral contact pressures and areas were measured with a Fuji pressure-sensitive film, and the changes in in situ strain in the peripatellar retinaculum were measured with four differential variable reluctance transducers. Tibial rotation had a significant effect on patellofemoral joint biomechanics. The data showed an inverse relationship between increasing knee flexion angles and the change in patellofemoral contact pressures and in situ strain with tibial rotation. At higher knee flexion angles, the patella is well-seated in the trochlear groove and the function of the peripatellar retinaculum is minimized and less affected by tibial rotations.