Hidden Conditional Random Fields for Visual Speech Recognition

In this paper we present the application of Hidden Conditional Random Fields (HCRFs) to modelling speech for visual speech recognition. HCRFs may be easily adapted to model long range dependencies across an observation sequence. As a result visual word recognition performance can be improved as the model is able to take more of a contextual approach to generating state sequences. Results are presented from a speaker-dependent, isolated digit, visual speech recognition task using comparisons with a baseline HMM system. We firstly illustrate that word recognition rates on clean video using HCRFs can be improved by increasing the number of past and future observations being taken into account by each state. Secondly we compare model performances using various levels of video compression on the test set. As far as we are aware this is the first attempted use of HCRFs for visual speech recognition.

Mast cell mediator release in nonasthmatic subjects after endobronchial adenosine challenge

Background: Adenosine 5′-monophosphate (AMP) has been shown to cause bronchoconstriction in atopic subjects but to have no effect on nonatopic nonasthmatic subjects. Endobronchial AMP challenge has previously been shown to cause mast cell mediator release in asthmatic subjects, but it is unknown whether a similar response occurs in atopic nonasthmatic and nonatopic nonasthmatic control subjects who have no response to inhalation AMP challenge.

Objective: This study examined the change in mast cell–derived products after endobronchial saline challenge and AMP challenge in subjects with and without a positive inhalation response to AMP.

Methods: Inhalation challenge with AMP challenge was performed in normal, atopic nonasthmatic, and atopic asthmatic subjects. Levels of mast cell mediators were measured after endobronchial adenosine challenge and after placebo endobronchial saline challenge.

Results: There were significant increases in histamine, tryptase, protein, and prostaglandin D2 levels (P = .02, P = .02, P = .01, and P = .01, respectively) after AMP challenge compared with after saline challenge in nonatopic nonasthmatic subjects. There was no significant increase in any mediator in either of the other 2 groups.

Conclusion: This study suggests dissociation between mediator release and bronchoconstriction in response to AMP.

A two compartment in vitro release model for the testing of HIV microbicide formulations

BACKGROUND: In vitro release testing of vaginal formulations is usually performed in a one-compartment model (OCM) where the release medium, usually comprising pH-adjusted water, an aqueous surfactant solution or a solvent-water solution, provides sink conditions throughout the release experiment. Although this model is useful in evaluating the effect of formulation parameters upon release, it rarely reflects in vivo conditions. Here we report use of a two-compartment diffusion cell model (TCM, comprising a small volume donor, a large volume receptor, and separated by a model epithelial membrane) to more closely mimic in vivo vaginal release and tissue absorption following administration of a UC781 vaginal ring.

METHODS: Macaque-sized matrix silicone elastomer vaginal rings containing 100mg UC781 were prepared by injection molding, and in vitro release testing performed using both OCM (20mL simulated vaginal fluid, SVF) and TCM (5mL SVF in donor cell and variable quantities of Tween 80; silicone elastomer membrane; 100mL 3:2 ethanol/water in receptor cell). In the TCM, drug levels were measured by HPLC in both donor and receptor cells, representing fluid and tissue levels respectively. Rings containing 100mg UC781 and 10% w/w Tween 80 were also manufactured and tested.

RESULTS: The amount of UC781 released from rings was significantly influenced by the choice of release model. Greatest release (56mg/14 days) was measured in the ethanol/water OCM, compared with no measurable release into SVF only. Increasing the concentration of Tween 80 in the SVF medium (1, 3 and 5% w/w) led to increased UC781 release (11, 16 and 18mg, respectively), demonstrating that vaginal fluid solubility of UC781 may be rate-determining in vivo. In the TCM, UC781 accumulates in the receptor cell medium over time, despite not being measured in the donor medium containing the ring device. Incorporation of Tween 80 directly into the ring provided enhanced release in both donor and receptor cells.

CONCLUSIONS: Release of UC781 was influenced by the choice of release medium and the inclusion of Tween 80 in the ring. Although use of SVF-only in the OCM indicated no measurable UC781 release from rings, data from the TCM confirms that UC781 is not only released but is also capable of penetrating across the model epithelial membrane. The TCM may therefore provide a more representative in vitro release model for mimicking in vivo absorption.

Adenosine induces histamine release from human bronchoalveolar lavage mast cells

Previous studies have shown that in vitro adenosine enhances histamine release from activated human lung mast cells obtained by enzymic dispersion of lung parenchyma. However, adenosine alone has no effect on histamine release from these cells. Given the evidence for direct activation of mast cells after endobronchial challenge with adenosine and previous studies indicating that mast cells obtained at bronchoalveolar lavage are a better model for asthma studies than those obtained by enzymic dispersion of lung tissue, the histamine-releasing effect of adenosine was examined on lavage mast cells. Bronchoalveolar lavage fluid was obtained from patients attending hospital for routine bronchoscopy (n = 54). Lavage cells were challenged with adenosine or adenosine receptor agonists (20 min, 37 degrees C) and histamine release determined using an automated fluorometric assay. Endogenous adenosine levels were also measured in lavage fluid (n = 9) via an HPLC method. Adenosine alone caused histamine release from ravage mast cells in 37 of 54 patients with a maximal histamine release of 20.56 +/- 2.52% (range 5.2-61 %). The adenosine receptor agonists (R)-N-6-(2-phenylisopropyl)adenosine, 5'-N-ethylcarboxamido-adenosine and CGS21680 also induced histamine release from lavage mast cells. Preincubation of lavage mast cells with the adenosine receptor antagonist xanthine amine congener caused significant inhibition of the response to adenosine (P = 0.007). There was an inverse correlation between endogenous adenosine levels in the lavage fluid and the maximal response to in vitro adenosine challenge of the lavage cells. The findings of the present study indicate a means by which adenosine challenge of the airways can induce bronchoconstriction and support a role for adenosine in the pathophysiology of asthma. The results also suggest that cells obtained from bronchoalveolar ravage fluid may provide the ideal model for the testing of novel, adenosine receptor, targeted therapies for asthma.

Histamine release from bronchoalveolar lavage cells from asthmatic subjects after allergen challenge and relationship to the late asthmatic response

Background


Metachromatic cells obtained from asthmatic subjects demonstrate increased spontaneous and stimulated histamine release in vitro. Their ability to synthesize and store proinflammatory cytokines has focused renewed interest on their role in asthma.


Objective: The late asthmatic response provides a useful model of clinical asthma. The aim of the study was to examine metachromatic cell derived mediators and histamine releasability in vitro after in vivo allergen exposure in atopic subjects with and without asthma and relate them to the type of physiological response observed.

Methods: Bronchoalveolar lavage (BAL) cells were obtained 4 h after challenge from asthmatics exhibiting a single early response (EAR, n = 5), a dual response (LAR, n = 7), unchallenged (basal, n = 5), atopic non-asthmatic (ANA, n = 6) and non-atopic non-asthmatics (normal, n = 5). BAL histamine and tryptase concentrations and in vitro histamine release (HR) after stimulation with anti-IgE, allergen, A23187, conconavalin A and substance P were compared.

Results:Metachromatic cell numbers were lower in normal controls compared with all asthmatic groups and in LAR compared with EAR. Metachromatic cell derived mediators were higher in asthmatic compared with normal subjects. Spontaneous HR in LAR (20.5 ± 5.0%) was lower than EAR (29.5 ± 3.9%) and ANA (30.2 ± 1.4%) (P < 0.05). No differences were seen in stimulated HR between EAR and LAR. HR in ANA stimulated with anti-IgE was greater than LAR (P < 0.05). HR in ANA stimulated with anti-IgE was greater than LAR (P < 0.05). After stimulation with ionophore A23187 (1 μM), release was greater in LAR compared with basal (P < 0.05) and no different at 5 μM. All subject groups responded to substance P (SP) but was significantly more in the asthmatic subjects compared to normal controls (P < 0.05). Allergen challenge did not modify the response of asthmatic subjects to SP.

Conclusion: Functional differences in metachromatic cell reactivity are present in atopic subjects 4 h after in vivo allergen exposure which relate to the physiological response observed after this time and suggest that there is ongoing metachromatic cell degranulation subjects who subsequently develop LAR.