SUMMARY1. Substance P (6-25-25 p-mole) produced dose-dependent flare and wheal responses when injected intradermally into the volar surface of the human forearm.2. The maximum flare response was obtained within the first 3 min of injection and declined thereafter. The wheal response reached a maximum after 12 min following the injection.3. Only those peptides having one or more basic residues in the N-terminal region were effective in producing a flare reaction. Eledoisin-related peptide and SP1 , were 17 and 7 times less active than substance P respectively, whilst [D-pro2, D-phe7, D-trp9]SP1 -1 was twice as active. The N-terminal tetrapeptide, SP1 4 and eledoisin were inactive in the dose range tested.4. Wheal-producing activity was not dependent on the presence of basic residues and the rank order of relative potencies was: physalaemin ( response in the dose range tested.5. Substance P was approximately equi-active with poly-L-arginine in the production of wheal and flare and both of these agents were about 10 times more potent than histamine. Adenosine triphosphate (25-400 n-mole) produced dose-dependent wheal and flare responses and was 10,000 times less potent than substance P. Pre-treatment of the subjects with the H, histamine antagonist, chlorpheniramine, (20 mg i.v.) reduced the wheal and flare responses to substance P.6. Local anaesthetic injection into the skin reduced the spread of the flare response but did not affect the development of the wheal response.7. Pre-treatment of the skin with capsaicin reduced the flare but not the wheal response to intradermal injection of histamine.8. The results are discussed in relation to the mechanism of the 'axon reflex' vasodilatation in skin. This is thought to involve mast cells in addition to substance P-containing primary afferent neurones.

SUMMARY1. Substance P (SP) induces histamine release from isolated rat peritoneal mast cells at concentrations of 0-1-10 /tM.2. Inhibitors of glycolysis and oxidative phosphorylation prevent the release of histamine induced by SP. 3. Cells heated to 47 'C for 20 min release histamine when treated with an agent causing cell lysis but fail to release histamine in response to SP.4. SP does not release histamine by interacting with cell-bound IgE.5. Histamine release by SP is rapid, with more than 90 % of the response occurring within 1 min of the addition of the peptide to mast cells at 37 'C. 6. Substance P, unlike antigen-antibody or compound 48/80, does not show enhanced release of histamine when calcium (0 1-1 mM) is present in the extracellular medium but calcium increases the response to SP when the ion is added after the peptide. Extracellular calcium (01-1 mM), magnesium (1-10 mM) and cobalt (0-01-0-1 mM) all inhibit SP-induced histamine release when added before the peptide.Pre-treatment of the cells with EDTA (10 mM) and washing in calcium-free medium inhibits the histamine release induced by SP.7. Histamine release induced by SP was optimum at an extracellular pH of 7-2. 8. A number of peptides structurally related to SP were examined for histaminereleasing activity. At the concentrations tested, the N-terminal dipeptides Lys-Pro and Arg-Pro, tuftsin, physalaemin, eledoisin, SP31,, Phe7]-SP6 -1 were all found to be inactive. The relative activities of the other peptides were:

In many image and video collections, we have access only to partially labeled data. For example, personal photo collections often contain several faces per image and a caption that only specifies who is in the picture, but not which name matches which face. Similarly, movie screenplays can tell us who is in the scene, but not when and where they are on the screen. We formulate the learning problem in this setting as partially-supervised multiclass classification where each instance is labeled ambiguously with more than one label. We show theoretically that effective learning is possible under reasonable assumptions even when all the data is weakly labeled. Motivated by the analysis, we propose a general convex learning formulation based on minimization of a surrogate loss appropriate for the ambiguous label setting. We apply our framework to identifying faces culled from web news sources and to naming characters in TV series and movies. We experiment on a very large dataset consisting of 100 hours of video, and in particular achieve 6% error for character naming on 16 episodes of LOST.

BackgroundWhile physical activity has been shown to improve cognitive performance and well-being, office workers are essentially sedentary. We compared the effects of physical activity performed as (i) one bout in the morning or (ii) as microbouts spread out across the day to (iii) a day spent sitting, on mood and energy levels and cognitive function.MethodsIn a randomized crossover trial, 30 sedentary adults completed each of three conditions: 6 h of uninterrupted sitting (SIT), SIT plus 30 min of moderate-intensity treadmill walking in the morning (ONE), and SIT plus six hourly 5-min microbouts of moderate-intensity treadmill walking (MICRO). Self-perceived energy, mood, and appetite were assessed with visual analog scales. Vigor and fatigue were assessed with the Profile of Mood State questionnaire. Cognitive function was measured using a flanker task and the Comprehensive Trail Making Test. Intervention effects were tested using linear mixed models.ResultsBoth ONE and MICRO increased self-perceived energy and vigor compared to SIT (p < 0.05 for all). MICRO, but not ONE, improved mood, decreased levels of fatigue and reduced food cravings at the end of the day compared to SIT (p < 0.05 for all). Cognitive function was not significantly affected by condition.ConclusionsIn addition to the beneficial impact of physical activity on levels of energy and vigor, spreading out physical activity throughout the day improved mood, decreased feelings of fatigue and affected appetite. Introducing short bouts of activity during the workday of sedentary office workers is a promising approach to improve overall well-being at work without negatively impacting cognitive performance.Trial registration NCT02717377, registered 22 March 2016.

Pictorial structure (PS) models are extensively used for part-based recognition of scenes, people, animals and multi-part objects. To achieve tractability, the structure and parameterization of the model is often restricted, for example, by assuming tree dependency structure and unimodal, data-independent pairwise interactions. These expressivity restrictions fail to capture important patterns in the data. On the other hand, local methods such as nearest-neighbor classification and kernel density estimation provide nonparametric flexibility but require large amounts of data to generalize well. We propose a simple semi-parametric approach that combines the tractability of pictorial structure inference with the flexibility of non-parametric methods by expressing a subset of model parameters as kernel regression estimates from a learned sparse set of exemplars. This yields query-specific, image-dependent pose priors. We develop an effective shape-based kernel for upper-body pose similarity and propose a leave-one-out loss function for learning a sparse subset of exemplars for kernel regression. We apply our techniques to two challenging datasets of human figure parsing and advance the state-of-the-art (from 80% to 86% on the Buffy dataset [8]), while using only 15% of the training data as exemplars. DisciplinesComputer Sciences AbstractPictorial structure (PS) models are extensively used for part-based recognition of scenes, people, animals and multi-part objects. To achieve tractability, the structure and parameterization of the model is often restricted, for example, by assuming tree dependency structure and unimodal, data-independent pairwise interactions. These expressivity restrictions fail to capture important patterns in the data. On the other hand, local methods such as nearest-neighbor classification and kernel density estimation provide nonparametric flexibility but require large amounts of data to generalize well. We propose a simple semi-parametric approach that combines the tractability of pictorial structure inference with the flexibility of non-parametric methods by expressing a subset of model parameters as kernel regression estimates from a learned sparse set of exemplars. This yields query-specific, image-dependent pose priors. We develop an effective shape-based kernel for upper-body pose similarity and propose a leave-one-out loss function for learning a sparse subset of exemplars for kernel regression. We apply our techniques to two challenging datasets of human figure parsing and advance the state-of-the-art (from 80% to 86% on the Buffy dataset [8]), while using only 15% of the training data as exemplars.

Movies and TV are a rich source of diverse and complex video of people, objects, actions and locales "in the wild". Harvesting automatically labeled sequences of actions from video would enable creation of large-scale and highly-varied datasets. To enable such collection, we focus on the task of recovering scene structure in movies and TV series for object tracking and action retrieval. We present a weakly supervised algorithm that uses the screenplay and closed captions to parse a movie into a hierarchy of shots and scenes. Scene boundaries in the movie are aligned with screenplay scene labels and shots are reordered into a sequence of long continuous tracks or threads which allow for more accurate tracking of people, actions and objects. Scene segmentation, alignment, and shot threading are formulated as inference in a unified generative model and a novel hierarchical dynamic programming algorithm that can handle alignment and jump-limited reorderings in linear time is presented. We present quantitative and qualitative results on movie alignment and parsing, and use the recovered structure to improve character naming and retrieval of common actions in several episodes of popular TV series. Abstract. Movies and TV are a rich source of diverse and complex video of people, objects, actions and locales "in the wild". Harvesting automatically labeled sequences of actions from video would enable creation of large-scale and highlyvaried datasets. To enable such collection, we focus on the task of recovering scene structure in movies and TV series for object tracking and action retrieval. We present a weakly supervised algorithm that uses the screenplay and closed captions to parse a movie into a hierarchy of shots and scenes. Scene boundaries in the movie are aligned with screenplay scene labels and shots are reordered into a sequence of long continuous tracks or threads which allow for more accurate tracking of people, actions and objects. Scene segmentation, alignment, and shot threading are formulated as inference in a unified generative model and a novel hierarchical dynamic programming algorithm that can handle alignment and jump-limited reorderings in linear time is presented. We present quantitative and qualitative results on movie alignment and parsing, and use the recovered structure to improve character naming and retrieval of common actions in several episodes of popular TV series.

No abstract

ISRCTN00061628.