Dynamic multilayer growth: Parallel vs. sequential approaches.

The decision of when to add a new hidden unit or layer is a fundamental challenge for constructive algorithms. It becomes even more complex in the context of multiple hidden layers. Growing both network width and depth offers a robust framework for leveraging the ability to capture more information from the data and model more complex representations. In the context of multiple hidden layers, should growing units occur sequentially with hidden units only being grown in one layer at a time or in parallel with hidden units growing across multiple layers simultaneously? The effects of growing sequentially or in parallel are investigated using a population dynamics-inspired growing algorithm in a multilayer context. A modified version of the constructive growing algorithm capable of growing in parallel is presented. Sequential and parallel growth methodologies are compared in a three-hidden layer multilayer perceptron on several benchmark classification tasks. Several variants of these approaches are developed for a more in-depth comparison based on the type of hidden layer initialization and the weight update methods employed. Comparisons are then made to another sequential growing approach, Dynamic Node Creation. Growing hidden layers in parallel resulted in comparable or higher performances than sequential approaches. Growing hidden layers in parallel promotes growing narrower deep architectures tailored to the task. Dynamic growth inspired by population dynamics offers the potential to grow the width and depth of deeper neural networks in either a sequential or parallel fashion.

Embodied working memory during ongoing input streams.

Sensory stimuli endow animals with the ability to generate an internal representation. This representation can be maintained for a certain duration in the absence of previously elicited inputs. The reliance on an internal representation rather than purely on the basis of external stimuli is a hallmark feature of higher-order functions such as working memory. Patterns of neural activity produced in response to sensory inputs can continue long after the disappearance of previous inputs. Experimental and theoretical studies have largely invested in understanding how animals faithfully maintain sensory representations during ongoing reverberations of neural activity. However, these studies have focused on preassigned protocols of stimulus presentation, leaving out by default the possibility of exploring how the content of working memory interacts with ongoing input streams. Here, we study working memory using a network of spiking neurons with dynamic synapses subject to short-term and long-term synaptic plasticity. The formal model is embodied in a physical robot as a companion approach under which neuronal activity is directly linked to motor output. The artificial agent is used as a methodological tool for studying the formation of working memory capacity. To this end, we devise a keyboard listening framework to delineate the context under which working memory content is (1) refined, (2) overwritten or (3) resisted by ongoing new input streams. Ultimately, this study takes a neurorobotic perspective to resurface the long-standing implication of working memory in flexible cognition.

Discrimination of Motion Direction in a Robot Using a Phenomenological Model of Synaptic Plasticity.

Recognizing and tracking the direction of moving stimuli is crucial to the control of much animal behaviour. In this study, we examine whether a bio-inspired model of synaptic plasticity implemented in a robotic agent may allow the discrimination of motion direction of real-world stimuli. Starting with a well-established model of short-term synaptic plasticity (STP), we develop a microcircuit motif of spiking neurons capable of exhibiting preferential and nonpreferential responses to changes in the direction of an orientation stimulus in motion. While the robotic agent processes sensory inputs, the STP mechanism introduces direction-dependent changes in the synaptic connections of the microcircuit, resulting in a population of units that exhibit a typical cortical response property observed in primary visual cortex (V1), namely, direction selectivity. Visually evoked responses from the model are then compared to those observed in multielectrode recordings from V1 in anesthetized macaque monkeys, while sinusoidal gratings are displayed on a screen. Overall, the model highlights the role of STP as a complementary mechanism in explaining the direction selectivity and applies these insights in a physical robot as a method for validating important response characteristics observed in experimental data from V1, namely, direction selectivity.

Rapid Homogeneous Immunoassay to Quantify Gemcitabine in Plasma for Therapeutic Drug Monitoring.

BACKGROUND: Gemcitabine (2',2'-difluoro-2'-deoxycytidine) is a nucleoside analog used as a single agent and in combination regimens for the treatment of a variety of solid tumors. Several studies have shown a relationship between gemcitabine peak plasma concentration (Cmax) and hematological toxicity. An immunoassay for gemcitabine in plasma was developed and validated to facilitate therapeutic drug monitoring (TDM) by providing an economical, robust method for automated chemistry analyzers. METHODS: A monoclonal antibody was coated on nanoparticles to develop a homogenous agglutination inhibition assay. To prevent ex vivo degradation of gemcitabine in blood, tetrahydrouridine was used as a sample stabilizer. Validation was conducted for precision, recovery, cross-reactivity, and linearity on a Beckman Coulter AU480. Verification was performed on an AU5800 in a hospital laboratory. A method comparison was performed with (LC-MS/MS) liquid chromatography tandem mass spectrometry using clinical samples. Selectivity was demonstrated by testing cross-reactivity of the major metabolite, 2',2'-difluorodeoxyuridine. RESULTS: Coefficients of variation for repeatability and within-laboratory precision were <8%. The deviation between measured and assigned values was <3%. Linear range was from 0.40 to 33.02 µ/mL (1.5-125.5 µM). Correlation with validated LC-MS/MS methods was R = 0.977. The assay was specific for gemcitabine: there was no cross-reactivity to 2',2'-difluorodeoxyuridine, chemotherapeutics, concomitant, or common medications tested. Tetrahydrouridine was packaged in single-use syringes. Gemcitabine stability in whole blood was extended to 8 hours (at room temperature) and in plasma to 8 days (2-8°C). CONCLUSIONS: The assay demonstrated the selectivity, test range, precision, and linearity to perform reliable measurements of gemcitabine in plasma. The addition of stabilizer improved the sample handling. Using general clinical chemistry analyzers, gemcitabine could be measured for TDM.

Evaluation of eicosanoid concentrations in stored units of canine packed red blood cells.

OBJECTIVE To evaluate eicosanoid concentrations in freshly prepared canine packed RBCs (PRBCs) and to assess changes in eicosanoid concentrations in PRBC units over time during storage and under transfusion conditions. DESIGN Prospective study. SAMPLE 25 plasma samples from 14 healthy Greyhounds. PROCEDURES Plasma samples were obtained during PRBC preparation (donation samples), and the PRBC units were then stored at 4°C until used for transfusion (≤ 21 days later; n = 17) or mock transfusion if expired (22 to 24 days later; 8). Immediately prior to use, 100 mL of saline (0.9% NaCl) solution was added to each unit and a pretransfusion sample was collected. A posttransfusion sample was collected after transfusion or mock transfusion. Concentrations of arachidonic acid, prostaglandin (PG) F2α, PGE2, PGD2, thromboxane B2, 6-keto-PGF1α, and leukotriene B4 were measured by liquid chromatography-mass spectrometry and analyzed statistically. RESULTS Median arachidonic acid concentration was significantly decreased in posttransfusion samples, compared with the concentration in donation samples. Median PGF2α, 6-keto-PGF1α, and leukotriene B4 concentrations were significantly increased in pretransfusion samples, compared with those in donation samples. Median PGF2α, thromboxane B2, and 6-keto-PGF1α concentrations were significantly increased in posttransfusion samples, compared with those in pretransfusion samples. Duration of PRBC storage had significant associations with pretransfusion and posttransfusion arachidonic acid and thromboxane B2 concentrations. CONCLUSIONS AND CLINICAL RELEVANCE Concentrations of several proinflammatory eicosanoids increased in PRBC units during storage, transfusion, or both. Accumulation of these products could potentially contribute to adverse transfusion reactions, and investigation of the potential association between eicosanoid concentrations in PRBCs and the incidence of transfusion reactions in dogs is warranted.

Extreme ionization leading to Coulomb explosion of small palladium and zirconium oxide clusters and reactivity with carbon monoxide.

Reported herein are strong-field ionization studies of small, neutral Pd(x)O(y) and Zr(x)O(y) clusters made using ultrafast laser pulses (~100 fs) centered at 624 nm. An enhancement in ionization of nearly 1.5 orders of magnitude lower in laser intensity than predicted from literature values is observed for both systems due to clustering. The change in enhancement upon addition of carbon monoxide at different pressures was also studied. Enhancement of high charge states of palladium was found to decrease upon CO addition, whereas in the case of the zirconium system, high charge states of zirconium were observed to increase. Pd and ZrO showed similar reactivity trends with CO and were found to have similar reactivity ratios in accord with their isovalent nature.

Influence of group 10 metals on the growth and subsequent Coulomb explosion of small silicon clusters under strong light pulses.

Growth and ionization patterns of small silicon clusters are studied using ultrafast pulses centered at 624 nm by varying the metal electron source for cluster formation using group 10 transition metals. The silicon-cluster size was observed to change as the electron source was varied from Pd

Ionization enhancement in silicon clusters and germanium atoms in the presence of zirconium.

Molecules/clusters have been shown to undergo an enhancement in ionization under ultrafast laser pulses. This enhancement results in the lowering of the laser intensity required to observe ion signal from higher atomic charge states resulting from Coulomb explosion of clusters. Here, we explore the effect of using an early-group transition metal as an electron source in the formation of small silicon clusters on the observed enhancement in ionization. Intensity selective scanning is used to measure the onset of ion signal for the atomic charge states of silicon, germanium, zirconium, and oxygen. Additionally, the kinetic energy released values for the resulting high charge states of silicon are measured and compared to those previously observed using a copper electron source. A significant increase in ionization enhancement is observed upon using zirconium metal, despite a decrease in cluster size. Germanium metal with zirconium is studied for comparison and shows a larger enhancement in ion signal than silicon, indicating that atomic mass may be significant.

Ionization and Coulomb explosion of small group 10 transition metal oxide clusters in strong light fields.

The ionization properties of small group 10 metal oxide clusters are explored using ultrafast pulses centered at 624 nm. Maximum atomic charge states resulting from Coulomb explosion were observed to be Ni(3+), Pd(3+), Pt(5+), and O(2+) species with similar ionization potentials ~30-35 eV. Ion signal as a function of laser intensity of each charge state of Ni, Pd, Pt, and O resulting from Coulomb explosion was mapped and compared to that predicted from semi-classical tunneling theory using sequential ionization potentials to quantify observed enhancements in ionization. The saturation intensity (I(sat)) of each charge state is measured and compared to previous studies on group 5 transition metal oxides. The atomic charge states of nickel showed a large enhancement in ionization compared to palladium and platinum, reflective of the differing bonding properties of each metal with oxygen. Results indicate that nickel oxide clusters undergo a greater extent of ionization enhancement as a result of multiple ionization mechanisms. The ionization enhancement behavior of each metal oxide species is explored herein.

Strong-field ionization and coulomb explosion of chlorine weakly bound to small water clusters.

Clusters exhibit an enhancement in ionization rates under intense, ultrafast laser pulses compared to their molecular/atomic counterparts. Studies of ionization enhancement of weakly bound molecules to clusters have not been previously characterized and quantified. We demonstrate that weakly bound ClO to (H(2)O)(n) (n = 1-12) clusters and weakly bound HCl to (H(2)O)(n) (n = 1-12) clusters produce high atomic charge states of chlorine via Coulomb explosion. Density functional theory (DFT) was used to qualitatively compare the interaction energy of ClO with respect to the number of water molecules as well as HCl with respect to the number of water molecules. The chlorine ion signal intensity for each atomic charge state was observed to be dependent on the molecule-cluster bond strength. The observed ionization enhancement was quantified using semiclassical tunneling theory, and it was found that the Cl(3+-5+) and O(2+) charge states are enhanced in ionization. Possible mechanisms of ionization enhancement are explored for weakly bound chlorine species.

Delocalized electronic behavior observed in transition metal oxide clusters under strong-field excitation.

Heterogeneously composed clusters are exposed to intensity resolved, 100 fs laser pulses to reveal the energy requirements for the production of the high charge states of both metal and nonmetal ions. The ionization and fragmentation of group V transition metal oxide clusters are here examined with laser intensities ranging nearly four orders in magnitude (∼3 × 10(11) W/cm(2) to ∼2 × 10(15) W/cm(2)) at 624 nm. The ionization potentials of the metal atoms are measured using both multiphoton ionization and tunneling ionization models. We demonstrate that the intensity selective scanning method can be utilized to measure the low ionization potentials of transition metals (∼7 eV). The high charge states demonstrate an enhancement in ionization that is three orders of magnitude lower in laser intensity than predicted for the atomic counterparts. Finally, the response from the various metals and the oxygen is compared to elucidate the mechanism of enhanced ionization that is observed. Specifically, the sequence of ion appearances demonstrates delocalized electron behavior over the entire cluster.

Influence of clustering and molecular orbital shapes on the ionization enhancement in ammonia.

The Coulomb explosion of clusters is known to be an efficient source for producing multiply charged ions through an enhanced ionization process. However, the factors responsible for obtaining these high charge states have not been previously explored in detail and remain poorly understood. By comparing intensity-resolved visible laser excitation experiments with semi-classical theory over a range spanning both multiphoton and tunneling ionization regimes, we reveal the mechanism in which extreme ionization proceeds. Under laser conditions that can only singly ionize individual molecules, ammonia clusters generate ions depleted of all valence electrons. The geometries of the molecular orbitals are revealed to be important in driving the ionization, and can be entirely emptied at the energy requirement for removal of the first electron in the orbital. The results are in accord with non-sequential ionization arising from electrons tunneling from three separate molecular orbitals aided through the ionization ignition mechanism.

Dual potency anti-HER2/neu and anti-EGFR anthracycline immunoconjugates in chemotherapeutic-resistant mammary carcinoma combined with cyclosporin A and verapamil P-glycoprotein inhibition.

Immunoconjugates of epirubicin were synthesized with monoclonal antibodies against the epidermal growth factor receptors, HER2/neu and EGFR, by creating a sulfhydryl-reactive epirubicin intermediate applying heterobifunctional succinimidyl-4-(N-maleimidomethyl)cyclohexane-1-carboxylate (SMCC), which was introduced at alpha-monoamide groups of the epirubicin carbohydrate moiety. In parallel, N-succinimidyl-S-acetylthioacetate (SATA) was used to incorporate a sulfhydryl group into immunoglobulin at the terminal amine position of -lysine amino acid residues. Eprirubicin-SMCC-SATA-IgG immunoconjugates were produced by reacting epirubicin-SMCC and SATA-IgG at appropriate molar ratios. Epirubicin-(anti-HER2/neu) and epirubicin-(anti-EGFR) had greater potency against chemotherapeutic-resistant SKBr-3 mammary carcinoma than did epirubicin at epirubicin-equivalent concentrations. Epirubicin-(anti-HER2/neu) was more potent than epirubicin-(anti-EGFR), and a synergistic level of antineoplastic activity was detected with an epirubicin immunoconjugate 50/50 combination. Competitive P-glycoprotein inhibition with cyclosporin A or verapamil enhanced the potency of the epirubicin immunoconjugate 50/50 combination. Minor levels of antineoplastic activity were detected only with an immunoglobulin 50/50 combination of anti-HER2/neu and anti-EGFR. The investigations represent a potential strategy for enhancing the selective internalization, intracellular deposition, and antineoplastic potency of chemotherapeutics in multidrug-resistant neoplasias.