A Neural Network Architecture for Autonomous Learning, Recognition, and Prediction in a Nonstationary World

In a constantly changing world, humans are adapted to alternate routinely between attending to familiar objects and testing hypotheses about novel ones. We can rapidly learn to recognize and narne novel objects without unselectively disrupting our memories of familiar ones. We can notice fine details that differentiate nearly identical objects and generalize across broad classes of dissimilar objects. This chapter describes a class of self-organizing neural network architectures--called ARTMAP-- that are capable of fast, yet stable, on-line recognition learning, hypothesis testing, and naming in response to an arbitrary stream of input patterns (Carpenter, Grossberg, Markuzon, Reynolds, and Rosen, 1992; Carpenter, Grossberg, and Reynolds, 1991). The intrinsic stability of ARTMAP allows the system to learn incrementally for an unlimited period of time. System stability properties can be traced to the structure of its learned memories, which encode clusters of attended features into its recognition categories, rather than slow averages of category inputs. The level of detail in the learned attentional focus is determined moment-by-moment, depending on predictive success: an error due to over-generalization automatically focuses attention on additional input details enough of which are learned in a new recognition category so that the predictive error will not be repeated. An ARTMAP system creates an evolving map between a variable number of learned categories that compress one feature space (e.g., visual features) to learned categories of another feature space (e.g., auditory features). Input vectors can be either binary or analog. Computational properties of the networks enable them to perform significantly better in benchmark studies than alternative machine learning, genetic algorithm, or neural network models. Some of the critical problems that challenge and constrain any such autonomous learning system will next be illustrated. Design principles that work together to solve these problems are then outlined. These principles are realized in the ARTMAP architecture, which is specified as an algorithm. Finally, ARTMAP dynamics are illustrated by means of a series of benchmark simulations.

Gene expression signatures of radiation response are specific, durable and accurate in mice and humans.

BACKGROUND: Previous work has demonstrated the potential for peripheral blood (PB) gene expression profiling for the detection of disease or environmental exposures. METHODS AND FINDINGS: We have sought to determine the impact of several variables on the PB gene expression profile of an environmental exposure, ionizing radiation, and to determine the specificity of the PB signature of radiation versus other genotoxic stresses. Neither genotype differences nor the time of PB sampling caused any lessening of the accuracy of PB signatures to predict radiation exposure, but sex difference did influence the accuracy of the prediction of radiation exposure at the lowest level (50 cGy). A PB signature of sepsis was also generated and both the PB signature of radiation and the PB signature of sepsis were found to be 100% specific at distinguishing irradiated from septic animals. We also identified human PB signatures of radiation exposure and chemotherapy treatment which distinguished irradiated patients and chemotherapy-treated individuals within a heterogeneous population with accuracies of 90% and 81%, respectively. CONCLUSIONS: We conclude that PB gene expression profiles can be identified in mice and humans that are accurate in predicting medical conditions, are specific to each condition and remain highly accurate over time.

Low-loss directional cloaks without superluminal velocity or magnetic response.

The possibility of making an optically large (many wavelengths in diameter) object appear invisible has been a subject of many recent studies. Exact invisibility scenarios for large (relative to the wavelength) objects involve (meta)materials with superluminal phase velocity [refractive index (RI) less than unity] and/or magnetic response. We introduce a new approximation applicable to certain device geometries in the eikonal limit: piecewise-uniform scaling of the RI. This transformation preserves the ray trajectories but leads to a uniform phase delay. We show how to take advantage of phase delays to achieve a limited (directional and wavelength-dependent) form of invisibility that does not require loss-ridden (meta)materials with superluminal phase velocities.

An evaluation of remifentanil-sevoflurane response surface models in patients emerging from anesthesia: model improvement using effect-site sevoflurane concentrations.

INTRODUCTION: We previously reported models that characterized the synergistic interaction between remifentanil and sevoflurane in blunting responses to verbal and painful stimuli. This preliminary study evaluated the ability of these models to predict a return of responsiveness during emergence from anesthesia and a response to tibial pressure when patients required analgesics in the recovery room. We hypothesized that model predictions would be consistent with observed responses. We also hypothesized that under non-steady-state conditions, accounting for the lag time between sevoflurane effect-site concentration (Ce) and end-tidal (ET) concentration would improve predictions. METHODS: Twenty patients received a sevoflurane, remifentanil, and fentanyl anesthetic. Two model predictions of responsiveness were recorded at emergence: an ET-based and a Ce-based prediction. Similarly, 2 predictions of a response to noxious stimuli were recorded when patients first required analgesics in the recovery room. Model predictions were compared with observations with graphical and temporal analyses. RESULTS: While patients were anesthetized, model predictions indicated a high likelihood that patients would be unresponsive (> or = 99%). However, after termination of the anesthetic, models exhibited a wide range of predictions at emergence (1%-97%). Although wide, the Ce-based predictions of responsiveness were better distributed over a percentage ranking of observations than the ET-based predictions. For the ET-based model, 45% of the patients awoke within 2 min of the 50% model predicted probability of unresponsiveness and 65% awoke within 4 min. For the Ce-based model, 45% of the patients awoke within 1 min of the 50% model predicted probability of unresponsiveness and 85% awoke within 3.2 min. Predictions of a response to a painful stimulus in the recovery room were similar for the Ce- and ET-based models. DISCUSSION: Results confirmed, in part, our study hypothesis; accounting for the lag time between Ce and ET sevoflurane concentrations improved model predictions of responsiveness but had no effect on predicting a response to a noxious stimulus in the recovery room. These models may be useful in predicting events of clinical interest but large-scale evaluations with numerous patients are needed to better characterize model performance.

Inhibition-Induced Forgetting Results from Resource Competition between Response Inhibition and Memory Encoding Processes.

UNLABELLED: Response inhibition is a key component of executive control, but its relation to other cognitive processes is not well understood. We recently documented the "inhibition-induced forgetting effect": no-go cues are remembered more poorly than go cues. We attributed this effect to central-resource competition, whereby response inhibition saps attention away from memory encoding. However, this proposal is difficult to test with behavioral means alone. We therefore used fMRI in humans to test two neural predictions of the "common resource hypothesis": (1) brain regions associated with response inhibition should exhibit greater resource demands during encoding of subsequently forgotten than remembered no-go cues; and (2) this higher inhibitory resource demand should lead to memory encoding regions having less resources available during encoding of subsequently forgotten no-go cues. Participants categorized face stimuli by gender in a go/no-go task and, following a delay, performed a surprise recognition memory test for those faces. Replicating previous findings, memory was worse for no-go than for go stimuli. Crucially, forgetting of no-go cues was predicted by high inhibitory resource demand, as quantified by the trial-by-trial ratio of activity in neural "no-go" versus "go" networks. Moreover, this index of inhibitory demand exhibited an inverse trial-by-trial relationship with activity in brain regions responsible for the encoding of no-go cues into memory, notably the ventrolateral prefrontal cortex. This seesaw pattern between the neural resource demand of response inhibition and activity related to memory encoding directly supports the hypothesis that response inhibition temporarily saps attentional resources away from stimulus processing. SIGNIFICANCE STATEMENT: Recent behavioral experiments showed that inhibiting a motor response to a stimulus (a "no-go cue") impairs subsequent memory for that cue. Here, we used fMRI to test whether this "inhibition-induced forgetting effect" is caused by competition for neural resources between the processes of response inhibition and memory encoding. We found that trial-by-trial variations in neural inhibitory resource demand predicted subsequent forgetting of no-go cues and that higher inhibitory demand was furthermore associated with lower concurrent activation in brain regions responsible for successful memory encoding of no-go cues. Thus, motor inhibition and stimulus encoding appear to compete with each other: when more resources have to be devoted to inhibiting action, less are available for encoding sensory stimuli.

Inter-annual variability of precipitation constrains the production response of boreal Pinus sylvestris to nitrogen fertilization

© 2015 Published by Elsevier B.V.Tree growth resources and the efficiency of resource-use for biomass production determine the productivity of forest ecosystems. In nutrient-limited forests, nitrogen (N)-fertilization increases foliage [N], which may increase photosynthetic rates, leaf area index (L), and thus light interception (IC). The product of such changes is a higher gross primary production and higher net primary production (NPP). However, fertilization may also alter carbohydrate partitioning from below- to aboveground, increasing aboveground NPP (ANPP). We analyzed effects of long-term N-fertilization on NPP, and that of long-term carbon storing organs (NPPS) in a Pinus sylvestris forest on sandy soil, a wide-ranging forest type in the boreal region. We based our analyses on a combination of destructive harvesting, consecutive mensuration, and optical measurements of canopy openness. After eight-year fertilization with a total of 70gNm^-2, ANPP was 27±7% higher in the fertilized (F) relative to the reference (R) stand, but although L increased relative to its pre-fertilization values, IC was not greater than in R. On the seventh year after the treatment initiation, the increase of ANPP was matched by the decrease of belowground NPP (78 vs. 92gCm^-2yr^-1; ~17% of NPP) and, given the similarity of IC, suggests that the main effect of N-fertilization was changed carbon partitioning rather than increased canopy photosynthesis. Annual NPPS increased linearly with growing season temperature (T) in both treatments, with an upward shift of 70.2gCm^-2yr^-1 by fertilization, which also caused greater amount of unexplained variation (r²=0.53 in R, 0.21 in F). Residuals of the NPPS-T relationship of F were related to growing season precipitation (P, r²=0.48), indicating that T constrains productivity at this site regardless of fertility, while P is important in determining productivity where N-limitation is alleviated. We estimated that, in a growing season average T (11.5±1.0°C; 33-year-mean), NPPS response to N-fertilization will be nullified with P 31mm less than the mean (325±85mm), and would double with P 109mm greater than the mean. These results suggest that inter-annual variation in climate, particularly in P, may help explaining the reported large variability in growth responses to fertilization of pine stands on sandy soils. Furthermore, forest management of long-rotation systems, such as those of boreal and northern temperate forests, must consider the efficiency of fertilization in terms of wood production in the context of changes in climate predicted for the region.

Validation of the Food-Linked Virtual Response task.

This research validates a computerized dietary selection task (Food-Linked Virtual Response or FLVR) for use in studies of food consumption. In two studies, FLVR task responses were compared with measures of health consciousness, mood, body mass index, personality, cognitive restraint toward food, and actual food selections from a buffet table. The FLVR task was associated with variables which typically predict healthy decision-making and was unrelated to mood or body mass index. Furthermore, the FLVR task predicted participants' unhealthy selections from the buffet, but not overall amount of food. The FLVR task is an inexpensive, valid, and easily administered option for assessing momentary dietary decisions.

Optical-CT 3D Dosimetry Using Fresnel Lenses with Minimal Refractive-Index Matching Fluid.

Telecentric optical computed tomography (optical-CT) is a state-of-the-art method for visualizing and quantifying 3-dimensional dose distributions in radiochromic dosimeters. In this work a prototype telecentric system (DFOS-Duke Fresnel Optical-CT Scanner) is evaluated which incorporates two substantial design changes: the use of Fresnel lenses (reducing lens costs from $10-30K t0 $1-3K) and the use of a 'solid tank' (which reduces noise, and the volume of refractively matched fluid from 1 ltr to 10 cc). The efficacy of DFOS was evaluated by direct comparison against commissioned scanners in our lab. Measured dose distributions from all systems were compared against the predicted dose distributions from a commissioned treatment planning system (TPS). Three treatment plans were investigated including a simple four-field box treatment, a multiple small field delivery, and a complex IMRT treatment. Dosimeters were imaged within 2 h post irradiation, using consistent scanning techniques (360 projections acquired at 1 degree intervals, reconstruction at 2mm). DFOS efficacy was evaluated through inspection of dose line-profiles, and 2D and 3D dose and gamma maps. DFOS/TPS gamma pass rates with 3%/3mm dose difference/distance-to-agreement criteria ranged from 89.3% to 92.2%, compared to from 95.6% to 99.0% obtained with the commissioned system. The 3D gamma pass rate between the commissioned system and DFOS was 98.2%. The typical noise rates in DFOS reconstructions were up to 3%, compared to under 2% for the commissioned system. In conclusion, while the introduction of a solid tank proved advantageous with regards to cost and convenience, further work is required to improve the image quality and dose reconstruction accuracy of the new DFOS optical-CT system.

Is observed variability in the long-term results of the Continuous Plankton Recorder survey a response to climate change?

In the more than 50 years that the Continuous Plankton Recorder (CPR) survey has operated on a regular monthly basis in the north-east Atlantic and North Sea, large changes have been witnessed in the planktonic ecosystem. These changes have taken the form of long-term trends in abundance for certain species or stepwise changes for others, and in many cases are correlated with a mode of climatic variability in the North Atlantic, either: (1) the North Atlantic Oscillation (NAO), a basin-scale atmospheric alteration of the pressure field between the Azores high pressure cell and the Icelandic Low; or (2) the Gulf Stream Index (GSI), which measures the latitudinal position of the north wall of the Gulf Stream. Recent work has shown that the changes in the GSI are coupled with the NAO and Pacific Southern Oscillation with a 2 year lag. The plankton variability is also possibly linked to changes observed in the distribution and flux of water masses in the surface, intermediate and deep waters of the North Atlantic. For example, in the last two decades, the extent and location of the formation of North Atlantic Deep Water, Labrador Sea Intermediate Water and Norwegian Sea intermediate and upper-layer water has altered considerably. This paper discusses the extent to which observed changes in plankton abundance and distribution may be linked to this basin-scale variability in hydrodynamics. The results are also placed within the context of global climate warming and the possible effects of the observed melting of Arctic permafrost and sea ice on the subpolar North Atlantic.

Coccolithophore bloom size variation in response to the regional environment of the subarctic North Atlantic

Several environmental/physical variables derived from satellite and in situ data sets were used to understand the variability of coccolithophore abundance in the subarctic North Atlantic. The 7-yr (1997–2004) time-series analysis showed that the combined effects of high solar radiation, shallow mixed layer depth (<20 m), and increased temperatures explained >89% of the coccolithophore variation. The June 1998 bloom, which was associated with high light intensity, unusually high sea-surface temperature, and a very shallow mixed layer, was found to be one of the most extensive (>995,000 km2) blooms ever recorded. There was a pronounced sea-surface temperature shift in the mid-1990s with a peak in 1998, suggesting that exceptionally large blooms are caused by pronounced environmental conditions and the variability of the physical environment strongly affects the spatial extent of these blooms. Consequently, if the physical environment varies, the effects of these blooms on the atmospheric and oceanic environment will vary as well.

The response of phytoplankton to climate variability associated with the North Atlantic Oscillation

Inter-annual variability in the timing of phytoplankton spring bloom and phytoplankton community structure in the central North Atlantic Ocean was quantified using ocean color data and continuous plankton recorder (CPR) data. This variability was related to the North Atlantic Oscillation using correlation analysis and multivariate auto-regression models. The initiation of the spring bloom derived from CPR phytoplankton color index data is similar to that derived from satellite chlorophyll, and exhibits a nominal correlation with the sea surface temperature (SST) and the North Atlantic Oscillation (NAO). The extrapolated spring bloom timing suggested later initiation of blooms in the mid-1980s and earlier initiation of blooms in the 1990s. The climatological phytoplankton community structure in the central North Atlantic is dominated by diatoms, except for a shift in community composition favoring dinoflagellates in August. The ratio of diatoms to total phytoplankton abundance and the ratio of dinoflagellates to total phytoplankton abundance are both closely correlated with the NAO and SST. The extended time series of phytoplankton community structure between 1985 and 2009, deduced from the time series of SST and NAO over the same interval, showed a decadal shift away from diatoms towards dinoflagellates. The linkages between the NAO, and changes in stratification and phytoplankton processes occur over a larger scale than previously observed.

The meso-scale response of subarctic North Pacific seabird community structure to lower trophic level abundance and diversity

Understanding the mechanisms that structure communities and influence biodiversity are fundamental goals of ecology. To test the hypothesis that the abundance and diversity of upper-trophic level predators (seabirds) is related to the underlying abundance and diversity of their prey (zooplankton) and ecosystem-wide energy availability (primary production), we initiated a monitoring program in 2002 that jointly and repeatedly surveys seabird and zooplankton populations across a 7,500 km British Columbia-Bering Sea-Japan transect. Seabird distributions were recorded by a single observer (MH) using a strip-width technique, mesozooplankton samples were collected with a Continuous Plankton Recorder, and primary production levels were derived using the appropriate satellite parameters and the Vertically Generalized Production Model (Behrenfeld and Falkowski 1997). Each trophic level showed clear spatio-temporal patterns over the course of the study. The strongest relationship between seabird abundance and diversity and the lower trophic levels was observed in March/April ('spring') and significant relationships were also found through June/July ('summer'). No discernable relationships were observed during the September/October ('fall') months. Overall, mesozooplankton abundance and biomass explained the dominant portion of seabird abundance and diversity indices (richness, Simpson's Index, and evenness), while primary production was only related to seabird richness. These findings underscore the notion that perturbations of ocean productivity and lower trophic level ecosystem constituents influenced by climate change, such as shifts in timing (phenology) and synchronicity (match-mismatch), could impart far-reaching consequences throughout the marine food web.

BRCA1 and GADD45 mediated G2/M cell cycle arrest in response to antimicrotubule agents

BRCA1 is a tumour suppressor gene implicated in the predisposition to early onset breast and ovarian cancer. We have generated cell lines with inducible expression of BRCA1 to evaluate its role in mediating the cellular response to various chemotherapeutic drugs commonly used in the treatment of breast and ovarian cancer. Induction of BRCA1 in the presence of Taxol and Vincristine resulted in a dramatic increase in cell death; an effect that was preceded by an acute arrest at the G2/M phase of the cell cycle and which correlated with BRCA1 mediated induction of GADD45. A proportion of the arrested cells were blocked in mitosis suggesting activation of both a G2 and a mitotic spindle checkpoint. In contrast, no specific interaction was observed between BRCA1 induction and treatment of cells with a range of DNA damaging agents including Cisplatin and Adriamycin. Inducible expression of GADD45 in the presence of Taxol induced both G2 and mitotic arrest in these cells consistent with a role for GADD45 in contributing to these effects. Our results support a role for both BRCA1 and GADD45 in selectively regulating a G2/M checkpoint in response to antimicrotubule agents and raise the possibility that their expression levels in cells may contribute to the toxicity observed with these compounds.

Multivariate prediction of clarified butter composition using Raman spectroscopy

Raman spectroscopy has been used to predict the abundance of the FA in clarified butterfat that was obtained from dairy cows fed a range of levels of rapeseed oil in their diet. Partial least squares regression of the Raman spectra against FA compositions obtained by GC showed good prediction for the five major (abundance >5%) FA with R-2=0.74-0.92 and a root mean SE of prediction (RMSEP) that was 5-7% of the mean. In general, the prediction accuracy fell with decreasing abundance in the sample, but the RMSEP was 1.25%. The Raman method has the best prediction ability for unsaturated FA (R-2=0.85-0.92), and in particular trans unsaturated FA (best-predicted FA was 18:1 tDelta9). This enhancement was attributed to the isolation of the unsaturated modes from the saturated modes and the significantly higher spectral response of unsaturated bonds compared with saturated bonds. Raman spectra of the melted butter samples could also be used to predict bulk parameters calculated from standard analyzes, such as iodine value (R-2=0.80) and solid fat content at low temperature (R-2=0.87). For solid fat contents determined at higher temperatures, the prediction ability was significantly reduced (R-2=0.42), and this decrease in performance was attributed to the smaller range of values in solid fat content at the higher temperatures. Finally, although the prediction errors for the abundances of each of the FA in a given sample are much larger with Raman than with full GC analysis, the accuracy is acceptably high for quality control applications. This, combined with the fact that Raman spectra can be obtained with no sample preparation and with 60-s data collection times, means that high-throughput, on-line Raman analysis of butter samples should be possible.

BRCA1 mRNA Expression Levels Predict for Overall Survival in Ovarian Cancer after Chemotherapy.

We investigated whether BRCA1 mRNA expression levels may represent a biomarker of survival in sporadic epithelial ovarian cancer following chemotherapy treatment. EXPERIMENTAL DESIGN: The effect of loss of BRCA1 expression on chemotherapy response in ovarian cancer was measured in vitro using dose inhibition assays and Annexin V flow cytometry. Univariate and multivariate analyses were done to evaluate the relationship between BRCA1 mRNA expression levels and survival after chemotherapy treatment in 70 fresh frozen ovarian tumors. RESULTS: We show that inhibition of endogenous BRCA1 expression in ovarian cancer cell lines results in increased sensitivity to platinum therapy and decreased sensitivity to antimicrotubule agents. In addition, we show that patients with low/intermediate levels of BRCA1 mRNA have a significantly improved overall survival following treatment with platinum-based chemotherapy in comparison with patients with high levels of BRCA1 mRNA (57.2 versus 18.2 months; P = 0.0017; hazard ratio, 2.9). Furthermore, overall median survival for higher-BRCA1-expressing patients was found to increase following taxane-containing chemotherapy (23.0 versus 18.2 months; P = 0.12; hazard ratio, 0.53). CONCLUSIONS: We provide evidence to support a role for BRCA1 mRNA expression as a predictive marker of survival in sporadic epithelial ovarian cancer.