Biased ART: A Neural Architecture that Shifts Attention Toward Previously Disregarded Features Following an Incorrect Prediction

Memories in Adaptive Resonance Theory (ART) networks are based on matched patterns that focus attention on those portions of bottom-up inputs that match active top-down expectations. While this learning strategy has proved successful for both brain models and applications, computational examples show that attention to early critical features may later distort memory representations during online fast learning. For supervised learning, biased ARTMAP (bARTMAP) solves the problem of over-emphasis on early critical features by directing attention away from previously attended features after the system makes a predictive error. Small-scale, hand-computed analog and binary examples illustrate key model dynamics. Twodimensional simulation examples demonstrate the evolution of bARTMAP memories as they are learned online. Benchmark simulations show that featural biasing also improves performance on large-scale examples. One example, which predicts movie genres and is based, in part, on the Netflix Prize database, was developed for this project. Both first principles and consistent performance improvements on all simulation studies suggest that featural biasing should be incorporated by default in all ARTMAP systems. Benchmark datasets and bARTMAP code are available from the CNS Technology Lab Website: http://techlab.bu.edu/bART/.

Distributed Hypothesis Testing, Attention Shifts and Transmitter Dynatmics During the Self-Organization of Brain Recognition Codes

BP (89-A-1204); Defense Advanced Research Projects Agency (90-0083); National Science Foundation (IRI-90-00530); Air Force Office of Scientific Research (90-0175, 90-0128); Army Research Office (DAAL-03-88-K0088)

Facilitation shifts paradigms and can amplify coastal Restoration efforts

Restoration has been elevated as an important strategy to reverse the decline of coastal wetlands worldwide. Current practice in restoration science emphasizes minimizing competition between outplanted propagules to maximize planting success. This paradigm persists despite the fact that foundational theory in ecology demonstrates that positive species interactions are key to organism success under high physical stress, such as recolonization of bare substrate. As evidence of how entrenched this restoration paradigm is, our survey of 25 restoration organizations in 14 states in the United States revealed that >95% of these agencies assume minimizing negative interactions (i.e., competition) between outplants will maximize propagule growth. Restoration experiments in both Western and Eastern Atlantic salt marshes demonstrate, however, that a simple change in planting configuration (placing propagules next to, rather than at a distance from, each other) results in harnessing facilitation and increased yields by 107% on average. Thus, small adjustments in restoration design may catalyze untapped positive species interactions, resulting in significantly higher restoration success with no added cost. As positive interactions between organisms commonly occur in coastal ecosystems (especially in more physically stressful areas like uncolonized substrate) and conservation resources are limited, transformation of the coastal restoration paradigm to incorporate facilitation theory may enhance conservation efforts, shoreline defense, and provisioning of ecosystem services such as fisheries production.

Exercise elicits phase shifts and acute alterations of melatonin that vary with circadian phase

To examine the immediate phase-shifting effects of high-intensity exercise of a practical duration (1 h) on human circadian phase, five groups of healthy men 20-30 yr of age participated in studies involving no exercise or exposure to morning, afternoon, evening, or nocturnal exercise. Except during scheduled sleep/dark and exercise periods, subjects remained under modified constant routine conditions allowing a sleep period and including constant posture, knowledge of clock time, and exposure to dim light intensities averaging (±SD) 42 ± 19 lx. The nocturnal onset of plasma melatonin secretion was used as a marker of circadian phase. A phase response curve was used to summarize the phase-shifting effects of exercise as a function of the timing of exercise. A significant effect of time of day on circadian phase shifts was observed (P < 0.004). Over the interval from the melatonin onset before exercise to the first onset after exercise, circadian phase was significantly advanced in the evening exercise group by 30 ± 15 min (SE) compared with the phase delays observed in the no-exercise group (-25 ± 14 min, P < 0.05). Phase shifts in response to evening exercise exposure were attenuated on the second day after exercise exposure and no longer significantly different from phase shifts observed in the absence of exercise. Unanticipated transient elevations of melatonin levels were observed in response to nocturnal exercise and in some evening exercise subjects. Taken together with the results from previous studies in humans and diurnal rodents, the current results suggest that 1) a longer duration of exercise exposure and/or repeated daily exposure to exercise may be necessary for reliable phase-shifting of the human circadian system and that 2) early evening exercise of high intensity may induce phase advances relevant for nonphotic entrainment of the human circadian system.

Regime shifts in marine ecosystems: detection prediction and management

Regime shifts are abrupt changes between contrasting, persistent states of any complex system. The potential for their prediction in the ocean and possible management depends upon the characteristics of the regime shifts: their drivers (from anthropogenic to natural), scale (from the local to the basin) and potential for management action (from adaptation to mitigation). We present a conceptual framework that will enhance our ability to detect, predict and manage regime shifts in the ocean, illustrating our approach with three well-documented examples: the North Pacific, the North Sea and Caribbean coral reefs. We conclude that the ability to adapt to, or manage, regime shifts depends upon their uniqueness, our understanding of their causes and linkages among ecosystem components and our observational capabilities.

Temporal changes in plankton of the North Sea: community shifts and environmental drivers

This paper analyses long-term and seasonal changes in the North Sea plankton community during the period 1970 to 2008. Based on Continuous Plankton Recorder (CPR) data covering 38 yr, major changes in both phytoplankton and zooplankton abundance and community structure were identified. Regime changes were detected around 1978, 1989 and 1998. The first 2 changes have been discussed in the literature and are defined as a cold episodic event (1978) and a regime shift towards a warm dynamic regime (1989). The effect of these 2 regime changes on plankton indicators was assessed and checked against previous studies. The 1998 change represents a shift in the abundance and seasonal patterns of dinoflagellates and the dominant zooplankton group, the neritic copepods. Furthermore, environmental factors such as air temperature, wind speed and the North Atlantic water inflow were identified as potential drivers of change in seasonal patterns, and the most-likely environmental causes for detected changes were assessed. We suggest that a change in the balance of dissolved nutrients driven by these environmental factors was the cause of the latest change in plankton community structure, which in turn could have affected the North Sea fish community.

Rapid biogeographical plankton shifts in the North Atlantic Ocean

Large-scale biogeographical changes in the biodiversity of a key zooplankton group (calanoid copepods) were detected in the north-eastern part of the North Atlantic Ocean and its adjacent seas over the period 1960–1999. These findings provided key empirical evidence for climate change impacts on marine ecosystems at the regional to oceanic scale. Since 1999, global temperatures have continued to rise in the region. Here, we extend the analysis to the period 1958–2005 using all calanoid copepod species assemblages (nine species assemblages based on an analysis including a total of 108 calanoid species or taxa) and show that this phenomenon has been reinforced in all regions. Our study reveals that the biodiversity of calanoid copepods are responding quickly to sea surface temperature (SST) rise by moving geographically northward at a rapid rate up to about 23.16 km yr−1. Our analysis suggests that nearly half of the increase in sea temperature in the northeast Atlantic and adjacent seas is related to global temperature rises (46.35% of the total variance of temperature) while changes in both natural modes of atmospheric and oceanic circulation explain 26.45% of the total variance of temperature. Although some SST isotherms have moved northwards by an average rate of up to 21.75 km yr−1 (e.g. the North Sea), their movement cannot fully quantify all species assemblage shifts. Furthermore, the observed rates of biogeographical movements are far greater than those observed in the terrestrial realm. Here, we discuss the processes that may explain such a discrepancy and suggest that the differences are mainly explained by the fluid nature of the pelagic domain, the life cycle of the zooplankton and the lesser anthropogenic influence (e.g. exploitation, habitat fragmentation) on these organisms. We also hypothesize that despite changes in the path and intensity of the oceanic currents that may modify quickly and greatly pelagic zooplankton species, these organisms may reflect better the current impact of climate warming on ecosystems as terrestrial organisms are likely to significantly lag the current impact of climate change.

Causes and prediction of abrupt climate-driven ecosystem shifts in the North Atlantic

Warming of the global climate is now unequivocal and its impact on Earth’ functional units has become more apparent. Here, we show that marine ecosystems are not equally sensitive to climate change and reveal a critical thermal boundary where a small increase in temperature triggers abrupt ecosystem shifts seen across multiple trophic levels. This large-scale boundary is located in regions where abrupt ecosystem shifts have been reported in the North Atlantic sector and thereby allows us to link these shifts by a global common phenomenon. We show that these changes alter the biodiversity and carrying capacity of ecosystems and may, combined with fishing, precipitate the reduction of some stocks of Atlantic cod already severely impacted by exploitation. These findings offer a way to anticipate major ecosystem changes and to propose adaptive strategies for marine exploited resources such as cod in order to minimize social and economic consequences.