Mather et al. have proposed an intriguing theory to explain how norepinephrine (NE) release and subsequent noradrenergic modulatory actions are focused in neural circuits by concomitant “priority” stimulus-driven release of glutamate. In doing so they confront a question that has perplexed the field for some time, that is, how to account for selectivity of signal processing in noradrenergic terminal fields and focused perception of salient events when tonic discharge from the broadly projecting NE-containing nucleus locus coeruleus (LC) is elevated, as would occur during generalized arousal. Here we focus on how the theory applies to NE modulation of sensory signal processing. Given the results of four decades of published work, we would expect increases in LC–NE output to promote enhanced neuronal and neural network responses to sensory-driven afferent inputs (Berridge & Waterhouse Reference Berridge and Waterhouse2003), actions that have been linked to improved performance of sensory-guided behavioral tasks (Aston-Jones et al. Reference Aston-Jones, Rajkowski and Cohen1999; Rajkowski et al. Reference Rajkowski, Majczynski, Clayton and Aston-Jones2004).

Until recently, conventional wisdom was that the LC–NE efferent network was broadly distributed from a relatively small number of brainstem neurons with homogeneous physiological properties. Given this, the presumption was that NE was released uniformly and simultaneously across all terminal fields within the forebrain, cerebellum, and spinal cord for as long as LC is discharging, either tonically or phasically. If that were the case, neuronal and neural circuit responsiveness to sensory driven afferent inputs would be increased throughout the central nervous system without any bias in favor of one sensory signal versus another. If responsiveness to synaptically driven inputs is elevated everywhere and for every modality, what has been gained? Is there a way for the LC–NE system to selectively differentiate sensory signals from the constant stream of information that is presented to the nervous system from the periphery? Mather and colleagues' GANE (glutamate amplifies noradrenergic effects) theory is timely in so far as it appropriately confronts these issues.

An idea similar to the current theory was suggested by Marrocco et al. (Reference Marrocco, Lane, McClurkin, Blaha and Alkire1987) after they observed a correlation between catecholamine release in monkey visual cortex and coincident light-evoked activity in geniculostriate projections to ocular dominance columns. These authors postulated a local interaction between NE fibers and geniculostriate afferents: one that created a local hotspot for NE release within the visual cortex and, thus, preferentially promoted modulation of synaptic transmission at this site. Akin to Marrocco and colleagues' proposal, the GANE theory argues that locally released glutamate provides the means for amplifying release of NE from tonically or phasically active LC–NE fibers.

The GANE theory accounts for many, but not all of the well-documented attributes and operational capacities of the LC–NE system, particularly those demonstrated in sensory networks. The authors exhaustively reviewed an extensive literature including many reports that support the core of their proposal: a positive feedback mechanism whereby synaptic release of glutamate amplifies NE release from nearby noradrenergic axons and results in enhanced responsiveness of neurons and glia to glutamate neurotransmission at this local site of interaction. The process relies on a delicate balance and interplay between receptor-mediated actions that are triggered as extracellular concentrations of NE and glutamate change. The temporal and spatial dynamics of these interactions are postulated, but experimental evidence to support the details of these mechanisms is lacking. For example, to date, the extracellular tissue concentrations of NE that yield the range of modulatory actions demonstrated in vivo and in vitro have been only crudely approximated based upon the results of microdialysis studies (Berridge and Abercrombie Reference Berridge and Abercrombie1999; Florin-Lechner et al. Reference Florin-Lechner, Druhan, Aston-Jones and Valentino1996). As illustrated in many studies, LC–NE modulatory effects are expressed according to an inverted-U dose–response curve, rising to optimal facilitation of cellular and behavioral events as LC–NE activity increases and then falling to suppression of neuronal responsiveness and disrupted task performance as NE concentrations and LC discharge increase further (Aston-Jones et al. Reference Aston-Jones, Rajkowski and Cohen1999; Devilbiss & Waterhouse Reference Devilbiss and Waterhouse2000). At what point along this inverted-U function is the glutamate–NE interaction operating? At some level of glutamate release, the facilitating actions of NE would be expected to diminish along the right side of the function.

Other aspects of LC–NE action require attention in the GANE model. For example, GANE relies on β-receptor mediated modulation of excitatory synaptic transmission and minimizes a role for α₁-receptors, despite evidence in sensory circuits that α₁-receptor activation augments postsynaptic responses to excitatory inputs (Mouradian et al. Reference Mouradian, Seller and Waterhouse1991; Nai et al. Reference Nai, Dong, Linster and Ennis2010; Rogawski & Aghajanian Reference Rogawski and Aghajanian1982). In this same vein, β-receptor activation has been reported to enhance postsynaptic responses to γ-aminobutyric acid (GABA) (Cheun & Yeh Reference Cheun and Yeh1992; Waterhouse et al. Reference Waterhouse, Moises, Yeh and Woodward1982). The latter mechanism could account for the enhanced lateral inhibition that is invoked as a means of establishing additional contrast between a priority stimulus-driven hotspot and its tissue surround. Evidence for increased lateral inhibition and focusing of the zone of neural excitation created by prioritized inputs can be found in studies that examined the impact of the LC–NE system on off-beam inhibition in the cerebellar cortex (Moises et al. Reference Moises, Waterhouse and Woodward1983) and receptive field properties of visually responsive neurons (Waterhouse et al. Reference Waterhouse, Azizi, Burne and Woodward1990). Finally, a “gating” effect of the LC–NE system on otherwise subthreshold synaptic inputs has been demonstrated in multiple brain circuits (Waterhouse et al. Reference Waterhouse, Sessler, Cheng, Woodward, Azizi and Moises1988). This action of NE would serve to recruit additional neurons into a sensory response pool, as opposed to suppression of hotspot surround activity.

In summary, the model works reasonably well in support of circumstances in which a noradrenergically innervated circuit is called on to process behaviorally imperative information that should be prioritized for immediate response and/or accentuated for future retrieval and experience-guided action, for example, fear-related or reward-generating stimuli. This theoretical construct is extremely valuable in providing the platform for generating testable hypotheses about glutamate-facilitated noradrenergic transmission in sensory and cognitive circuits that process potentially imperative stimuli. However, it is well to remember that much of the evidence for NE modulatory actions in sensory circuits has been demonstrated in anesthetized or controlled waking conditions where stimuli are behaviorally irrelevant. How does GANE account for bottom-up differentiation and prioritization of sensory signals in the absence of behavioral relevance? We applaud the model but look forward to resolution of how GANE integrates with conventional NE modulation to differentiate signals amidst the constant stream of incoming information from the sensory surround at both early-stage sensory relays and areas of higher-order sensory/cognitive integration.