2002 Special issueNeuromodulation of decision and response selection
Introduction
Neuromodulators such as norepinephrine (NE) and dopamine are often assumed to mediate the attentional control of cognitive processing (Arnsten, 1998, Cohen and Servan-Schreiber, 1992). In a previous article (Usher, Cohen, Servan-Schreiber, Rajkowski, & Aston-Jones, 1999) we have presented neurophysiological and behavioral data (see also Aston-Jones, Rajkowski, Kubiak, & Alexinsky (1994)) showing that changes in NE responses of the brain nucleus Locus Coeruleus (LC) correlate with behavioral responses in awake monkey and we presented a neurocomputational model addressing the attentional mechanism of NE neuromodulation. In this paper, we extend and develop this computational model for NE neuromodulation to a wider domain of cognitive tasks that require decision and response selection. The first domain corresponds to perceptual choice situations, being a direct extension of the LC model presented in Usher et al. (1999). To do this we review data on attentional phases of LC responses in the LC, and show how a simplified model for LC response, functionally similar with the one derived by Gilzenrat, Holmes, Rajkowski, Aston-Jones, and Cohen (2002), can account for a variety of behavioral findings and leads to novel predictions. The second domain corresponds to a more complex cognitive process that involves memory maintenance followed by a probe-triggered response selection.
Section snippets
Attentional control and LC responses in perceptual choice
Previous behavioral studies of LC neuromodulation involved a simple go–nogo task (Aston-Jones et al., 1994). In this task monkeys are presented with a long sequence of stimuli, one type of which is designated as a ‘target’ that requires a lever-release, while all the other types are designated as ‘nontargets’ (or distractors) requiring to withhold responses. Typically, the target has a low frequency and the monkey needs to maintain vigilance in order to respond as fast as possible to targets
Model
The functional LC model described in this section is embedded in a behavioral model for perceptual choice and for selection from memory. Since the same LC model is used in both of these tasks, we first describe its mechanism and its attentional modes that are designed to account for the neurophysiological data in the LC described earlier.
The functional model for the LC is not attempted to be neurophysiologically realistic but only to provide an equivalent functional behavior to that found in
Behavioral model for the perceptual task
In this section, the model is applied to a straightforward extension of the go/nogo task described earlier—a forced choice between two response alternatives, both targets that require a response and which need to be distinguished from distractors that require the withholding of response (e.g. ‘respond to digits: 1 vs. 2 with the corresponding button and ignore letters’). The computational model used to address this task is shown in Fig. 3. It includes an input layer, a layer of perceptual
Response selection and memory maintenance
The cognitive task described earlier can be fully characterized as a choice of one out of many alternatives. In a recent work (Usher & McClelland, 2001) we have proposed that lateral inhibition between competing representations is a powerful mechanism suited to perform this type of winner-take-all selection. There are other cognitive tasks, however, where a winner-take-all selection is not advantageous. A simple example for such situations is a short-term memory (STM) test, where participants
General discussion
Attentional modulation of LC responses can be considered at two levels, a molar/descriptive level and a micro/mechanistic level. At the molar level, the LC modulation has been described by a gain hypothesis (Cohen & Servan-Schreiber, 1992). Accordingly, NE enhances or reduces the cortical responsivity in the high and low range, respectively. The aspect of neuromodulation explored in this paper is the dynamical modification of network parameters and input efficiency, which when triggered by
Acknowledgements
We wish to thank Jonathan D. Cohen and Gary Aston-Jones for insightful discussions on LC, NE and neuromodulation. EJD is a recipient of a departmental PhD studentship from the Birkbeck College.
References (28)
Catecholamine modulation of prefrontal cortical cognitive function
Trends in Cognitive Sciences
(1998)- et al.
Modulation of Gaba-A receptor-activated current by norepinephrine in cerebellar Purkinje cells
Neuroscience
(1992) - et al.
Noradrenergic enhancement of gaba-induced resistance changes in layer-V regular spiking neurons of rat somatosensory cortex
Brain Research
(1995) - et al.
Differential modulatory effects of norepinephrine on synaptically driven responses of layer V barrel field cortical neurons
Brain Research
(2000) - et al.
Excitatory and inhibitory interactions in localized populations of model neurons
Biological Cybernetics
(1972) Firing rate models for neural populations
- et al.
Comparison of current-discharge relationships of pyramidal neurons from the visual-cortex—in-vitro and in the anesthetized cat
Journal of Neurophysiology
(1995) - et al.
Estimates of the net excitatory currents evoked by visual stimulation of identified neurons in cat visual cortex
Cerebral cortex
(1998) - et al.
Quantitative study of attractor neural network retrieving at low spike rates: I substrate—Spikes, rates and neuronal gain
Network
(1991) - et al.
Distinctive features, categorical perception and probability learning: Some applications of a neural model
Psychological Review
(1977)
Locus-Coeruleus neurons in monkey are selectively activated by attended cues in a vigilance task
Journal of Neuroscience
Conflict monitoring and cognitive control
Psychological Review
Context, cortex, and dopamine: A connectionist approach to behavior and biology in schizophrenia
Psychological Review
The magical number 4 in short-term memory: A reconsideration of mental storage capacity
Behavioral and Brain Sciences
Cited by (29)
The timescale of perceptual evidence integration can be adapted to the environment
2013, Current BiologyCitation Excerpt :Physiological evidence suggests that such network interactions during decision-making span multiple regions that are widely distributed across the brain [29–31]. The adaptive changes in integration timescales found in the present study may, therefore, reflect the flexible adjustment of such large-scale decision networks, which might be achieved by neuromodulation [32–34]. Finally, our study sets the stage for future neurophysiological studies of the biophysical circuit mechanisms underlying perceptual evidence integration.
Positive and negative congruency effects in masked priming: A neuro-computational model based on representation, attention, and conflict
2009, Brain ResearchCitation Excerpt :The current model, in addition to being biologically compelling, showed many dynamic effects in RT and error patterns that have not been shown previously (such as the changes in RT and the size of priming effects through time). The model is based on previous models that have been used to simulate different tasks such as target detection and simple decisions in monkeys and humans (Gilzenrat et al., 2002; Usher et al., 1999; Usher and Davelaar, 2002; Nieuwenhuis et al., 2005). The present model is similar to some other previous neuro-computational models, especially those employed to simulate the attentional blink (Mathis and Mozer, 1996; Nieuwenhuis et al., 2005).
What has been learned from computational models of attention
2006, Neural NetworksMetalearning and neuromodulation
2002, Neural NetworksSimplified dynamics in a model of noradrenergic modulation of cognitive performance
2002, Neural Networks