Attention to specific regions of sensory space, or specific features of an object, can be rapidly shifted at will or in response to a stimulus and results in a significant increase in signal detectability (e.g., salience) and decrease in reaction times. Extracellular recordings in the visual system often reveal attention to be associated with an increase in neuronal responsiveness, especially to less salient stimuli (such as a low-contrast grating) (Reynolds et al. 2000; McAdams and Maunsell 1999). Through what mechanisms could attention result in a rapid change in the excitability of individual as well as larger groups of cortical neurons? The release of neuromodulatory agents, such as acetylcholine or norepi-nephrine, is unlikely to be responsible since these have far too slow a time course to underlie the rapid changes in excitability associated with shifts in attention. The leading hypothesis is that attentional mechanisms involve rapid reconfiguration of neuronal networks through shifts in the synaptic bombardment of key elements of the network that corresponds to the stimulus region or feature that is being attended to. Our results suggest that increasing the synaptic bombardment of a cortical cell with a depolarizing barrage of EPSPs and IPSPs may result in enhancements of neuronal excitability that are similar to those observed in some attentional paradigms (Reynolds et al. 2000). These shifts in responsiveness are naturally stronger for weak, versus strong stimuli (see Figure 16.7).
Attention can also result in an increase in neuronal "gain" for all stimuli, meaning that the spike rate output for each stimulus is increased by the same percentage, regardless of the magnitude of the input (McAdams and Maunsell 1999). How might such an increase in neuronal gain be achieved? It has been proposed that the rapid removal of barrages of synaptic potentials may underlie attentional changes in gain: if the incoming PSPs are perfectly balanced between excitation and inhibition, then there will be no net change in membrane potentials, and the reduced neuronal conductance will make the cell more responsive to its other inputs (Chance et al. 2002). However, this proposal has two unusual features. First, it implies that the neurons representing all of the large parameter space that is unattended are constantly bombarded with synaptic activity so as to keep their responsiveness low. Second, this model of attention requires a precise and ongoing balance of incoming EPSPs and IPSPs so that the membrane potential of the cell is unaffected by changes in background PSP rate (even though the membrane potential itself is constantly changing). It is difficult to imagine exactly how this precise balance between EPSPs, IPSPs, and membrane potential, even as it fluctuates, could be obtained so precisely. Rather, we propose that the network dynamics of neurons representing the attended object or attended spatial location are altered to facilitate a "pop-out" effect (Figure 16.8). We envision a dynamic interaction of a facilitatory attentional template of the attended object with the neurons that represent that object leading to an
Was this article helpful?