- Email: [email protected]
Phase Resetting as a Mechanism for Supramodal Attentional Control
Neuron
Previews REFERENCES Buzsaki, G., and Draguhn, A. (2004). Science 304, 1926–1929. Castro-Alamancos, M.A. (2002). J. Neurosci. 22, 9651–9655.
Haider, B., and McCormick, D.A. (2009). Neuron 62, 171–189.
Poulet, J.F.A., and Petersen, C.C.H. (2008). Nature 454, 881–885.
Kerr, J.N., Greenberg, D., and Helmchen, F. (2005). Proc. Natl. Acad. Sci. USA 102, 14063–14068.
Sakata, S., and Harris, K.D. (2009). Neuron 64, this issue, 404–418.
Crochet, S., and Petersen, C.C.H. (2006). Nat. Neurosci. 9, 608–610.
Lee, A.K., Manns, I.D., Sakmann, B., and Brecht, M. (2006). Neuron 51, 399–407.
de Kock, C.P., and Sakmann, B. (2009). Proc. Natl. Acad. Sci. USA 106, 16446–16450.
Lefort, S., Tomm, C., Sarria, J.C.F., and Petersen, C.C.H. (2009). Neuron 61, 301–316.
Ferezou, I., Haiss, F., Gentet, L.J., Aronoff, R., Weber, B., and Petersen, C.C.H. (2007). Neuron 56, 907–923.
Otazu, G.H., Tai, L.H., Yang, Y., and Zador, A.M. (2009). Nat. Neurosci. 12, 646–654.
Sanchez-Vives, M.V., and McCormick, (2000). Nat. Neurosci. 3, 1027–1034.
D.A.
Timofeev, I., Grenier, F., Bazhenov, M., Sejnowski, T.J., and Steriade, M. (2000). Cereb. Cortex 10, 1185–1199. Volgushev, M., Chauvette, S., Mukovski, M., and Timofeev, I. (2006). J. Neurosci. 26, 5665– 5672.
Phase Resetting as a Mechanism for Supramodal Attentional Control Christoph Kayser1,* 1Max Planck Institute for Biological Cybernetics, Spemannstrasse 38, 72076 Tu ¨ bingen, Germany *Correspondence: [email protected] DOI 10.1016/j.neuron.2009.10.022
Attentional modulation and cross-modal integration might partly rely on the same neurophysiological mechanisms. As a new study by Lakatos et al. in this issue of Neuron shows, attended stimuli in one sensory modality not only modulate oscillatory activity within the primary cortex of the same modality but also reset the phase of ongoing oscillations in primary cortices of other modalities. Imagine being at a cocktail party, trying to understand a friend who is telling you about his latest daring exploits. In the background, there is the din of noisy chatter and loud music is playing. To better understand what your friend is saying, your brain has two possibilities: to use directed attention to tune out irrelevant sounds or to enhance the acoustic speech signal by integrating visual information about your friend’s lip movements. Both attentional selection and cross-modal integration are important mechanisms to structure sensory information and enhance our perceptual abilities. As a study in this issue of Neuron shows (Lakatos et al., 2009), when operating in primary sensory cortices, both attentional selection and cross-modal integration might rely on the same neurophysiological mechanism— modulation of ongoing oscillatory activity. A role of oscillations in attention is suggested by reports of different rhythms
being enhanced during conditions of focused attention (Fries, 2009; Schroeder and Lakatos, 2009). A role of oscillations in cross-modal integration, in addition, is suggested by studies investigating early multisensory influences. For example, when stimuli presented to one modality alter activity in primary cortices of another modality, this often results in enhanced oscillatory activity (Ghazanfar and Schroeder, 2006; Senkowski et al., 2008). In the new study, Lakatos and colleagues now reason that if attention and crossmodal influences in primary sensory cortices are mediated via similar neurophysiological mechanisms, this would provide a unique means for supramodal attentional control to regulate both selection and binding of cross-modal information at the same time. To test this experimentally, they recorded neural activity in primary visual and auditory cortices of monkeys performing an intermodal selec-
300 Neuron 64, November 12, 2009 ª2009 Elsevier Inc.
tion task. This allowed them to compare responses to visual and auditory stimuli within each of these cortical areas when the respective stimulus was either attended or ignored. To investigate the involvement of oscillatory activity, they recorded laminar profiles of field potentials and used these to calculate one-dimensional current source density profiles (CSD). Similar to local field potentials, CSDs reflect rhythmic subthreshold activity, such as synaptic potentials and membrane fluctuations, but are less influenced by volume conduction and allow activity to be localized to the different cortical laminae. Overall, Lakatos and colleagues obtained very comparable results whether studying activity in visual or auditory cortex. When recording from the same sensory cortex as the target stimulus (e.g., a visual stimulus when recording from visual cortex), CSDs revealed
Neuron
Previews
Figure 1. Impact of Attended Stimuli on Ongoing Oscillations Colored traces depict idealized theta oscillations in supragranular field potentials in visual cortex, shaped to reflect the main findings by Lakatos and colleagues. Color denotes the oscillation phase (separated in quarter cycles). (A) Responses to attended visual stimuli comprised both an evoked response (here depicted as increase in amplitude) as well as phase resetting: before stimulus onset, the phase of individual oscillation cycles varies across trials, while after stimulus presentation phase values become highly consistent across trials. This is achieved by setting the phase to the same value in different trials—hence the name ‘‘phase resetting.’’ Ignored visual stimuli (attention deployed to the auditory modality) also caused an evoked response but only weakly affected the phase of ongoing oscillations (not shown in figure). (B) Attended acoustic stimuli induced phase resetting in visual cortex, hence demonstrating cross-modal influences in a primary cortex. However, auditory induced responses did not encompass increases in the amplitude of oscillations but only induced phase resetting. Ignored acoustic stimuli did not affect oscillations in visual cortex (not shown). Very consistent results were also found when recording in auditory cortex. In this case, attended acoustic stimuli caused phase resetting and evoked responses, while attended visual stimuli only induced phase resetting. Together, this demonstrates that attention to one modality can affect ongoing oscillatory activity in primary cortices of the same and other sensory modalities.
well-known patterns of feed-forward processing: activity emerged in the cortical input layer and subsequently spread to supra- and infragranular layers. Importantly, when this ‘‘dominant’’ modality was attended, attention enhanced the evoked response and rendered the phase of ongoing oscillations more consistent across repeats of the stimulus (Figure 1A). The latter effect, called ‘‘phase resetting,’’ implies that the precise timing of individual oscillation cycles becomes highly consistent across trials and demonstrates that attention can prominently alter the precise timing of ongoing oscillations. When activity was recorded from the other sensory cortex as the stimulus was presented (e.g., an auditory stimulus when recording from visual cortex), a striking result emerged: attended stimuli still elicited a significant change in ongoing activity, but ignored stimuli did not. Importantly, this response to attended stimuli in another modality involved phase resetting, but not changes in oscillation amplitude (Figure 1B). Effectively, this puts the attended modality in a position of supramodal control: stimuli in the attended modality not only evoke a response and modulate oscillatory ac-
tivity within the primary cortex of the same modality but also set the oscillatory ‘‘context’’ in primary cortices of other modalities to a well-defined state. By virtue of this mechanism, one sensory modality—the leading sense at each moment in time—might orchestrate sensory integration and attentional selection across modalities (Lakatos et al., 2009). Previous studies demonstrated that the phase of slow oscillations indeed can be regarded as neurophysiological ‘‘context’’ that determines the impact of sensory stimuli. Electrophysiological studies found that the amplitude of stimulus-evoked responses depends on the phase of ongoing oscillations at the time of stimulus presentation (Schroeder et al., 2008). And recent behavioral studies indicate that prestimulus phase also modulates the chance of faint stimuli to be detected by human observers (as reviewed in Wyart and Sergent, 2009). Hence, phase resetting by supramodal attention might be an effective means to modulate the perceptual saliency of individual stimuli. As important detail, the results of Lakatos et al. show that phase resetting within one cortical area by attended stimuli in
another modality is stronger than by ignored stimuli in the same modality. This property might be crucial for phase resetting to help to correctly bind attended sensory information across modalities. At the cocktail party, for example, attention toward the lip movements controls the phase of ongoing oscillations in auditory cortex, and this phase control is stronger than the influence of distracting sounds. As a result, visual attentional control over ongoing oscillations in auditory cortex might prevent distracting sounds from taking control over activity in auditory cortex. It may not come as a surprise that attentional selection and sensory integration are linked by the same neurophysiological mechanism, considering their similar impact on processing and perception. Both mechanisms enhance the representation of ‘‘chosen’’ stimuli, and both exert their strongest influences when the respective unisensory responses are weak. For example, attentional modulation of single neurons is strongest for stimuli weakly driving the neuron (Reynolds et al., 2000), and cross-modal influences obey a similar principle of inverse effectiveness (Stein and Stanford, 2008). Future work could elucidate whether
Neuron 64, November 12, 2009 ª2009 Elsevier Inc. 301
Neuron
Previews these similar functional characteristics indeed arise from the same underlying neural mechanisms. Phase resetting in the Lakatos study was most prominent in theta (4–8 Hz) and gamma (>30 Hz) frequency bands, in concordance with studies implicating these frequencies in cross-modal binding and attention (Fries, 2009; Schroeder and Lakatos, 2009; Senkowski et al., 2008). Gamma synchronization has also been implicated in affording efficient neural communication by aligning the windows of optimal excitability between two neural populations (Fries, 2009). It will be interesting to test whether and how phase resetting in a single trial might be orchestrated across different sensory cortices, in a manner that promotes such aligned temporal windows for efficient neural communication. Phase resetting of oscillations constitutes a manipulation of temporal aspects of neural activity. One might hence speculate whether attentional influences in primary cortices at the level of individual neurons might also be more strongly expressed in temporal measures of activity than for example in firing rates. As a very recent study shows, cross-modal influences in auditory cortex enhance the reliability of firing rates and increase the precision of temporal spike patterns (Kayser et al., 2009a). Believing in shared mechanisms for attention and sensory integration hence only fosters such speculations.
It will also be revealing to understand the molecular and biophysical processes underlying phase resetting. Modulation of oscillatory activity and attention have both been attributed to cholinergic modulation (Deco and Thiele, 2009). Brain regions implicated in attentional control project to cholinergic centers and thalamic matrix systems, which provide diffusely projecting modulatory inputs to different sensory areas (Zikopoulos and Barbas, 2007). These diffuse projections preferentially terminate in supragranular cortical layers, in good accordance with the finding that phase resetting was strongest in these layers (Lakatos et al., 2009). Furthermore, it might not be mere coincidence that the membrane potential of neurons in supragranular layers shows spontaneous oscillations at near theta frequencies (Sun and Dan, 2009). Finally, it will be interesting to integrate these findings with models of sensory coding that assign a role as temporal frame of reference to slow oscillations. Practically, such an intrinsic frame of reference can increase the sensory information provided by neural responses and can enhance the robustness of neural representations to sensory noise, as exemplified by a recent study in monkey auditory cortex (Kayser et al., 2009b). In such a framework of neural coding, the impact of supramodal phase resetting could be directly quantified in terms of information gain for specific neural representations. This might provide further
302 Neuron 64, November 12, 2009 ª2009 Elsevier Inc.
insights into how the ‘‘leading sense’’ links attention and cross-modal integration, allowing us to form a coherent percept across sensory modalities. REFERENCES Deco, G., and Thiele, A. (2009). Eur. J. Neurosci. 30, 347–354. Fries, P. (2009). Annu. Rev. Neurosci. 32, 209–224. Ghazanfar, A.A., and Schroeder, C.E. (2006). Trends Cogn. Sci. 10, 278–285. Kayser, C., Logothetis, N., and Panzeri, S. (2009a). Curr. Biol., in press. Kayser, C., Montemurro, M.A., Logothetis, N., and Panzeri, S. (2009b). Neuron 61, 597–608. Lakatos, P., O’Connell, M.N., Barczak, A., Mills, A., Javitt, D.C., and Schroeder, C.E. (2009). Neuron 64, this issue, 419–430. Reynolds, J.H., Pasternak, T., and Desimone, R. (2000). Neuron 26, 703–714. Schroeder, C.E., and Lakatos, P. (2009). Trends Neurosci. 32, 9–18. Schroeder, C.E., Lakatos, P., Kajikawa, Y., Partan, S., and Puce, A. (2008). Trends Cogn. Sci. 12, 106–113. Senkowski, D., Schneider, T.R., Foxe, J.J., and Engel, A.K. (2008). Trends Neurosci. 31, 401–409. Stein, B.E., and Stanford, T.R. (2008). Nat. Rev. Neurosci. 9, 255–266. Sun, W., and Dan, Y. (2009). Proc. Natl. Acad. Sci. USA 106, 17986–17991. Wyart, V., and Sergent, C. (2009). J. Neurosci. 29, 12839–12841. Zikopoulos, B., and Barbas, H. (2007). Rev. Neurosci. 18, 417–438.
Previews REFERENCES Buzsaki, G., and Draguhn, A. (2004). Science 304, 1926–1929. Castro-Alamancos, M.A. (2002). J. Neurosci. 22, 9651–9655.
Haider, B., and McCormick, D.A. (2009). Neuron 62, 171–189.
Poulet, J.F.A., and Petersen, C.C.H. (2008). Nature 454, 881–885.
Kerr, J.N., Greenberg, D., and Helmchen, F. (2005). Proc. Natl. Acad. Sci. USA 102, 14063–14068.
Sakata, S., and Harris, K.D. (2009). Neuron 64, this issue, 404–418.
Crochet, S., and Petersen, C.C.H. (2006). Nat. Neurosci. 9, 608–610.
Lee, A.K., Manns, I.D., Sakmann, B., and Brecht, M. (2006). Neuron 51, 399–407.
de Kock, C.P., and Sakmann, B. (2009). Proc. Natl. Acad. Sci. USA 106, 16446–16450.
Lefort, S., Tomm, C., Sarria, J.C.F., and Petersen, C.C.H. (2009). Neuron 61, 301–316.
Ferezou, I., Haiss, F., Gentet, L.J., Aronoff, R., Weber, B., and Petersen, C.C.H. (2007). Neuron 56, 907–923.
Otazu, G.H., Tai, L.H., Yang, Y., and Zador, A.M. (2009). Nat. Neurosci. 12, 646–654.
Sanchez-Vives, M.V., and McCormick, (2000). Nat. Neurosci. 3, 1027–1034.
D.A.
Timofeev, I., Grenier, F., Bazhenov, M., Sejnowski, T.J., and Steriade, M. (2000). Cereb. Cortex 10, 1185–1199. Volgushev, M., Chauvette, S., Mukovski, M., and Timofeev, I. (2006). J. Neurosci. 26, 5665– 5672.
Phase Resetting as a Mechanism for Supramodal Attentional Control Christoph Kayser1,* 1Max Planck Institute for Biological Cybernetics, Spemannstrasse 38, 72076 Tu ¨ bingen, Germany *Correspondence: [email protected] DOI 10.1016/j.neuron.2009.10.022
Attentional modulation and cross-modal integration might partly rely on the same neurophysiological mechanisms. As a new study by Lakatos et al. in this issue of Neuron shows, attended stimuli in one sensory modality not only modulate oscillatory activity within the primary cortex of the same modality but also reset the phase of ongoing oscillations in primary cortices of other modalities. Imagine being at a cocktail party, trying to understand a friend who is telling you about his latest daring exploits. In the background, there is the din of noisy chatter and loud music is playing. To better understand what your friend is saying, your brain has two possibilities: to use directed attention to tune out irrelevant sounds or to enhance the acoustic speech signal by integrating visual information about your friend’s lip movements. Both attentional selection and cross-modal integration are important mechanisms to structure sensory information and enhance our perceptual abilities. As a study in this issue of Neuron shows (Lakatos et al., 2009), when operating in primary sensory cortices, both attentional selection and cross-modal integration might rely on the same neurophysiological mechanism— modulation of ongoing oscillatory activity. A role of oscillations in attention is suggested by reports of different rhythms
being enhanced during conditions of focused attention (Fries, 2009; Schroeder and Lakatos, 2009). A role of oscillations in cross-modal integration, in addition, is suggested by studies investigating early multisensory influences. For example, when stimuli presented to one modality alter activity in primary cortices of another modality, this often results in enhanced oscillatory activity (Ghazanfar and Schroeder, 2006; Senkowski et al., 2008). In the new study, Lakatos and colleagues now reason that if attention and crossmodal influences in primary sensory cortices are mediated via similar neurophysiological mechanisms, this would provide a unique means for supramodal attentional control to regulate both selection and binding of cross-modal information at the same time. To test this experimentally, they recorded neural activity in primary visual and auditory cortices of monkeys performing an intermodal selec-
300 Neuron 64, November 12, 2009 ª2009 Elsevier Inc.
tion task. This allowed them to compare responses to visual and auditory stimuli within each of these cortical areas when the respective stimulus was either attended or ignored. To investigate the involvement of oscillatory activity, they recorded laminar profiles of field potentials and used these to calculate one-dimensional current source density profiles (CSD). Similar to local field potentials, CSDs reflect rhythmic subthreshold activity, such as synaptic potentials and membrane fluctuations, but are less influenced by volume conduction and allow activity to be localized to the different cortical laminae. Overall, Lakatos and colleagues obtained very comparable results whether studying activity in visual or auditory cortex. When recording from the same sensory cortex as the target stimulus (e.g., a visual stimulus when recording from visual cortex), CSDs revealed
Neuron
Previews
Figure 1. Impact of Attended Stimuli on Ongoing Oscillations Colored traces depict idealized theta oscillations in supragranular field potentials in visual cortex, shaped to reflect the main findings by Lakatos and colleagues. Color denotes the oscillation phase (separated in quarter cycles). (A) Responses to attended visual stimuli comprised both an evoked response (here depicted as increase in amplitude) as well as phase resetting: before stimulus onset, the phase of individual oscillation cycles varies across trials, while after stimulus presentation phase values become highly consistent across trials. This is achieved by setting the phase to the same value in different trials—hence the name ‘‘phase resetting.’’ Ignored visual stimuli (attention deployed to the auditory modality) also caused an evoked response but only weakly affected the phase of ongoing oscillations (not shown in figure). (B) Attended acoustic stimuli induced phase resetting in visual cortex, hence demonstrating cross-modal influences in a primary cortex. However, auditory induced responses did not encompass increases in the amplitude of oscillations but only induced phase resetting. Ignored acoustic stimuli did not affect oscillations in visual cortex (not shown). Very consistent results were also found when recording in auditory cortex. In this case, attended acoustic stimuli caused phase resetting and evoked responses, while attended visual stimuli only induced phase resetting. Together, this demonstrates that attention to one modality can affect ongoing oscillatory activity in primary cortices of the same and other sensory modalities.
well-known patterns of feed-forward processing: activity emerged in the cortical input layer and subsequently spread to supra- and infragranular layers. Importantly, when this ‘‘dominant’’ modality was attended, attention enhanced the evoked response and rendered the phase of ongoing oscillations more consistent across repeats of the stimulus (Figure 1A). The latter effect, called ‘‘phase resetting,’’ implies that the precise timing of individual oscillation cycles becomes highly consistent across trials and demonstrates that attention can prominently alter the precise timing of ongoing oscillations. When activity was recorded from the other sensory cortex as the stimulus was presented (e.g., an auditory stimulus when recording from visual cortex), a striking result emerged: attended stimuli still elicited a significant change in ongoing activity, but ignored stimuli did not. Importantly, this response to attended stimuli in another modality involved phase resetting, but not changes in oscillation amplitude (Figure 1B). Effectively, this puts the attended modality in a position of supramodal control: stimuli in the attended modality not only evoke a response and modulate oscillatory ac-
tivity within the primary cortex of the same modality but also set the oscillatory ‘‘context’’ in primary cortices of other modalities to a well-defined state. By virtue of this mechanism, one sensory modality—the leading sense at each moment in time—might orchestrate sensory integration and attentional selection across modalities (Lakatos et al., 2009). Previous studies demonstrated that the phase of slow oscillations indeed can be regarded as neurophysiological ‘‘context’’ that determines the impact of sensory stimuli. Electrophysiological studies found that the amplitude of stimulus-evoked responses depends on the phase of ongoing oscillations at the time of stimulus presentation (Schroeder et al., 2008). And recent behavioral studies indicate that prestimulus phase also modulates the chance of faint stimuli to be detected by human observers (as reviewed in Wyart and Sergent, 2009). Hence, phase resetting by supramodal attention might be an effective means to modulate the perceptual saliency of individual stimuli. As important detail, the results of Lakatos et al. show that phase resetting within one cortical area by attended stimuli in
another modality is stronger than by ignored stimuli in the same modality. This property might be crucial for phase resetting to help to correctly bind attended sensory information across modalities. At the cocktail party, for example, attention toward the lip movements controls the phase of ongoing oscillations in auditory cortex, and this phase control is stronger than the influence of distracting sounds. As a result, visual attentional control over ongoing oscillations in auditory cortex might prevent distracting sounds from taking control over activity in auditory cortex. It may not come as a surprise that attentional selection and sensory integration are linked by the same neurophysiological mechanism, considering their similar impact on processing and perception. Both mechanisms enhance the representation of ‘‘chosen’’ stimuli, and both exert their strongest influences when the respective unisensory responses are weak. For example, attentional modulation of single neurons is strongest for stimuli weakly driving the neuron (Reynolds et al., 2000), and cross-modal influences obey a similar principle of inverse effectiveness (Stein and Stanford, 2008). Future work could elucidate whether
Neuron 64, November 12, 2009 ª2009 Elsevier Inc. 301
Neuron
Previews these similar functional characteristics indeed arise from the same underlying neural mechanisms. Phase resetting in the Lakatos study was most prominent in theta (4–8 Hz) and gamma (>30 Hz) frequency bands, in concordance with studies implicating these frequencies in cross-modal binding and attention (Fries, 2009; Schroeder and Lakatos, 2009; Senkowski et al., 2008). Gamma synchronization has also been implicated in affording efficient neural communication by aligning the windows of optimal excitability between two neural populations (Fries, 2009). It will be interesting to test whether and how phase resetting in a single trial might be orchestrated across different sensory cortices, in a manner that promotes such aligned temporal windows for efficient neural communication. Phase resetting of oscillations constitutes a manipulation of temporal aspects of neural activity. One might hence speculate whether attentional influences in primary cortices at the level of individual neurons might also be more strongly expressed in temporal measures of activity than for example in firing rates. As a very recent study shows, cross-modal influences in auditory cortex enhance the reliability of firing rates and increase the precision of temporal spike patterns (Kayser et al., 2009a). Believing in shared mechanisms for attention and sensory integration hence only fosters such speculations.
It will also be revealing to understand the molecular and biophysical processes underlying phase resetting. Modulation of oscillatory activity and attention have both been attributed to cholinergic modulation (Deco and Thiele, 2009). Brain regions implicated in attentional control project to cholinergic centers and thalamic matrix systems, which provide diffusely projecting modulatory inputs to different sensory areas (Zikopoulos and Barbas, 2007). These diffuse projections preferentially terminate in supragranular cortical layers, in good accordance with the finding that phase resetting was strongest in these layers (Lakatos et al., 2009). Furthermore, it might not be mere coincidence that the membrane potential of neurons in supragranular layers shows spontaneous oscillations at near theta frequencies (Sun and Dan, 2009). Finally, it will be interesting to integrate these findings with models of sensory coding that assign a role as temporal frame of reference to slow oscillations. Practically, such an intrinsic frame of reference can increase the sensory information provided by neural responses and can enhance the robustness of neural representations to sensory noise, as exemplified by a recent study in monkey auditory cortex (Kayser et al., 2009b). In such a framework of neural coding, the impact of supramodal phase resetting could be directly quantified in terms of information gain for specific neural representations. This might provide further
302 Neuron 64, November 12, 2009 ª2009 Elsevier Inc.
insights into how the ‘‘leading sense’’ links attention and cross-modal integration, allowing us to form a coherent percept across sensory modalities. REFERENCES Deco, G., and Thiele, A. (2009). Eur. J. Neurosci. 30, 347–354. Fries, P. (2009). Annu. Rev. Neurosci. 32, 209–224. Ghazanfar, A.A., and Schroeder, C.E. (2006). Trends Cogn. Sci. 10, 278–285. Kayser, C., Logothetis, N., and Panzeri, S. (2009a). Curr. Biol., in press. Kayser, C., Montemurro, M.A., Logothetis, N., and Panzeri, S. (2009b). Neuron 61, 597–608. Lakatos, P., O’Connell, M.N., Barczak, A., Mills, A., Javitt, D.C., and Schroeder, C.E. (2009). Neuron 64, this issue, 419–430. Reynolds, J.H., Pasternak, T., and Desimone, R. (2000). Neuron 26, 703–714. Schroeder, C.E., and Lakatos, P. (2009). Trends Neurosci. 32, 9–18. Schroeder, C.E., Lakatos, P., Kajikawa, Y., Partan, S., and Puce, A. (2008). Trends Cogn. Sci. 12, 106–113. Senkowski, D., Schneider, T.R., Foxe, J.J., and Engel, A.K. (2008). Trends Neurosci. 31, 401–409. Stein, B.E., and Stanford, T.R. (2008). Nat. Rev. Neurosci. 9, 255–266. Sun, W., and Dan, Y. (2009). Proc. Natl. Acad. Sci. USA 106, 17986–17991. Wyart, V., and Sergent, C. (2009). J. Neurosci. 29, 12839–12841. Zikopoulos, B., and Barbas, H. (2007). Rev. Neurosci. 18, 417–438.