1. A Neural Circuit for Auditory Dominance over Visual Perception
You-Hyang Song, Jae-Hyun Kim, Hye-Won Jeong, Ilsong Choi, Daun Jeong, Kwansoo Kim, Seung-Hee Lee 
Neuron 
Volume 93, Issue 4, Pages e6 (February 2017)
DOI: /j.neuron
Copyright © 2017 Elsevier Inc. Terms and Conditions

2. Figure 1
Auditory-Dominant Perception in Mice Performing Auditory-Visual Discrimination Tasks
(A) Schematic of the discrimination and conflict tasks. VIS, visual stimulus; AUD, auditory stimulus.
(B) Task structure. Vertical lines indicate sensory stimuli and the horizontal black bar indicates the response window.
(C) Performance in detecting different stimulus intensities. Red lines, visual detection rates (n = 8); blue lines, auditory detection rates (n = 8); black and gray lines, unstimulated spontaneous lick rates (mean ± SEM).
(D) Learning to discriminate a flashing light from pure tones. Changes in lick rate and discriminability (d’) over the training period. (Top) VISgo-AUDnogo (n = 10). (Bottom) AUDgo-VISnogo (n = 9). Red lines, average lick rate to the flashing light; blue lines, average lick rate to the pure tone; gray lines, d’ of individual mice; black lines, average d’; shading and vertical lines, SEM.
(E) Lick rate in response to VIS (red), AUD (blue), or VIS+AUD (green) stimuli at different intensities (subjective salience levels 1–4) (VISgo-AUDnogo, n = 10; AUDgo-VISnogo, n = 8; mean ± SEM; ∗∗∗p < 0.001, n.s., not significant; paired t test).
See also Figures S1 and S2 and Movies S1 and S2.
Neuron  , e6DOI: ( /j.neuron )
Copyright © 2017 Elsevier Inc. Terms and Conditions

3. Figure 2
Virtual Perception Induced by Optogenetic Stimulation of the VC and AC
(A) Schematic of the artificial discrimination task (top) and actual discrimination task (bottom). Task structure was the same as that shown in Figure 1B.
(B) Coronal view of a VC brain slice from a CaMKIIα::ChR2 mouse. Green, ChR2-eYFP. Scale bar, 1,000 μm.
(C) Changes in multi-unit firing activity in the VC and AC in response to optogenetic stimulation (n = 3 mice; n.s., not significant; ∗∗p = ; ∗p = 0.019, paired t test; red, VC stimulation; blue, AC stimulation; horizontal and vertical lines, mean ± SEM).
(D) Learning to discriminate optogenetic stimulation of the VC or AC. Changes in lick rate and discriminability (d’) over the training period. (Top) VCgo-ACnogo (n = 9). (Bottom) ACgo-VCnogo (n = 9). Red lines, average lick rate to the VC stimulation; blue lines, average lick rate to the AC stimulation; gray lines, d’ of individual mice; black lines, average d’; shading and vertical lines, SEM.
(E) Discrimination of actual stimuli in mice trained to discriminate artificial VC-AC stimuli. (Top) VCgo-ACnogo (n = 7). (Bottom) ACgo-VCnogo (n = 4). ∗∗∗p < 0.001; ∗∗p =  and ; n.s., not significant, paired t test.
(F) Discrimination of artificial optogenetic stimuli in mice trained to discriminate actual sensory stimuli. (Top) VISgo-AUDnogo (n = 5). (Bottom) AUDgo-VISnogo (n = 6). ∗∗∗p < 0.001; n.s., not significant, paired t test. The sloping lines indicate the licking performance of individual mice, while the horizontal and vertical lines indicate mean ± SEM.
See also Figure S2 and Movie S3.
Neuron  , e6DOI: ( /j.neuron )
Copyright © 2017 Elsevier Inc. Terms and Conditions

4. Figure 3 AC-Dominant Perceptual Integration of Audiovisual Conflicts
(A) Schematic for testing VC-AC conflict resolution in the artificial discrimination task.
(B) Lick rates of mice discriminating VC and AC optogenetic stimuli. (Top) VCgo-ACnogo (n = 6). (Bottom) ACgo-VCnogo (n = 5). ∗∗∗p < 0.001; n.s., not significant, paired t test. Gray lines, lick rates of individual mice; black lines, average lick rates ± SEM.
(C) Schematic for testing the effect of VC or AC inactivation in the modality discrimination and conflict trials. Fluorescent images indicate the fluorophore-conjugated muscimol (FCM) injection sites in the VC and AC. Scale bars, 500 μm. Note that FCM expression covers most areas of VC and AC.
(D) Lick rates of mice discriminating a flashing visual stimulus from a pure tone auditory stimulus after muscimol injection into the VC (red) or AC (blue). Top, VISgo-AUDnogo (n = 7). Bottom, AUDgo-VISnogo (n = 7) (∗∗∗p < 0.001; n.s., not significant; paired t test). Black, no injection; red, muscimol injection into the VC; blue, muscimol injection into the AC; mean ± SEM. The three points on the right indicate the spontaneous lick rates of mice in catch trials (n = 4; mean ± SEM).
(E) Absolute changes in lick rates (|Δ Lick rate|) (left) or auditory dominance index in conflict trials (right) of the same mice in (D). Black, no injection; red, muscimol injection into VC (MUSVC); blue, muscimol injection into AC (MUSAC); dots, data from individual mice; lines, mean ± SEM (n.s., not significant, ∗∗∗p < 0.001, paired t test).
(F) Schematic for testing conflict resolution between actual visual and optogenetic AC stimulation in mice discriminating actual stimuli.
(G) Lick rates in response to VIS, AUD, VIS+AUD, or VIS+AC stimulation. Gray lines indicate AC stimulation using varying blue light intensities (1, 1.5, and 2.3 mW). The black line indicates pure tone stimulation. VISgo-AUDnogo, n = 5; AUDgo-VISnogo, n = 8; mean ± SEM; ∗∗∗p < 0.001; n.s., not significant; paired t test.
Neuron  , e6DOI: ( /j.neuron )
Copyright © 2017 Elsevier Inc. Terms and Conditions

5. Figure 4 AC Stimulation Does Not Suppress VC Activity
(A) Schematic of multi-unit recordings in VC or AC with optogenetic stimulation.
(B) An example raster plot of AC multi-unit spikes during optogenetic stimulation of AC (ACstim), VC (VCstim), and both (Dualstim). Blue shading indicates the blue-light stimulation period.
(C) Firing rate (FR) changes at different depths in the AC induced by blue-light stimulation. Gray dots, individual recordings at a specific depth; horizontal and vertical lines, mean ± SEM (n = 20 recordings; n.s., not significant; p = 0.233; paired t test).
(D and E) Same as (B) and (C), but for VC recording (n = 37 recordings; horizontal and vertical lines, mean ± SEM; n.s., not significant; p = 0.156; paired t test).
(F) FR changes in the VC upon stimulation of the VC and AC with light of varying intensity (mean ± SEM; n = 4 mice; n.s., not significant; p = 0.709, paired t test).
Neuron  , e6DOI: ( /j.neuron )
Copyright © 2017 Elsevier Inc. Terms and Conditions

6. Figure 5
PTLp Activity Is Required for Auditory Dominance during Perception of Cross-Modal Conflicts
(A) (Left) Retrograde tracing schematic and a representative coronal view of the PTLp injected with red fluorescent retrobeads. (Right) Representative coronal views of the VC (left) and AC (right) labeled with retrobeads. Scale bars, 500 μm. Insets, magnified images of retrobead-labeled neurons in each area. Scale bars for insets, 25 μm.
(B) (Left) Anterograde tracing schematic. (Middle) Representative coronal views of the VC (top; red, tdTomato) and AC (bottom; green, eGFP). (Right) A representative view of the PTLp in which VC (red) and AC (green) axons converge. Scale bars, 500 μm.
(C) (Top) Schematic for muscimol inactivation of the PTLp. (Bottom) Representative coronal view of the FCM injection site in the PTLp. Scale bar, 500 μm.
(D) Lick rates of mice discriminating a flashing light from a sound (pure tone or white noise). (Top) VISgo-AUDnogo (n = 14; n = 10 for pure tone; n = 4 for white noise). (Bottom) AUDgo-VISnogo (n = 13; n = 9 for pure tone; n = 4 for white noise). Black, no injection; red, muscimol injection; blue, PBS injection; light lines, lick rates of individual mice; dark lines, average lick rates, mean ± SEM; ∗∗∗p < 0.001, n.s., not significant, paired t test.
(E) (Left) Changes in lick rates (Δ Lick rate) of the mice discriminating a flashing visual stimulus from a pure tone auditory stimulus after muscimol injection (red) or PBS injection (blue) into the PTLp. (Right) Auditory dominance index in conflict trials without injection (black, None), with muscimol injection (red, MUS), or with PBS injection (blue, PBS) into PTLp (dots, individual mice; lines, mean ± SEM; n.s., not significant, ∗∗∗p < 0.001, paired t test with Bonferroni correction).
See also Figure S3 and Movies S4 and S5.
Neuron  , e6DOI: ( /j.neuron )
Copyright © 2017 Elsevier Inc. Terms and Conditions

7. Figure 6
Feedforward Inhibition via PV+ Interneurons in the PTLp Mediates Auditory Dominance
(A) Schematic for the cell-type-specific retrograde tracing using pseudotyped rabies virus.
(B) Starter cells in the PTLp of PV-Cre, SST-Cre, VIP-Cre, and CaMKIIα-Cre mice. Green, AAV-mediated eGFP expression; red, pRV-mediated tdTomato expression; yellow, starter cells co-labeled with eGFP and tdTomato. Scale bars, 500 μm. Insets, magnified images of representative starter cells; scale bars for insets, 50 μm.
(C) Neurons labeled with the rabies virus (red) in the VC and AC. Scale bars, 500 μm.
(D) Quantification of normalized AC (blue) and VC (red) inputs to PV+ (n = 6), SST+ (n = 4), and VIP+ (n = 4) interneurons, and CaMKIIα+ (n = 5) excitatory neurons in the PTLp. Bars, mean ± SEM; ∗∗∗p < 0.001, ∗p = 0.026; n.s., not significant; paired and unpaired t test.
(E–G) Optogenetic inactivation of PV+ neurons in the PTLp impaired auditory-dominant perception of mice in conflict trials.
(E) Schematics for optogenetic inactivation of PV+ interneurons in PTLp of mice discriminating a flashing light from a pure tone. Micrographs, representative coronal views of ArchT-eGFP (top) and eGFP (bottom) expression in PV+ neurons (red). Scale bars, 500 μm. Insets, magnified views of representative PV+ neurons expressing eGFP. Scale bars for insets, 25 μm.
(F) Lick rates of PV::ArchT and eGFP control mice under the green laser stimulation. (Top) VISgo-AUDnogo (n = 8). (Bottom) AUDgo-VISnogo (n = 8). Black, without light stimulation (OFF; n = 8); light green, with optogenetic inactivation of PV+ neurons (ArchT ON; n = 5); dark green, eGFP control with light stimulation (eGFP ON; n = 3); light lines, lick rates of individual mice; dark lines, average lick rates (mean ± SEM); ∗p = 0.015; ∗∗p = 0.005; n.s., not significant; paired t test.
(G) (Left) Δ Lick rate of the same mice in (F) upon laser stimulation in visual-only (VIS) or auditory-only (AUD) trials. (Right) Auditory dominance index in conflict trials. Black, without laser stimulation (OFF); light green, laser stimulation of PV::ArchT (ArchT ON); dark green, laser stimulation of eGFP controls (eGFP ON); dots, individual mice; lines, mean ± SEM (n.s., not significant; ∗∗∗p < 0.001, paired t test with Bonferroni correction).
See also Figures S4 and S5.
Neuron  , e6DOI: ( /j.neuron )
Copyright © 2017 Elsevier Inc. Terms and Conditions

8. Figure 7
Neural Representation of Auditory-Visual Competition in the PTLp
(A) Schematic of multi-unit recordings in the PTLp with optogenetic activation of VC and AC.
(B) Correlation of firing rate changes induced by ACstim, VCstim, or Dualstim for all responsive PTLp neurons. Circles, individual VC&AC-responding (green), AC-selective (blue), VC-selective (red), and dual-responding (black) neurons with increased (filled) or decreased (open) firing rate changes; lines, linear regression for the population.
(C–F) Number (top) and mean firing rate changes (middle and bottom; mean ± SEM) of PTLp neurons classified according to their responses to optogenetic stimulation of the AC (ACstim), the VC (VCstim), or both (Dualstim). Inc (middle), neurons with increased firing rate changes; Dec (bottom), neurons with decreased firing rate changes. Each column indicates neurons that respond to both VCstim and ACstim (C; VC&AC-responding; Inc, n = 82; Dec, n = 10; ∗∗p = ), only ACstim (D; AC-selective; Inc, n = 48; Dec, n = 11; ∗p = 0.048, ∗∗p = ), only VCstim (E; VC-selective; Inc, n = 34; Dec, n = 8; ∗∗p = , ∗p = 0.017, 0.013), and only Dualstim (F; Dual-responding; Inc, n = 8; Dec, n = 7; ∗p = 0.013, 0.021, ∗∗p = ). Bars, mean ± SEM; ∗∗∗p < 0.001; n.s., not significant; paired t test with Bonferroni correction.
(G) Schematic of multi-unit recordings in the PTLp with optogenetic inactivation of PV+ interneurons.
(H) Multisensory modulation index of AUD-selective and VIS-selective neurons (mean ± SEM). Green thunderbolts indicate green laser stimulation. (Left) Laser stimulation of PV+ neurons expressing ArchT-eGFP (ArchT; AUD-selective, n = 58; VIS-selective, n = 51). (Middle) Laser stimulation of PV+ neurons expressing eGFP (eGFP control; AUD-selective, n = 19; VIS-selective, n = 12). (Right) Laser stimulation of naive neurons (No expression; AUD-selective, n = 39; VIS-selective, n = 36). ∗∗∗p < 0.001; n.s., not significant; paired t test with Bonferroni correction.
(I) Schematic model of auditory-to-visual feedforward inhibition mediated by PV+ interneurons in the PTLp.
See also Figure S6.
Neuron  , e6DOI: ( /j.neuron )
Copyright © 2017 Elsevier Inc. Terms and Conditions

9. Figure 8
Activity of PV+ Neurons in the PTLp Is Necessary and Sufficient for Auditory Dominance over Vision in Audiovisual Conflicts
(A) Schematic for optogenetic activation of PV+ neurons in the PLTp.
(B) Optogenetic activation of PV+ neurons in the PTLp enhanced the incomplete auditory-dominant perception produced by mild AUD stimuli (A2) presented with stronger VIS stimuli (V3–V8). (Top) VISgo-AUDnogo (n = 3). (Bottom) AUDgo-VISnogo (n = 4). Black, without optogenetic stimulation; blue, with optogenetic stimulation; lines, means ± SEM; ∗p = 0.049, ∗∗p = , paired t test with Bonferroni correction.
(C) The task design for auditory-visual dual discrimination. Two types of visual stimuli (vertical versus horizontal drifting gratings) and two types of auditory stimuli (5 versus 10 kHz pure tones) were used for go/no-go tasks. After training mice to discriminate both visual and auditory stimuli, we presented congruent audiovisual stimuli (VISgo + AUDgo or VISnogo + AUDnogo) or incongruent audiovisual stimuli (VISgo + AUDnogo or VISnogo + AUDgo) in test trials.
(D) Schematic for optogenetic inactivation of PV+ neurons in the PTLp.
(E) Correct rates of PV::ArchT mice discriminating both visual and auditory stimuli with or without laser stimulation (n = 8). Green thunderbolts indicate the laser stimulation. Correct rates were quantified based on the auditory stimulus type (AUDgo versus AUDnogo). Bars, means ± SEM; ∗p = and ; n.s., not significant; paired t test with Bonferroni correction.
See also Figures S7 and S8.
Neuron  , e6DOI: ( /j.neuron )
Copyright © 2017 Elsevier Inc. Terms and Conditions