Article Figures & Data
Figures
- Figure 1.
Transient optogenetic suppression of auditory cortex impaired pure tone discrimination. A, Transient laser suppression used a 500-ms constant laser pulse at 300 mW/mm2 (9.5-mW total power) that completely overlapped with the 500-ms pure tone stimulus. Laser was delivered on a random 10% subset of trials. B, Electrophysiological validation of optogenetic suppression in separate mice. Left, Responses to best-frequency tones averaged across 46 tone-responsive neurons and 50 repetitions under control conditions (black) and during a 500-ms constant laser pulse at 200 mW/mm2 (cyan). Right, Effect of illumination on spontaneous activity during silence, averaged across 90 neurons and 50 repetitions. Both tone responses and spontaneous activity were significantly suppressed. C, Progression through training steps is shown for five example mice, aligned to the start of training. Performance was computed in a 200-trial sliding window. D, The first 10,000 total trials in the suppression condition (step 6) for an example mouse (mouse rt.15) show a consistent effect of the laser, with minimal effects on bias. A 200-trial sliding window was used to separately compute accuracy over time on control trials (black) and laser trials (cyan). Since this would require ∼2000 total trials before having a sufficient sample to measure laser performance, we used a smaller window of 50 trials to measure performance for the initial 100 laser trials, and a 200-trial window for subsequent laser trials. We also measured the bias on laser trials (gray), with 50% corresponding to equal left and right responses. E, Performance in overall percentage correct for each mouse for 10,000 total trials is shown by connected dots for control and laser trials. Bars represent mean performance for each trial type across all nine mice. Inset at right shows a Gardner–Altman estimation plot of effect size. The dot shows the paired Cohen’s d, the vertical bar shows the 95% confidence interval, and the gray shading shows the bootstrap sampling distribution of paired Cohen’s d. F, To show stimulus-related effects, we used ROC analysis to separate each mouse’s performance into “hit rate” (percentage correct on 13-kHz trials) and “false alarm rate” (100 − percent correct on 4-kHz trials). The dashed line represents chance performance (50% correct), ranging from 100% leftward (4 kHz) responses at the lower left corner to 100% rightward (13 kHz) responses at the upper right corner. Performance for each mouse on control trials is represented with a large open circle, connected by an arrow to the laser performance for the same mouse (cyan circle). Perfectly accurate behavior corresponds to the upper left corner, and any effect of the laser toward the dashed line indicates a decrease in performance, which is seen for all mice. Most mice showed little response bias, corresponding to the direction toward the lower right corner (perpendicular to the dashed line), but two mice showed stronger laser effects on response bias (one toward the lower left corner, and one toward the upper right corner). G, H, Mice that did not express ChR2 showed no significant effects of laser illumination. Format as in E, five mice, 2000–4000 total trials. Rewards on laser trials were either delivered randomly (H), or normally (G; i.e., for the correct response corresponding to the stimulus).
- Figure 2.
Electrolytic lesions of auditory cortex did not impair pure tone discrimination. A, The section of maximal lesion extent for each mouse (n = 5) is shown in cyan/blue, overlaid onto the boundaries of auditory cortex (red). B, Performance before and after lesion for an example mouse (mouse 6685). Vertical black dashed line indicates graduation to the final training step (6). In the 700 pre-lesion trials, overall performance was 94.25% correct. The lesion is indicated by the red dashed line at trial zero. In the 700 trials following the lesion, the mouse performed at 95.1%, showing a lack of impairment. Performance is computed in a 100-trial sliding window. Chance performance is 50%. C, Performance for each of five mice is shown in different colors. Format as in B. Traces are aligned to the time of the lesion, indicated by the red dashed line. Note that a median of 4 h elapsed at the time of lesion (red dashed line). Heavy black line shows mean across mice for 700 trials before and after lesion. D, Mean performance across all mice on the 700 trials pre-lesion and 700 trials post-lesion are represented as bars, with the performance of individual mice indicated by connected dots. Inset at right shows estimated effect size, format as in Figure 1E. E, Lesion extent for individual mice is plotted against their pre-lesion performance (black) and post-lesion performance (red). F, ROC analysis (same format as Fig. 1F). Open circles indicate pre-lesion performance for each mouse, which are connected to red circles indicating post-lesion performance. Lesions had negligible effects on accuracy or response bias.
- Figure 3.
Cycle suppression. A, We designed a cycle suppression protocol in which a 100-Hz laser pulse train (300 mW/mm2, 5 ms on/off or 50% duty cycle) was on for 20 consecutive trials, and then off for 20 consecutive trials, in alternating blocks. Blocks of 20 trials had median durations of 1–3 min. B, Electrophysiological validation of optogenetic suppression using a 100-Hz laser pulse train in separate mice. We used a fixed duration of 100 s (corresponding to the approximate duration of 20 trials) and measured the effect on spontaneous activity in silence (10 repetitions, n = 153 neurons). The suppression of spontaneous activity was significant and stable throughout the duration of the pulse train. C, Performance on each trial within the cycle, averaged across a minimum of 119 blocks, is shown for individual mice in gray, with the mean across all mice in black (n = 10 mice). Cyan bars indicate the trial blocks with the laser on. The data are duplicated to display two complete cycles. Mice showed a sharp initial drop in performance on the first trial in an on-cycle, followed by a rapid but partial recovery. Performance did not return to baseline performance by the end of the 20-trial block. D, Median reaction time on each trial within the cycle, averaged across all cycles, is shown for individual mice in gray, with the median across all mice in black. Reaction times were unaffected by laser. E, Median 20-trial cycle duration for the on and off cycles of each mouse. Median on-cycle duration and off-cycle duration were tightly correlated and nearly equal, indicating that mice showed individual differences in trial rate but no difference between response rates during on-cycles and off-cycles. Gray line is a linear regression, dashed line is unity. F, After testing on 20-trial cycles, we re-tested all 10 mice with transient 500-ms suppression (as in Fig. 1A). We compared the effect size for cycle suppression to the effect size for transient 500-ms suppression, for each mouse. We quantified effect size for cycle suppression as the drop in performance from the last trial in the off-cycle to the first trial in an on-cycle. Effect size was tightly correlated between the two suppression protocols, indicating that mice showed stable individual differences in the effects of suppression. The effect size for transient suppression was roughly double the effect size for cycle onset. Gray line is a linear regression. G, To test cycle suppression over a longer time course, we extended the cycle protocol shown in A to span 200 trials with the laser pulse train on, alternating with 200 trials with laser off. Performance on each trial within the cycle is shown in same format as B, except that data are binned (in 10-trial bins). Mice did not show a full recovery to control performance even after 200 trials (median duration 20.5 min).
- Figure 4.
Sustained optogenetic suppression impairs pure tone discrimination. A, Performance for each of five mice before and after 4 h of sustained optogenetic suppression (format as in Fig. 2C; each mouse is shown in a different color). Note that 4 h elapsed during sustained suppression (cyan dashed line), after which suppression was maintained throughout a second session of behavior. We repeated this experiment one to four times per mouse, resulting in 15 sessions for the five mice. Because mice ran a different number of trials in each session, we averaged across mice by including only trials contained in all 15 sessions (black line, which is limited to the duration of the shortest session). The gray line shows the average across the 14 sessions not including the shortest one. B, Mean performance before and after sustained suppression. Format as in Figure 2D. Connected dots show each mean performance for each of 15 sessions (using trials from the black segments in A). Inset at right shows estimated effect size, format as in Figure 1E.
Tables
Training step Description Advancement criterion 1. Surgery Fiber implantation 1 d of water restriction after recovery 2. Free drinks Ports give water, no stimulus Trial rate (cannot trigger the same port repeatedly) 3. Requests rewarded Rewards given for center-port trial requests and for correct responses Trial rate (7 trials in 30 s) 4. Only correct responses rewarded Rewards only given for correct responses 400 trials completed 5. Longer penalty Increased timeout for incorrect responses, no correction trials Sustained performance > ∼85% 6. Optogenetic suppression Laser on for 10% of trials N/A
In this issue
