Contributions of Distinct Auditory Cortical Inhibitory Neuron Types to the Detection of Sounds in Background Noise

Location of optical fibers. Coronal brain slice showing the location of optical fibers implanted above the auditory cortex, such that light was concentrated on the AUDp. The primary field is surrounded by a dorsal and a ventral field (AUDd and AUDv, respectively). Laser powers were chosen such that areas outside the auditory cortex were unlikely to be affected. SS, Somatosensory cortex; TEa, temporal association area; HC, hippocampus. Scale bar, 1 mm.

Inactivation of the auditory cortex impaired the detection of signals masked by noise. A, Top, Optogenetic stimulation in freely moving PV::ChR2 mice. Laser activation of PV⁺ cells in the auditory cortex silences auditory cortical activity. Bottom, For a random 25% of trials in each session, the laser turned on at the same time as the sound and lasted for 100 ms after the sound turned off. B, Example psychometric curve from one PV::ChR2 mouse trained to report the presence of the tone by going to the right (trials pooled across multiple sessions). The activation of auditory cortical PV⁺ cells reduced the probability of the mouse reporting the presence of the tone for all SNRs. Error bars show 95% confidence intervals. C, Like B, but for a PV::ChR2 mouse trained to go to the left to report the presence of the tone. As before, the activation of auditory cortical PV⁺ cells reduced the probability of the tone being reported, despite the opposite action being required to report tones. D, Silencing the auditory cortex during the signal detection task significantly reduced sensitivity to the pure tone signal (p = 0.0117, Wilcoxon signed-rank test; N = 8 mice). Each pair of points corresponds to one mouse. E, Silencing the auditory cortex during the signal detection task significantly reduced the hit rate (p = 0.0117, Wilcoxon signed-rank test). F, Silencing the auditory cortex during the signal detection task significantly reduced the false alarm rate (p = 0.0251, Wilcoxon signed-rank test). G, The effect of silencing the auditory cortex (No AC) on d′ was not correlated with the effect of visual distraction (Control) on d′ (p = 0.593, r = 0.224, linear correlation coefficient). Each point corresponds to one mouse. H, The effect of silencing the auditory cortex on hit rate was not correlated with the effect of visual distraction on hit rate (p = 0.157, r = 0.710, linear correlation coefficient). I, The effect of silencing the auditory cortex on false alarm rate was not correlated with the effect of visual distraction on false alarm rate (p = 0.589, r = 0.227, linear correlation coefficient).

Inactivation of the auditory cortex yielded consistent impairments across sessions. Signal detection performance (d′) during each behavioral session (dots) compared with the pooled performance across all trials (horizontal lines) for each PV::ChR2 mouse (M1–M8).

Inactivation of the auditory cortex did not affect the timing of behavior. A, Silencing the auditory cortex during the signal detection task did not affect the time spent sampling the sound in the center port (p = 0.0925, Wilcoxon signed-rank test; N = 8 mice). Each pair of points corresponds to one mouse. B, Silencing the auditory cortex during the signal detection task did not affect the time spent moving toward the reward port (p = 0.779, Wilcoxon signed-rank test). Each point corresponds to one mouse. C, The effect of silencing the auditory cortex (No AC) on time spent sampling the sound in the center port was not correlated with the effect of visual distraction (control) on sampling time (p = 0.675, r = 0.177; linear correlation coefficient). D, The effect of silencing the auditory cortex on time spent moving toward the reward port was not correlated with the effect of visual distraction on time spent moving toward the reward port (p = 0.686, r = –0.171; linear correlation coefficient).

Contributions of distinct inhibitory neuron types to auditory signal detection. A, Inactivation of PV⁺ cells during the signal detection task. Laser was presented for a random 25% of trials each session. B, Example psychometric curve from one PV::ArchT mouse (trials pooled across multiple sessions). Inactivation of auditory cortical PV⁺ cells decreased the probability of the mouse reporting hearing the signal when it was present, though false alarm rate was unaffected. Error bars show 95% confidence intervals. C, Inactivating auditory cortical PV⁺ cells during the signal detection task significantly reduced sensitivity to the signal (p = 0.0046, Wilcoxon signed-rank test; N = 13 mice). D, Inactivating auditory cortical PV⁺ cells during the signal detection task significantly reduced the hit rate (p = 0.033, Wilcoxon signed-rank test). E, Inactivating auditory cortical PV⁺ cells during the signal detection task did not significantly affect the false alarm rate (p = 0.861, Wilcoxon signed-rank test). F, The effect of inactivating auditory cortical PV⁺ cells (No PV) on d′ was not correlated with the effect of visual distraction (control) on d′ (p = 0.609, r = 0.157; linear correlation coefficient). G, The effect of inactivating auditory cortical PV⁺ cells on hit rate was not correlated with the effect of visual distraction on hit rate (p = 0.657, r = –0.136; linear correlation coefficient). H, The effect of inactivating auditory cortical PV⁺ cells on false alarm rate was not correlated with the effect of visual distraction on false alarm rate (p = 0.981, r = –0.0073; linear correlation coefficient). I, Inactivation of SOM⁺ cells during the signal detection task. Laser was presented for a random 25% of trials each session. J, Like B, but for one SOM::ArchT mouse. Inactivation of auditory cortical SOM⁺ cells reduced the probability of the mouse reporting hearing the signal when it was present, though false alarm rate was unaffected. K, Inactivating auditory cortical SOM⁺ cells during the signal detection task significantly reduced sensitivity to the signal (p = 0.0051, Wilcoxon signed-rank test; N = 10 mice). L, Inactivating auditory cortical SOM⁺ cells during the signal detection task significantly reduced the hit rate (p = 0.0051, Wilcoxon signed-rank test). M, Inactivating auditory cortical SOM⁺ cells during the signal detection task did not significantly affect the false alarm rate (p = 0.878, Wilcoxon signed-rank test). N, The effect of inactivating auditory cortical SOM⁺ cells (No SOM) on d′ was not correlated with the effect of visual distraction (control) on d′ (p = 0.546, r = 0.218; linear correlation coefficient). O, The effect of inactivating auditory cortical SOM⁺ cells on hit rate was correlated with the effect of visual distraction on hit rate (p = 0.0494, r = 0.633; linear correlation coefficient). P, The effect of inactivating auditory cortical SOM⁺ cells on false alarm rate was not correlated with the effect of visual distraction on false alarm rate (p = 0.0725, r = 0.590; linear correlation coefficient).

Effects of disrupting auditory cortical inhibition on each behavioral session. Signal detection performance (d′) during each behavioral session (dots) compared with the pooled performance across all trials (horizontal lines) for each PV::ArchT mouse (PV1–PV13) and each SOM::ArchT mouse (SOM1–SOM10). B, Baseline (laser off); M, manipulation (laser on).