Layer 5 and 6b extratelencephalic neurons encode distinct sound features in auditory cortex

Abstract

Extratelencephalic (ET) neurons in layers (L)5 and 6b of the auditory cortex (ACtx) provide major corticofugal outputs to subcortical structures and contribute to auditory learning and experience-dependent plasticity. Although both ET subtypes innervate overlapping downstream targets, they differ in morphology, intrinsic physiology, and molecular identity, suggesting distinct functional roles. However, their in vivo response properties remain incompletely characterized. Here, we used a projection-defined viral strategy to express GCaMP8s selectively in L5 and L6b ET neurons in the ACtx of male and female mice and recorded calcium activity during presentation of diverse acoustic stimuli. L5 ET neurons predominantly exhibited sound-evoked excitation, higher response sparseness, and greater trial-to-trial reliability than L6b ET neurons; these differences were most pronounced for pure tones and sinusoidally amplitude-modulated (sAM) noise. In contrast, L6b ET neurons frequently exhibited sound-evoked suppression, particularly for more complex stimuli (including sAM noise and spectrotemporal ripples). Unsupervised clustering identified separable temporal response motifs and tuning profiles. L5 ET neurons favored single-peaked frequency tuning and monotonically increasing intensity tuning, whereas L6b ET neurons exhibited complex frequency tuning, monotonically decreasing intensity tuning, and stronger pairwise functional coupling (noise correlations). Together, these results support complementary corticofugal processing streams, with L5 ET neurons conveying more selective acoustic feature representations and L6b ET neurons conveying more integrative corticofugal signals. Significance Statement The primary auditory cortex sends extensive descending projections to subcortical structures, shaping how sounds are processed. These projections arise from distinct extratelencephalic (ET) neurons in cortical layers 5 and 6b, yet how these populations encode sound has remained unclear. Using projection-specific two-photon imaging in awake mice, we show that layer 5 ET neurons convey sparse, reliable, and selective representations of acoustic features, whereas layer 6b ET neurons more often show sound-evoked suppression, complex tuning, lower reliability, and stronger coupling to one another. These findings reveal two complementary descending channels, one relaying precise feature information and the other broader, integrative signals, that together enrich cortical control over subcortical auditory processing.