Distinct nonlinear spectrotemporal integration in primary and secondary auditory cortices

Abstract

Animals sense sounds through hierarchical neural pathways that ultimately reach higher-order cortices to extract complex acoustic features, such as vocalizations. Elucidating how spectrotemporal integration varies along the hierarchy from primary to higher-order auditory cortices is a crucial step in understanding this elaborate sensory computation. Here we used two-photon calcium imaging and two-tone stimuli with various frequency-timing combinations to compare spectrotemporal integration between primary (A1) and secondary (A2) auditory cortices in mice. Individual neurons showed mixed supralinear and sublinear integration in a frequency-timing combination-specific manner, and we found unique integration patterns in these two areas. Temporally asymmetric spectrotemporal integration in A1 neurons enabled their discrimination of frequency-modulated sweep directions. In contrast, temporally symmetric and coincidence-preferring integration in A2 neurons made them ideal spectral integrators of concurrent multifrequency sounds. Moreover, the ensemble neural activity in A2 was sensitive to two-tone timings, and coincident two-tones evoked distinct ensemble activity patterns from the linear sum of component tones. Together, these results demonstrate distinct roles of A1 and A2 in encoding complex acoustic features, potentially suggesting parallel rather than sequential information extraction between these regions.

Figure 1.. Quantification of spectrotemporal interaction using two-tone stimuli.

(a) Two-photon imaging setup. Auditory areas were first mapped by intrinsic signal imaging, which was used to guide the chronic window implantation. Bottom left, thresholded intrinsic signal responses to pure tones superimposed on cortical vasculature imaged through the skull. Bottom right, in vivo two-photon image of L2/3 neurons in A1. (b) Sound stimulus schematic showing the relationship between frequency and time for each of the two 20 ms tones. The Center tone was matched to the best frequency of the neuronal population within the field of view. (c) Responses to each dF-dT pair and single tone presentations in a representative A1 (top) and A2 (bottom) neuron. Traces are average across five trials. Inset schematics show the spectrotemporal relationship between the two presented tones. (d) Calculation of LI for neuronal responses marked with arrowheads from (c). LI > 0 (red arrowhead) indicates supralinear integration of two tones compared to the linear sum of both frequency components, whereas LI < 0 (blue arrowhead) indicates sublinear integration. (e) Spectrotemporal interaction maps showing the LI across dF-dT pairs for neuron 1 (A1) and neuron 2 (A2).

Figure 2.. Spectrotemporal interaction maps of A1 and A2 cells in representative mice

(a) Intrinsic signal image superimposed on cortical vasculature imaged through a glass window in a representative mouse. Yellow square represents the A1 two-photon imaging field of view. (b) Spectrotemporal interaction maps for example A1 neurons in the same mouse as (a) show mixed supralinear and sublinear interactions across dF-dT pairs (c) Average spectrotemporal interaction map across all A1 neurons in the same mouse. n = 121 neurons (d) Intrinsic signal image in a representative mouse with A2 two-photon imaging. (e) Same as (b), but for example neurons in A2. (f) Same as (c) but across A2 neurons in the same mouse as (d) and (e). n = 35 neurons.

Figure 3.. A1 and A2 neurons integrate two-tone stimuli with distinct spectrotemporal combinations.

(a) Spectrotemporal integration maps across all A1 and A2 cells. A1, n = 9 mice, 809 responsive cells. A2, n = 8 mice, 357 responsive cells. (b) Left, summary data comparing normalized response magnitudes in A1 and A2. Right, summary data comparing linearity index in A1 and A2. A1: n = 2596 cell-dF pairs, A2: n = 1574 cell-dF pairs. Data are mean ± SEM. (c) Fraction of neurons with statistically significant supralinearity (facilitative interaction) and sublinearity (suppressive interaction) for each dF-dT pair in A1. (d) Same as (c), but for A2. (e) A1 and A2 neurons classified by their preference for two-tone timings. The fraction of neurons preferring coincident over shifted stimuli was significantly higher in A2 than A1, Chi-square test, p = 1.11 × 10⁻¹⁶. (f) A cumulative probability plot of asymmetry index for all sound-responsive cells in A1 and A2. ***p = 9.60 × 10⁻⁷, Wilcoxon rank sum test.

Figure 4.. Asymmetry in suppressive spectrotemporal interaction accounts for FM direction selectivity.

(a) Top, FM sweep tuning of a representative L2/3 pyramidal cell in A1. Traces are average responses across five trials. Insets at the bottom show the schematics of frequency versus time representations. Bottom, a two-tone spectrotemporal interaction map for the same neuron. Yellow boxes: Upward region, blue boxes: Downward region. (b) Fraction of responsive cells at six absolute FM rates in A1 and A2. A1: n = 6 mice, 993 cells; A2: n = 7 mice, 434 cells. ***p < 0.001 for all speeds, Chi-square test with Bonferroni correction. (c) Average (solid line) and SEM (shading) of absolute DSI at each FM rate in A1 and A2. A1: 391 sweep-responsive cells; A2: n = 241 sweep-responsive cells. (d) Top, DSI of pyramidal cells averaged across 10-40 oct/sec has a strong correlation with linearity index bias for suppressive interactions (Bias_supp), but not facilitative interactions (Bias_fac). p = 0.0006, two-sided t-test. Red line, regression curve. n = 220 cells responsive to both FM sweeps and two tones. Bottom, p and R values of the correlation between DSI and linearity index bias separated by FM rate. *p < 0.05, **p < 0.01. p values are adjusted for multiple comparisons with Bonferroni correction. (e) Same as (d), but for A2. n = 175 cells responsive to both FM sweeps and two tones.

Figure 5.. Ensemble activity patterns show distinct integrative functions between A1 and A2.

(a) Out of single-tone non-responsive neurons in L2/3, 53% (A1) and 53.7% (A2) responded to either coincident or shifted two tones. (b) Schematic showing ensemble neuronal activity vectors in high-dimensional space for two-tone, individual tones (“Tone_cent” and “Tone_dF”). and the linear sum of individual tones (“linear sum’). (c) Correlation coefficient between single-tone versus two-tone (black lines) and between linear sum versus two-tone (red lines) representations across dTs in A1 (left) and A2 (right). Solid line: average, shading: SEM. (d) Box plots showing correlation coefficients between two-tone representation and linear sum representation separately for coincident and shifted two-tone stimuli. Box: 25th to 75th percentiles. Whiskers: 99.3% coverage. Red lines: median. Blue crosses: outliers. Shifted: n = 64 dF-dT pairs, Coincident: n = 8 dF-dT pairs. *p < 0.05, ***p < 0.001, Two-way ANOVA followed by Tukey’s honest significance test.